A novel way to massively improve the SQ of computer audio streaming59 minutes ago, dtossan said:
Hey Nenon, saw your original post back in july. Glad you remind us.
A rough est to make a 1 foot Mundorf Silver/Gold cable would be around USD 100, right.
That sounds like an amazing step up compared to a 1 foot Ghent Neotech 7N for US$79.
What other parts to buy to make a 1 foot JSSG360 version? Will PM you.
I don't want to derail this thread with DIY cables, there are plenty of other threads on that, but for completeness, here is what you need for 1 foot cable:
Wire: https://www.partsconnexion.com/MUNDORF-72180.html - 2 feet
Cotton: https://www.partsconnexion.com/COTTUBE-72533.html - 2 feet
Whatever connectors you need, i.e. https://www.vhaudio.com/oyaide-dc.html
Techflex cable sleeving to decorate the cable: https://www.vhaudio.com/heatshrink.html#techflex - 1 foot + a little extra
And some heatshrink: https://www.vhaudio.com/heatshrink.html#heatshrink - 2 pieces to go over the connectors
I use WBT solder: https://www.vhaudio.com/heatshrink.html#solder - 1 foot is plenty
For the JSSG360 shielding, you would need some tinned copper braid. Plenty of options on that:
the 1/4'' should be fine for a two wires cable.
and some plumbers tape: https://www.amazon.com/gp/product/B06XCWQ4D2/
There is a lot of info how to do the shielding, please search the forum.
Making the cable:
Cut two wires to your preferred length.
Insert each wire in its own cotton sleeve.
Slightly twist the two wires together. BTW, I just realized I have not twisted mine.
Insert the twisted wires in the first tinned copper braid.
Wrap the tinned copper braid with plumbers tape.
Insert in the second tinned copper braid.
Do the JSSG360 shielding - connect the first braid and the second braid on each side.
Insert in the techflex cable sleeving.
Add heatshrink at both ends.
Solder the connectors (don't forget to insert the back part of the connector before soldering).
Heat the heatshrink in place.
Hope that helps. Please share your impressions if you try it.
A novel way to massively improve the SQ of computer audio streaming7 hours ago, ElviaCaprice said:
Great job Roy,
What a contrast to my economical sCLK-EX direct USB build.
NUC (used $100)
tXUSBexp (used $150)
Chord 2Qute (used $900)
2-old HDD enclosures ($20)
1 PPA SATA chord (used $20)
Omega Super 8XRS Speakers ($2420)
2-DIY canare speaker/interconnect cables ($16)
1KVA Topaz Isolation Transformer (used $200)
1 tripplite 20V power strip, 12 outlets ($30)
2- LPS-1's ($760)
Paul Hynes SR7 MR4 ($1950)
5TB 2.5" HDD - (Ebay $130)
sCLK-EX loaded, installation, shipping, master clock option - ($1185)
sCLK OCX-10 Reference (once released) - ($1500)???
USB mouse and keyboard - ($15)
ASUS monitor I've been using for a few years, all PC's attached.
I do my own optimization on Windows 10, no need for others software, other than JRiver.
Shocking!! still comes to ($9431) give or take on final cost of SOtM master clock and some DIY Canare DC chords. And I thought I was being economically wise in choosing my components carefully for the biggest bang in SQ. Nice little exercise for others to do from Server/streamer to Speaker or headphones, what ever is your preferred listening means.
Thanks, Mark. Your system is well thought out and so good for you. In my case, the cost was painfully higher mainly because of so many things that I bought that ended up never making the final cut and so I have a lot of money invested in parts that are now just lying around (i.e. motherboards, CPUs, various RAM, various PCIe cards, USB hubs, the Adnaco, network switches, FMCs, routers, etc.). This is the price I've paid for knowledge but at least I've now satisfied just about all of my curiosities.
A novel way to massively improve the SQ of computer audio streaming13 hours ago, Johnseye said:
Have you tried or considered Barry Diament's roller bearing isolation solution?
No, I have never heard of those but I do find ball bearing-type solutions from the likes of Stillpoints and Symposium Acoustics to generally be more effective then the rubber or sorbothane-type footers.13 hours ago, Johnseye said:
What AO filters and signatures are you using?
It depends. For large orchestral music or music with a lot of atmosphere, I am liking 4D. It provides the most depth. However, when I want tighter imaging, I'll go down to 3C. Signatures 1 and 2 sound too mechanical for my setup.13 hours ago, Johnseye said:
Why did you replace both ethernet clocks on the motherboard?
Because I could do so while only utilizing 1 clock from my sCLK-EX board. I don't plan to build 2 servers although I have 2 listening rooms (home office with desktop setup and my main listening room). What I envision eventually is to bridge both LAN ports and then run CAT6/7 cabling from one of these bridged ports directly to my home office where I will install a SOtM trifecta. That way, both systems can benefit from this server. This is where the Adnaco solution could have worked well, however, my server only has 1 PCIe slot.13 hours ago, Johnseye said:
Do you believe the sound of the extracted data of the CDs are improved by your configuration?
Yes. While it makes no sense, it would appear that what's good for streaming is also good for CD ripping. Whether its due to noise that gets embedded into the stream that makes its way into the rip, I'm not sure.
Paul Pang advocates a certain type of DVD-ROM drive for ripping CDs, specifically an ASUS model with a MediaTek chip that utilizes no internal clock of its own (presuming this clock adds noise to the stream).
I bought a portable USB version of one of these these drives and I will run some tests to see if it makes a difference.13 hours ago, Johnseye said:
Have you incorporated any type of grounding?
Yes. As recently posted, I have tried grounding my DAVE via an Entreq Poseidon with Atlantis grounding cables via DAVE's XLR output, a Synergistic Research grounding block with HiDef grounding cable and most recently with Sound Galleries latest D2 grounding blocks and they made no difference. I have 2 of these D2 grounding blocks and have applied them to my music server but also to the spare USB port on my tX-USBultra and they result in improvements in those positions although the bigger difference is on my server.
A novel way to massively improve the SQ of computer audio streaming8 hours ago, seeteeyou said:
Many thanks for posting everything in great detail.
Lee should have mentioned that fiber optic would add jitter and that's why SOtM would prefer filtering instead of isolation.
Is that why you'd rather forgo the option of Adnaco-S1B even if that were giving you extra PCIe slots plus the advantage of distancing tX-USBexp cards from other components?
Anyways, it's already a challenge to cram everything on top of Tranquility Base XL UEF so that's really no biggie.
BTW, you've gotta take a picture when Rob Watts drops his jaw after listening to your latest and greatest rig.
Yes, you're correct. Lee does not like optical solutions due to high jitter. As it was Lee that personally modified the Adnaco I sent him, he got a chance to listen to it and even with its 3 clocks replaced, he told me he thought it was a step back because It was sounding flat. It was at that point that he encouraged me to try his tX-USBexp.
I knew I needed to hear it for myself and so he went ahead and completed the mod of the Adnaco per my request which included the replacing of several capacitors. Once I received the modified Adnaco back, I felt it had definitely improved and I liked the smoother and more solid image it was portraying. Around the time I received the modified Adnaco back, I was also able to audition the ISO Regen and I found the ISO Regen to be so good that in its stock form, it was equivalent to the modified Adnaco. Around this time, I was able to access an SOtM tX-USBexp (non Ultra) and so I compared it to the modified Adnaco. I found Lee to be correct, even the non-ultra tX-USBexp was sounding more impressive to me than the modified Adnaco and when I added the ISO Regen to the tX-USBexp, this combo was easily better than the Adnaco. Better yet, tX-USBexp + ISO Regen only utilized 2 clocks while the Adnaco solution by itself utilized 3 clocks although this combo costs more than the Adnaco.
Having said that, the Adnaco definitely has merit and it does some nice things. For those who have a thin sounding system, you might really like what the Adnaco does and regardless of jitter, optical isolation definitely lowers the noise floor. For those who wish to keep their server in a different room or even a different building, you can purchase very long optical cables for very little money and still get excellent sound while remaining "straight USB." If you follow the Adnaco with a tX-USBultra, you come reasonably close to what the SOtM trifecta provides.
A novel way to massively improve the SQ of computer audio streaming
Several have asked about the specifics of my current build. It has been in continual flux until recently but finally, I have reached my final build. I don't claim it to be the best there is but it is the best that I know how to put together and the best that I have thus far heard. To be honest, this is not what I envisioned it would look like when I first started this thread back in January but for those who have followed along, as you know, there were exciting and unexpected revelations that came to light and so this digital front end has become a much more ambitious build than I initially expected it to be. At so many points, it was sounding so good that I felt like finalizing my build but it was hard to stop when each successive gain seemed so significant. Along the way, many of you made constructive suggestions, some that I had never considered, and for that, I am grateful. They didn't always work out but the lessons learned and the perspective gained have proven to be invaluable. To that extent, I consider this build to have been a community effort and so I am happy to share what I have learned.
Server chassis - Streacom FC9 Alpha ($285)
I have used Streacom cases on numerous occasions in the past as I have found them to be attractive, well-constructed and minimally resonant. I chose the version that can hold a CD-ROM drive but eventually decided that I wanted nothing that would vibrate within the chassis as I felt this might adversely impact the sCLK-EX board and so I opted for an external CD-ROM drive to rip my CDs instead. This also provided me much needed space inside. If I could do it over, I would have gone with the cleaner look of the FC9 Alpha without the CD slot as pictured below (although I went with black instead of silver):
Streacom makes smaller cases but I went with this larger case for flexibility. As you can see in the photo, when used with a mini-ITX motherboard, I was able to install my sCLK-EX board adjacent to the motherboard which allowed me to use very short 15cm clock cables. Also installed adjacent to the motherboard is my Intel X25-E SATA II OS drive.
For those who want to go for the ultimate, Stillpoints makes standoffs that supposedly incorporate the same level of isolation as their Ultra Minis that could have been used to isolate my motherboard, SSD, and sCLK-EX board from the chassis. At $45 per standoff and since 12 standoffs were necessary, that would have cost me $540 for standoffs alone and so I passed. Instead, I bought a trio of Stillpoints Ultra Minis ($375 although I found a used set of 3 for $250) to use under the chassis and I have confirmed that these absolutely make a difference with respect to detail clarity (a larger difference than the inexpensive Black Raviolis I was previously using) although the overall improvement is small.
SOtM eABS-200 paper ($120 for 1 sheet)
I have used this type of EMI-absorbing paper in the past with several builds, specifically the Stillpoints ERS sheets but I have found the SOtM eABS-200 to be considerably more impactful with regards to a calmer and blacker presentation. If you are finding your server to sound fatiguing and bright, you might find this paper to be very helpful. Unlike other products, this paper supposedly absorbs a broad range of EMI and converts it to heat, thereby actually removing EMI instead of just scattering it. This product is so powerful that Lee advised caution as he indicated it was was possible to use too much resulting in an overdamped sound. I applied it to most of my chassis, to my RAM, and to my SSD and I have heard only good things. I even applied it to the inside of my tX-USBultra chassis and its collective impact is very easy to hear. I have found it well worth the $120.
DFI BW171 motherboard ($331 which includes an embedded Celeron CPU)
Finding a suitable motherboard was extremely difficult and it became s a matter of finding the motherboard with the least compromises. DFI was willing to design a motherboard for me to my specifications, however, it would have cost $1,500 minimum to get the project off the ground. I discussed the prospect of doing this with Lee of SOtM while in Munich and while he expressed interest, he told me he believed he had already figured out how to filter out most of the noise from a noisy motherboard with his tX-USBexp card. While I suspect there would have been something to be gained by designing my own motherboard devoid of unnecessary noisy components and noisy switching regulators, the time, effort and expense didn't appeal to me and so I ended up going with Lee's tX-USBexp card instead.
Having built many servers over the years, none of the servers I had built could match the wonderful SQ I got from the $9k CAD CAT that I purchased in 2015. I opened this machine up and studied it carefully and realized that this was essentially a minimalist PC with detuned components (ie both the quad-core i7 and RAM were both detuned to run at 800MHz). At that time, for an i7-based PC with 4GB of RAM and 3TB of SSD storage to draw less than 20 watts was not a common thing to see. When I first got this machine, I went online to download some drivers and noticed that this thing ran as slow as molasses but boy, did this machine sound good. Heavily shielded SATA cables were used (made by Paul Pang) along with point-to-point DC cabling, EMI paper throughout, and finally a custom-specified PCIe-based USB card with a special clock. The whole thing was powered by a quad-rail custom HDPlex LPSU. I found this setup to be superior to a TotalDac d1-Server and to an Aurender N10 that I had on hand and so I bought the CAD CAT but I also knew this server could be improved upon (especially with regards to its power supply).
As far as choosing an actual motherboard for this build, based on my experience with my CAD CAT and as I had no intention of oversampling, I knew I wanted a low power setup. I was further convinced when I directly compared the SQ from a very powerful HP workstation (with dual Xeon CPUs, 64GB of RAM, and PCIe-based SSDs) against a powerful Mac Pro (with a 12-core Xeon, 64GB of RAM and 1TB of PCIe-SSD storage) and found that the SQ I got from a much less powerful stock Mac Mini with the OS running off of a low power SD card to be easily superior. With the Mac Mini modified to accept 12V power from a Paul Hynes SR7, it was no contest. The SR7 was such a difference maker that I chose to only look at options that allowed me to use my SR7.
Having tried various motherboards from Gigabyte, AsRock, SuperMicro and Dell, I ended up with the DFI BW171 for the following reasons:
1. SoC (System on a Chip) architecture that did away with the slower (higher latency) processor controller hub (PCH).
2. Mini-ITX form factor resulting in shorter data paths
3. Simple Celeron CPU with only 6w TDP
4. Could be powered from a 12V rail from my SR7
If there is a downside to these types of motherboards, they generally only contain 1 PCIe slot (usually a PCIe x4 or slower). If they contain an M.2 slot, they run off the slower SATA bus and not the faster (lower latency) NVMe bus.
There are other motherboards that could have fit the bill but I went with this company because they had presented me the option of designing a new motherboard from scratch for me if I found this motherboard to be unacceptable. Thus far, I have been very pleased. As an industrial motherboard, it has not only proven to be very reliable running 24/7 but even prior to modification, I found it to sound at least as good as my Mac Mini with SD card even though my OS was running off a noisier SSD. With this motherboard running Windows Server 2016 + Audiophile Optimizer (something that a Mac Mini with SD card cannot run), this machine was sounding even better.
Intel X25-E Sata II SSD ($80) + SOtM SATA II filter ($65) + Pachanko Reference SATA cable (about $250) for OS storage
Those of you who have followed this thread from early on know that in my testing, I had found the impact of the OS drive to be greater than a separate data storage drive with respect to SQ and this is probably due to the fact that OS drives are in perpetual use, especially with Windows, while a separate data storage drive is only in use up until the digital file has been buffered into memory, at which time it becomes idle.
Those of you who have followed this thread from the beginning also know that while I was initially quite happy with my Mac Mini with its MacOS running off of an SD card, I was also looking for a way to run WIndows Server 2012R2 + Audiophile Optimizer. Unfortunately, no recent flavor of Windows Server can boot from either an SD or compact flash card and so I purchased a 2nd Mac Mini but this time, it had a PCIe SSD card installed. I was successful in installing Windows Server 2012R2 + Audiophile Optimizer and while I was pleased with the improvement over MacOS, I also noticed a fatiguing HF harshness that was not present with my Mac Mini with the SD card and no matter what I did, I couldn't get rid of it. It was one of those annoyances that you may not initially notice is there but every time I went back to my other Mac Mini with the SD card, I found myself not wanting to switch back as the SSD setup was more irritating and fatiguing. While SSDs are acoustically silent, it turns out they are electrically very noisy, especially the faster and more powerful SATA III SSDs that are being sold today as they apparently emit noise in the 6GHz range that is very difficult to mitigate even when powered by something like an LPS-1. There are many who have come up to me and have suggested that they aren't hearing the harshness that I'm hearing with their SSDs but those that have had a chance to compare my 2 Mac Minis have easily picked up how much calmer and less fatiguing my Mac Mini with SD card sounds.
Based on this experience, I felt challenged to find a way to install Windows Server onto a compact flash card and I eventually succeeded but ultimately had to give it up since compact flash boot drives are only possible with motherboards that incorporate a BIOS that allows for the older PATA drives and unfortunately, none of the SoC motherboards I wanted to use allow for this legacy option. I considered the option of using a spinning hard drive for my OS which I had confirmed is electrically quieter than an SSD but I didn't like the idea that I would have a spinning and vibrating drive in the same chassis as SOtM's sensitive clock board. Over on the JPlay forum but also on Paul Pang's website were reports of how SATA II SSDs sounded better than SATA III SSDs with SLC sounding better than MLC or TLC and so I went ahead and purchased a NOS (new old stock) Intel X25-E 64GB SLC SSD for about $80 from EBay. This turned out to be a very good purchase as indeed, this lower power, slower SSD drive was sounding less harsh and fatiguing then my Samsung 850 EVO SSD. The downside of older SATA II SLC SSDs is that they're not available as high capacity drives and so they can function just fine as an OS drive but they don't have enough capacity to store much of a music collection. Based on the limited capacities of SLC SATA II SSDs, I elected to go with the Intel X25-E for my OS drive and with compact flash cards for the storage of music files.
Here is where things get really interesting. When combined with the Pachanko Reference SATA cable, with my DFI motherboard, I felt I had equaled the SQ I was getting from my Mac Mini with SD card as a boot drive and so I found myself quite content with this solution. While at Munich, May realized I was using a SATA II drive for my OS (and not a SATA III drive) and so she strongly suggested that I try SOtM's SATA II filter. I knew they made both a SATA II and a SATA III filter and I asked if the SATA III filter was the improved version. I was surprised to hear May tell me that their SATA II filter is better but unfortunately, this filter does not work with SATA III drives and so they were forced to design a new SATA III filter. What the SATA II filter apparently does that the SATA III filter does not is that it filters both the SSD's power and data lines (their SATA III filter only filters the power line).
Here is a picture of their SATA II filter:
Here is a picture of their SATA III filter:
Those that know Lee knows that he is a filter specialist. In the same way that his filtering methods have transformed products like the tX-USBultra but also SOtM's dCBL-CAT7 and their new USB cable, this SATA filter does the same for SATA II drives and the impact is astonishing with respect to a lower noise floor devoid of any HF harshness but also this more open soundstage. The impact of this filter is eye opening in terms of just how much noise OS SSD drives create. This $65 filter is definitely one of the stars of the show and something that I consider a "must have." I would say its impact is considerably greater than the Pachanko SATA cable or the Intel X25-E SSD.
Operating System - Windows Server 2016 Essentials ($124) + Audiophile Optimizer ($129)
I believe the full version costs more but I was able to purchase a license for the Essentials version for only $124 from an online vendor. Even the GUI version of WIndows Server 2016 sounds better than the Core version of Windows Server 2012R2 and so for now, I am sticking with this version. Combined with the latest AO beta and its digital filters and sound signatures, it is the best sounding OS solution I have yet heard for music playback for Roon but also for ripping CDs via dBPowerAmp. The benefit of Windows above and beyond the Linux OS's I have heard is software compatibility. With Windows, I know I can (or will be able to) run just about anything now and in the future. More than that, I am finding Chord's ASIO driver for Windows to be superior to the stock Linux and Mac drivers for my DAVE. With my sMS-200ultra or microRendu, any time I chose to play DSD128 or DSD256 files, I frequently got skips and pauses. Not a problem at all with Chord's ASIO driver for Windows as now I am capable of smooth native DSD playback of all my DSD tracks. Also, Windows Server 2016 has been super stable. I haven't had to reboot yet due to a lock up or some system instability since I installed it.
sCLK-EX board ($850 with 4 clock outputs activated)
Along with Chord's new Blu Mk 2, this device probably gets my vote for most revolutionary product of the year. This clock board has opened up new vistas I never knew were possible just a year ago and upon replacing clock after clock, I am just amazed at how my system has become transformed. Even without connecting to an external master clock, the impact of replacing 8 clocks in my setup using this board has just been astonishing. It is this clock board that makes my build different from any other server on the market today (except for SOtM's own sMS-1000SQ server). When people tell me their server has this or that, knowing that their system probably still contains a bunch of noisy clocks and having now experienced just how big the impact of replacing clocks can be, I know their server can still sound so much better.
With the sCLK-EX board in my server chassis, I am using 1 clock for the motherboard's 25Mhz system clock, 1 clock to concurrently replace both of the integrated LAN clocks, 1 clock to replace the stock clock in SOtM's tX-USBexp and tX-USBhubIN.
WIth the sCLK-EX board housed in my tX-USBultra, 1 clock is being used for the tX-USBultra itself, 1 clock is being used for my modified ISO-Regen, and 2 clocks are being used for my Netgear integrated internet modem/router/switch.
SOtM tX-USBexp ($350)
Considering the importance of this card in my system, I consider it's $350 price a bargain. As previously stated, according to Lee, this is his best product. The photo below compares the tX-USBexp to the tX-USBhubIN and you will see that the tX-USBexp is nearly twice as large with many more parts.
While this card has been around for awhile, I believe it has gone through improvements with time. While the newer tX-USBhubIN is still using a USB 2.0 chipset, the tX-USBexp is using what Lee considers to be a better sounding USB 3.0 chipset. This card also incorporates his best regulator circuit and better filtration. What is interesting is that SOtM is an OEM supplier for other music server builders and I noticed that both Antipodes and Baetis have chosen to use SOtM's tX-USBhubIN for their servers. When I asked Lee why they would do this, all he could say is that "they didn't know any better. I designed both of these cards and I know which one is better." With SOtM's own sMS-1000SQ, this is the card they use. What is interesting is both the tX-USBexp and the tX-USBhubIN cost the same.
SOtM tX-USBhubIN ($350)
Based on what Lee has told me, if I had 2 free PCIe slots, I would have gone with 2 tX-USBexp cards, especially since they cost the same. At the same time, I am extremely happy this option exists. Even though my motherboard has only 1 free PCIe slot, my Streacom chassis has 3 available slots in the back. This card occupies one of those free slots and then connects to my motherboard via its USB 2.0 header. In the same way that SOtM's SATA II filter significantly improves the SQ of my OS drive, this card does the same for my data storage drives. Compared against the motherboard's stock USB ports, there is greater immediacy but also a more open soundstage. Even with the clocks replaced on my router, using the identical track played from my compact flash drive connected to the USB port on the back of my router (you could consider this a NAS) vs the same compact flash card connected to the tX-USBhubIN, playback through the tX-USBhubIN sounds noticeably better.
Here's another benefit of having both the tX-USBexp and the tX-USBhubIN. When ripping a CD, I can now connect my external USB CD-ROM drive to my tX-USBultra USB port (which is directly connected to my tX-USBexp) and the files rip directly onto my compact flash drive connected to my tX-USBhubIN. Using dBpowerAmp on Windows Server 2016 + AO, these are the now very best CD rips I have ever made.
Lexar HR-1 Compact Flash Hub (approx $90 + $23 for each additional CF reader)
This device can be powered with a 5V iFi PSU (no difference heard compared against an LPS-1). It has the capacity for 4 readers (CF, SD, microSD, USB, etc) and works very well for its intended use. Other devices would probably sound just as good when connected to the tX-USBhubIN but I already had this one on hand. Lexar has discontinued their 512GB Compact Flash cards although I was able to buy several of them from B&H Photo for $199 each during their closeout sale.
I purchased mine used from another CA member. I didn't find it as resolving as the tX-USBultra (my main complaint with it is that it tends to flatten the sound) but it added a certain smoothness and a weight to my piano tracks that I found very much to my liking. In this sense, I am using the ISO-Regen as more of a tuning aid but I definitely like what it adds when placed between my tX-USBexp and my tX-USBultra. Replacing its clock has definitely improved it.
tX-USBultra with 12V option and 75ohm Master Clock connectors ($1,200)
If I didn't need the extra clock taps in the tX-USBultra for my router, I could have easily skipped the tX-USBultra (and ISO Regen) because my single box server paired with the REF10 is already so good, however, adding a reclocked ISO-Regen and tX-USBultra definitely adds further improvement and for those looking to go to infinity and beyond, I believe this gets you there. Having opened up my tX-USBultra, I can see why Lee couldn't fit the tX-USBexp into the chassis, it's just too big. Most of this chassis is occupied by the sCLK-EX.
Netgear C3000 cable modem / router / switch ($90)
This particular Netgear device that integrates a cable modem, router and switch into one small chassis was chosen for its low power characteristics. As I don't use this device for its wi-fi capabilities, not only did I save a fair amount of money not having to buy an expensive NightHawk (for up to $400) but I found this low power device (12V/1.5A) to sound a bit better than a $400 Nighthawk and so I sent it along to SOtM for modification. One thing that is interesting but not surprising, when I connect my Mac Mini to this device, even with the assistance of an SOtM dCBL-CAT cable and iSO-CAT6 LAN isolator, the improvement heard is there but it's small. The improvement is much larger when connected to my server that has it's LAN clock replaced. Undoubtedly, it has improved Tidal streaming and considering how little I paid for it, this was a very worthwhile thing to do. I am currently powering this router with my sPS-500 and it does a very good job.
Mutec REF10 + 2 Habst Cables (approx $4k)
The REF10 is the icing on the cake and it is what elevates my setup to simply otherworldly status. Even with the REF10 turned off, this new setup is already the very best I have heard in my system at home sound but with the REF10 turned on, the first thing that becomes immediately apparent is this sense of buttery smoothness that I have never before experienced (especially when combined with Blu Mk 2), even when compared against the best turntables I have heard. The benefits provided by each component with just the sCLK-EX in place are enhanced to a higher level with the REF10 activated but it is this smoothness (ie complete lack of glare) that is so mesmerizing.
It's quite possible that SOtM's new external master clock will be better, especially since it can be powered by something like an SR7 but if you compare one against the other, I suspect it will be a matter of splitting hairs meaning they both will sound very good. Where the real benefits become apparent are when you replace multiple clocks and especially when you replace every clock in your chain.
USB cables (SOtM USB cable with filter block - $1,000, Phasure Lush USB cable - approx $275, Uptone Audio USPCB adapter - $35)
Like many of you, I purchased a Lush USB cable from Phasure. Mine is 0.75m long and I paid 237 Euros. Unlike many of you, I don't care for this cable as much. It does some nice things with respect to tone and warmth but it flattens sound. The perceived loss of spatial resolution when compared against SOtM's new USB cable with filter block is quite stark and even greater than ISO Regen vs tX-USBultra. With my setup already so very smooth and tonally rich, I'm not sure the small benefits it imparts are worth the very significant downside although I haven't completely made up my mind to sell it. At the present time, I am using it to connect the ISO Regen to the tX-USBexp but I am thinking my Clarity Cables Natural is sounding better at this position. I am then using the USPCB cable came with the ISO Regen to connect it to my tX-USBultra. At this time, I am finding USPCB cable preferable to the Lush. I am using my very best USB cable, SOtM's latest USB cable with filter block to connect the tX-USBultra to the USB input on my Chord Blu Mk 2.
Regarding SOtM's new USB cable with filter block, I believe this cable will be officially showcased for the first time at RMAF in a few days. It is now the best USB cable I have heard although I suspect there are some who will still prefer the tonality of the Lush. It uses the same filter block as the dCBL-CAT7, however, I find its impact to be greater and I believe I know why. With the dCBL-CAT7, I was using it before the sMS-200ultra or before the reclocking switch. With SOtM's new USB cable, I can use this cable much closer to my DAC where it is imparting a greater effect. If I swap this cable with the USPCB adapter or worse yet, the Lush (meaning I place it back further in the chain), it's impact is significantly diminished. Placed right before the Blu Mk2, the soundstage opens up beautifully. I went with UPOCC silver as a conductor and the improved detail resolution over the Lush is quite evident.
Paul Hynes SR7
I would be remiss not to mention the SR7 as I consider it the foundation of my digital system behind only my DAVE and my Blu Mk 2. If I could get Paul to build all my power supplies, I would be a very happy man as this guy has a magic touch. Yes, there's a long wait for one and even now, I'm waiting patiently for him to build me another SR7 but I'll happily wait as long as I need to.
Other tweaks - Synergistic Research Tranquility Base XL UEF ($3,250)
I found one used for $1,300 and so I took a chance on it. It was used at an audio show and it had some obvious cosmetic wear but was in perfect operational condition. At the very least, it would serve as a means of mechanical isolation in my equipment cabinet and perched on top of a trio of Synergistic Research's MIG 2.0 footers, it did a very nice job. When you turned this device on, however, I found it to be transformational in its impact. I didn't think it was possible but detail clarity improved further and the improvement was not subtle. This impact was remarkable enough that I felt compelled to squeeze every electronic component I had onto this 20" x 23" base including the REF10, server, tX-USBultra, Chord DAVE and Blu Mk2. They just barely fit and while the whole arrangement isn't as aesthetically tasteful as what I had before, I cannot deny what this thing does and the resultant improvement in detail clarity. This thing supposedly addresses airborne EMI but also EMI in your electric components. Just when you thought you had sufficiently addressed EMI, this device tells you there's still so much more of it. I was sure this was voodoo when I first heard about it but having heard a brief demo at an audio show, I felt compelled to give it a shot at home. At $3,250, I would question its value but at the price I paid, I consider it a must have.
This will likely be my last post of this depth and detail. I hope to see some of you at RMAF in a few days.
A novel way to massively improve the SQ of computer audio streaming8 hours ago, Superdad said:
Funny, if you go to my 2013 report on drive interfaces--and do a search in it on PATA--you will see me remarking how the old IDE>Firewire400 drive interface I had was the best sounding of all (well third, behind SD card and the gold standard of RAMdisk).
That was a landmark article, Alex. I benefited greatly from it.
This is what I took home from it:
With respect to SQ, you concluded that USB 2.0 > USB 3.0 and Firewire 400 > 800. Based on this, I had concluded long ago that higher speeds = higher electrical noise.
You also concluded that SD card sounded best. The Mac Mini is a unique platform in that it places an SD card reader directly on a PCIe x1 bus and so it achieves what audiophiles desire and that is the combination of an electrically quiet storage medium on a low latency bus. Thunderbolt also sits on a PCIe bus (as does FireWire) and so for those who have no interest in tinkering, the Mac Mini is really an ideal setup. If only there was a way to really strip down the OS.
For those interested, on page 16 of this technician's manual is the block diagram for the 2010 Mac Mini.
A novel way to massively improve the SQ of computer audio streamingOn 3/17/2017 at 7:40 PM, greenleo said:
Given Imitche's findings, I guess Romaz's approach of using a Mobo that supports IDE mode is very appealing. I remember Pang Pang said IDE > SATA2 > SATA3 SQ-wise because of the electrical noise. Also, Pang insist that CF sounds best again because of the lowest electrical noise.
Are you now already using the 512GB CF? Any findings?
This is correct. Paul has been very adamant that IDE > SATA2 > SATA3 with regards to SQ based on electrical noise. Looking at current draw, from what I am finding, CF draws less than SD which draws less than SSD or HD but not always.
For example, the Apacer SLC CF cards that so many people seem to like (including Phil Hobi of AO and Marcin of JPlay) has a max draw of 310mA (or about 1w at 3.3V). SDXC cards can draw twice as much. My Samsung 850EVO SSD can draw up to 1.1A.
Some 2.5 inch hard drives that spin at 5400RPM can draw as little as 450mA at 5V and could be a decent option for storage (especially when externally powered by something like an LPS-1). A 3.5 inch hard drive at 12V could be a decent option also if you decide to power it from the same supply that powers your motherboard.
Here is an interesting finding. I have found that the older Intel X25-E SLC SSDs draw very little current. It is SATA2 which is ideal. They make a 64GB model that draws about 1.1w during typical server level of I/O activity (about 0.22mA). I have found a used one for $90 and so I have purchased it for comparison.
As for 512GB of CF storage, yes, this is too small a volume for most of us. I have found a card that can accept 2 of them (for up to 1TB of storage). I have also found a PCIE to SATA adapter (with a replaceable clock) that comes with 4 SATA ports and so not counting one port that would be used for the OS drive, this solution provides 3 free ports for storage drives (or the option of 3TB of CF storage). I have not yet purchased this 512GB CF card but with the 32GB CF card that I have, I am finding SQ to be a bit better when compared against a 64GB SDXC.
I think what is potentially as important is the bus that you connect to. As opposed to USB vs SATA vs Thunderbolt vs Firewire vs PCIe, it may be more important to know whether these buses go through the PCH (Platform Controller Hub) which is a hub that most buses must pass through before reaching the CPU. From what I can tell, only PCIe 3.0 X8 or x16 slots have direct access to the CPU without having to access the PCH and as most know, such slots were designed for use with powerful GPUs in SLI mode. This will, however, depend on the motherboard and so it is important that you review the block diagram of your motherboard. PCIe 3.0 operates at either 3.3v or 12v depending on the device but it appears that that's it.
With buses that have to go through PCH (this means USB, SATA, SATA Express, most Thunderbolt, Firewire, PCIe 1.0 or 2.0 and some 3.0), you not only have to go through the noisy switching regulator that supplies that bus (3.3V, 5V or 12V) but also an even noisier 1.05V regulator that supplies the PCH.
With the SATA to PCIe adapter and the Intel PCIe LAN card I plan to use, because I will be connecting them to the motherboard's PCIe 3.0 x8 slots that completely bypass PCH, I am hoping that there will be a noticeable improvement in SQ from the standpoint of less latency and less noise. We'll see.
A novel way to massively improve the SQ of computer audio streaming
So it would seem that moving up to this larger microATX ASRock motherboard will bring with it both good things and bad but hopefully, the positives will outweigh the negatives.
Two PCIE 3.0 slots that allow the most direct access to the CPU that is possible. These slots have the same low latency access to the CPU as RAM. All the other buses must go through the Platform Controller Hub to access the CPU. One slot will be used for my Intel dual LAN card and the other slot will be used for my SATA to PCIE card where my OS drive and one of my data drives will connect.
I can power the CPU directly with a 12V lead from my SR7. This should be a significant advantage over my Mac Mini and hopefully will help negate some potentially serious disadvantages.
My mATX motherboard will require a high current 5V, 3.3V and 5V standby rail in addition to possibly another legacy rail and so Paul Hynes has told me that I will not be able to use my SR7 to directly power these rails without a DC-ATX converter. This means I have placed an order for the HDPlex 300W-HiFi-DC-ATX converter based on the recommendation of many. From my search, I have not found another DC-ATX converter that is convincingly better. With this HDPlex 300W-HiFi-DC-ATX converter in place, I can then use my SR7 to feed it. I have had another SR7 on order since January and Paul has told me it should be finally ready to ship in the next week or two. Upon hearing my dilemma, Paul was kind enough to agree to convert my 12V rail to a 19V rail so that I can feed the HDPlex DC-ATX converter the voltage it requires. He has also agreed to fabricate for me a high performance 24-pin to 24-pin ATX cable so that I can connect the HDPlex DC-ATX converter to my motherboard. I really can't say enough great things about Paul. For those with custom LPSU needs, I can't recommend him more highly.
Anyway, now that I'm on this path, I did some digging around and here are some comparative values. They don't guarantee how good something will sound but they do provide perspective.
Since only noise floor data is provided by most manufacturers, I've gone ahead and listed what I have found. Noise is listed in mV and the lower the value, the better:
Generic switching PSU ~ 60mV
EVGA SuperNova 1.6kW Platinum ATX PSU (100% load) ~ 12.4 - 22.7mV
EVGA SuperNova 1.6kW Platinum ATX PSU (20% load) ~ 4.0 - 13.4mV
Corsair AX1200i 1.2kW Platinum ATX PSU (100% load) ~ 7.3 - 13.7mV
Corsair AX1200i 1.2kW Platinum ATX PSU (20% load) ~ 5.6 - 11.4mV
HDPlex 400W ATX LPSU (100% load) ~ 3-5mV
HDPlex 100W LPSU with LT1083 regulators (100% load) ~ 2-3mV
Paul Hynes SR7 (100% load from 10Hz to 100KHz) <4uV = <0.004mV
Compared against either HDPlex LPSUs, the SR7 is in the order of being nearly 1000x quieter and nearly 10,000x quieter than a generic switching PSU.
Unfortunately, because I will have to use the HDPlex 300W-HiFi-DC-ATX converter which uses regulators with measured ripple values of 10mV (more than 2000x noisier than the SR7), the ultra low noise floor of the SR7 will be buried by the much higher noise floor of the HDPlex DC-ATX converter. From a noise standpoint, almost any decent switching PSU may possibly sound just as good because this converter will be the limiting factor.
This then raises the question of just how noisy the numerous switching regulators buried within the motherboard are. Quite possibly, they may be even noisier than 10mV and so from a noise standpoint, especially with a motherboard's 3.3V rail (which is the noisiest rail on an ATX motherboard), all of this may be moot. Quite depressing.
1. We really need someone to design an audiophile-class motherboard without noisy switching regulators.
2. In the absence of such a motherboard, because noise is potentially additive, it probably still makes sense to use a low noise PSU but its benefits will be at least partially negated by the noise created by the motherboard (or the DC-ATX converter).
3. It would seem that ATX power supplies designed for high power output (specifically high current output) will generally also have a higher noise floor but in this regard, the HDPlex ATX LPSU stands out as being exceptional.
4. Noise aside, low impedance still matters. To my ears, this may be a more important quality because with the aid of grounding boxes, noise filters and galvanic isolating devices, there are ways to mitigate noise but no way to undo the suffocating impact of a high impedance PSU. It's quite possible the EVGA excels here but there's no way to know except through comparative listening since no measurements are available.
UpTone Audio EtherREGEN Listening Impressions23 hours ago, Superdad said:
It is, but there is the matter what screw of the JS-2 enclosure you would use as the chassis is anodized and while the screws are steel, the ones holding the cover are short and a special countersink. Putting a wire under them could be tricky and likely to scratch the lovely case finish. And if you pick one of the other underside stainless steel hex-head screws, you need to be mindful as some are short PCB screws, and the larger ones are for the transformer and you could have trouble retightening them from only one side.
I'd really rather people not mess with JS-2 screws. You must have many other good grounds in your system.
Well I was naughty. Just for experimental purposes. And before you ask, yes, there was continuity. Outcome: shunting is a must when powering both the ER and something connected to the A-side - in my case a Mac mini - from a JS-2. At least when you have copper input from an ISP router or any other component in your home network to the A-side as well.
What puzzles me is this: while shunting closed the gap a bit, I still only get stellar performance when I keep the Emo systems network isolator between the Mac mini and the ER. I was so hoping this would not be the case. That - if grounding was dealt with correctly - network related changes on the A-side would not affect what comes out on the B-side...
PFM Flea from Alex
The ultimate Power Supply Units for music servers (and other devices for cleaner power source)1 hour ago, Nenon said:
What is a "definite No,No"?
Don't feed power to both an Oscillator and an SSD from the same supply rail without further regulation (preferably) or further filtering of the power to the oscillator.
I use the attached, which is fairly similar to the P.F.M. Flea to separately power oscillators from an already low noise voltage regulator.
2019-06 Isolation Pt2
Sonore opticalRendu6 hours ago, MagnusH said:
Ok, that might well be the case. But once you add an optical transmission like fiber, that all goes away, so any clocks before the fiber won't matter as long as the data was delivered correctly. In fact, nothing before the fiber should matter at all, provided the data was delivered correctly (the FMC after the fiber will matter though).
All the optical does is block leakage, it doesn't get rid of clocking issues at all (it can actually make them worse). The fact that it is optical does not automatically apply some universal quantum time scheme that mystically aligns edges perfectly, If you send in a pulse, then another that is 50ns apart, then another at 51ns, then another at 49, that difference gets preserved at the receiver, the optical does not magically force all of them to be exactly 50ns.
The raw data coming out of the optical receiver goes into a chip that rebuilds the Ethernet signal using its own local clock, that is done with flip flops inside the chip, these flop flops behave just like any other flip flops, again no magic here. I was trying to avoid re-iterating what I have said before on this, but it looks like I'm going to have to do it anyway.
So how come this reclocking with a new clock is not perfect? As edges from the input stream go into a circuit each and every one of those edges creates a current pulse on the power and ground network inside the chip and on the board. The timing of that pulse is exactly related to the timing of the input data. The timing of the input data is directly related to the jitter on the clock producing the stream. This noise on the PG network changes the threshold voltage of anything receiving data inside the chip, especially the local clock going into the chip. This means the phase noise spectrum of the data coming in gets overlayed on top of the phase noise spectrum of the local clock. It's attenuated from what it is in the source box, but it is definitely still there.
THAT is how phase noise gets from one device to the next, EVEN over optical connections.
If you look at this in a system containing all uniformly bad clocks, you don't particularly see this, since they are all bad to begin with. BUT when you go from a bad to a very good clock you can definitely see this contamination of the really good clock by the overlaying of the bad clock. This is really hard to directly measure because most of the effect is happening inside the flop flop chip itself. You CAN see the effect on the data coming out of the flip flop.
This process happens all the way down the chain, Ethernet to USB, USB into DAC box, and inside the DAC chips themselves, finally winding up on the analog out.
Wherever reclocking is happening, how strong this overlay is depends primarily on the impedance of the power and ground network, both on boards and inside chips. A lower impedance PG network produces lower clock overlay, higher PG impedance give stronger overlay.
This is something that is difficult to find out about a particular chip, the impedance of the PG network is NEVER listed in the data sheets! I have somewhat of an advantage here having spent 33 years in the semiconductor industry, spending a lot of time designing PG networks in chips, I have some insight into which chips look like good candidates for low impedance PG networks.
On a side note, because Ethernet and USB are packet systems the receiving circuit CAN use a completely separate clock, the frequency just has to be close enough to handle the small number of bits in the packet. If it is a little to slow or too fast the difference is made up in the dead time between packets.
To reiterate none of this has ANYTHING to do with accurately reading bits, this is assumed. It IS all about high jitter on network clocks working its way down through reclockings to the DAC chips and hence to audio outs. All the work done on DACs in recent years has cleaned up the signals so dramatically that these effects are getting to be audible in many systems.
2019-06 Isolation Pt1
The understanding of "isolation" in digital audio has been my passion for at least 10 years. There is a LOT of misunderstanding on the subject floating around in audio circles. Here is a quick summary of my current understanding and how the current products fit in with this.
There seems to be TWO independent mechanisms involved: leakage current and clock phase noise. Various amounts of these two exist in any system. Different "isolation" technologies out there address one or the other, but very rarely both at the same time. Some technologies that attenuate one actually increase the other. Thus the massively confusing information out there.
Leakage current is a property of power supplies. It is the leakage of AC mains frequency (50/60 Hz) into the DC output. It is usually common mode (ie exists on BOTH the + and - wires at the same time, this makes it a bit difficult to see. There seems to be two different types, one that comes from linear supplies and is fairly easy to block, and an additional type that comes from SMPS and is MUCH harder to block. An SMPS contains BOTH types. They are BOTH line frequency.
Unfortunately in our modern times where essentially all computer equipment is powered by SMPS we have to deal with this situation of both leakage types coming down cables from our computer equipment. There are many devices on the market (I have designed some of them) for both USB and Ethernet, most can deal with the type from linear supplies but only a few can deal with the type from SMPS.
Optical connections (when the power supplies are completely isolated from each other) CAN completely block all forms of leakage, it is extremely effective. Optical takes care of leakage, but does not deal with the second mechanism.
Clock phase noise
Phase noise is a frequency measurement of "jitter", yes that term that is so completely mis-understood in audio circles that I'm not going to use it. Phase noise is a way to look at the frequency spectrum of jitter, the reason to use it is that there seems to be fairly decent correlation to sound quality. Note this has nothing to do with "pico seconds" or "femto seconds". Forget those terms, they do not directly have meaning in audio, what matters is the phase noise. Ynfortunately phase noise is shown on a graph, not a single number, so it is much harder to directly compare units. This subject is HUGE and I'm not going to go into any more detail here.
Different oscillators (the infamous "clocks" that get talked about) can have radically different phase noise. The level of phase noise that is very good for digital audio is very difficult to achieve and costs money. The corollary is that the cheap clocks used in most computer equipment (including network equipment) produce phase noise that is very bad for digital audio.
The important thing to understand is that ALL digital signals carry the "fingerprint" of the clock used to produce them. When a signal coming from a box with cheap clocks comes into a box (via Ethernet or USB etc) with a much better clock, the higher level of phase noise carried on the data signal can contaminate the phase noise of the "good" clock in the second box. Exactly how this happens is complicated, I've written about this in detail if you want to look it up and see what is going on.
The contamination is not complete, every time the signal gets "reclocked" by a much better clock the resulting signal carries an attenuated version of the first clock layered on top of the fingerprint of the second clock. The word "reclocked" just means the signal is regenerated by a circuit fed a different clock. It may be a better or a worse clock, reclocking doesn't always make things better!
As an example if you start with an Ethernet signal coming out of a cheap switch, the clock fingerprint is going to be pretty bad. If this goes into a circuit with a VERY good clock, the signal coming out contains a reduced fingerprint from the first clock layered on top of the good clock. If you feed THIS signal into another circuit with a very good clock, the fingerprint from the original clock gets reduced even further. But if you feed this signal into a box with a bad clock, you are back to a signal with a bad fingerprint.
The summary is that stringing together devices with GOOD clocking can dramatically attenuate the results of an upstream bad clock.
The latest devices form Sonore take on BOTH of these mechanisms that effect sound: optical for blocking leakage and multiple reclocking with very good clocks. The optical part should be obvious. A side benefit of the optical circuit is that is completely regenerates the signal with a VERY low phase noise clock, this is a one step reclocking. It attenuates effects from upstream circuits but does not completely get rid of them. This is where the opticalModule comes into play, if you put an opticalModule in the path to the opticalRendu you are adding another reclocking with VERY good clocking. The result is a very large attenuation of upstream effects. It's not completely zero, but it is close.
The fact that the opticalRendu is a one stage reclocking (which leaves some effects from upstage circuits) is why changing switches etc can still make a difference. Adding an OpticalModule between the switch and opticalRendu reduces that down to vanishingly small differences.
So an optical module by itself adds both leakage elimination and significant clock effects attenuation. TWO optical modules in series give you the two level reclocking .
An opticalRendu still has some significant advantages over say an ultraRendu fed by a single opticalModule, the circuitry inside the opticalRendu has been improved significantly over the ultraRendu. (it uses new parts that did not exist when the ultraRendu was designed). In addition the opticalRendu has the reclocking taking place a couple millimeters away from the processor which cuts out the effects of a couple connectors, transformers and cable. The result is the opticalRendu has some significant advantages.
An opticalModule feeding an ultraRendu does significantly improve it, but not as much as an opticalRendu. So you can start with an opticalModule, then when you can afford it add an opticalRendu, also fed by the opticalModule and get a BIG improvement.
I hope this gives a little clarity to the situation.
Raspberry Pi as a music server?
Raspberry Pi as a music server?2 hours ago, Giuanniello said:
Ok, I downloaded and install the LMS on my Mini which, by the way, runs a Core Duo 2 CPU which I upgraded from a Core Duo so it won't install any other OS higher than Snow Leopard but I am fine with it if it does what I wish it to do which is simply to act as a media streamer for music, movie wise I have them on the NAS and use Infuse to stream over the Apple TV4.
Now, once I have the server up and running, how can I control it off an iPhone to make things easy, I guess there is an iOS app but I can't manage to find it.
Congratulations! There are apps, but first try your iPhone web browser. Enter the IP address of your Mac and add port 9000 — for example, 126.96.36.199:9000 — and you'll open the original classic web interface. I suggest you install the Material Skin plug-in for an updated and mobile-friendly interface, after which you'd append /material/ to the address, for example 188.8.131.52:9000/material/
On the iPhone or on the Mac itself, click "Settings" at the bottom of the LMS page, and you'll find tabs for scanning your library, adding plug-ins, and more. To add the Material Skin plug-in, scroll down here for installation instructions: https://github.com/CDrummond/lms-material Go here for discussion and support: https://forums.slimdevices.com/showthread.php?109624-Announce-Material-Skin Other useful plug-ins include Music and Artist Information, What Was that Tune, Radio Paradise, and streaming services.
All of the above is free. On iOS the most popular app is not free but it is very good: iPeng. I was using an open-source app on Android, Squeezer, but have switched to the web-based Material Skin as it offers more features. There's lots of information on https://forums.slimdevices.com/ and there are many users here on CA as well.
Isolation & Reclocking
USB jitter - why does it matter?1 hour ago, Sound Hound said:
I'm putting together an 8 channel system with DDX amps for experimenting with ambisonics and multiway active setups.
since my background is computers and I've only recently forayed into audio, I'm a bit mystified by jitter and precision clocking.
I get that galvanic isolation and a separate, clean power source are important to the audio bits beyond the computer.
but I don't understand why the USB connection between such needs to have more than a reliable/accurate transfer of data.
does the jitter transmit stray signals into the latter stages? if not, then the only place high precision clocks are warranted is in driving the DAC or DDX stage.
surely any USB implementation is sufficient with the data adequately buffered.
reclockers?! iPurifiers?! pah - audio voodoo!
I'm cynical but ready to be enlightened!
Hi Sound Hound,
I have been working on this for years, I'm getting close to a complete end to end measurement, but test equipment to properly measure this stuff doesn't exist, I'm having to design and build my own as I go along. I can measure pieces of the chain now and the rest hopefully coming soon. Part of the slowness was getting laid off and retiring and moving to a new state. I now have a working lab again and am working on the next piece of test equipment.
The hypothesis goes thusly:
ALL crystal oscillators exhibit frequency change with power supply voltage change. This is known and well measured. A cyclical change in voltage causes a cyclical change in frequency which shows up in phase noise plots. For example if you apply a 100Hz signal to the power supply of the oscillator you will see a 100Hz spur in the phase noise plot.
A circuit that has a digital stream running through it will will generate noise on the power and ground planes of the PCB just from the transistors turning on and off that are processing that stream. This effect is very well known and measured. Combine this with the previous paragraph and you have jitter on the incoming data stream producing varying noise on the PG planes that modulates the clock increasing its jitter.
The above has been measured.
But shouldn't ground plane isolation and reclockers fix this? At first glance you would think so, but look carefully at what is happening. What is a reclocker? A flip flop. The incoming data with a particular phase noise profile goes through transistors inside the flip flop. Those transistors switching create noise on its internal PG traces, wires in the package and traces on the board. This noise is directly related to the phase noise profile of the incoming data. This PG noise changes the thresholds of the transistors that are clocking the data out thus overlaying the phase noise profile of the local clock with that of the clock used to generate the stream that is being reclocked. This process is hard to see, so I am working on a test setup that generates a "marker" in the phase noise of the incoming clock so it becomes easy to see this phase noise overlaying process.
This process has always been there but has been masked by the phase noise of the local clock itself. Now that we are using much lower phase noise local clocks this overlying is a significantly larger percentage of the total phase noise from the local clock.
Digital isolators used in ground plane isolation schemes don't help this. Jitter on the input to the isolator still shows up on the output, with added jitter from the isolators. This combination of original phase noise and that added by the isolator is what goes into the reclocking flip flop, increasing the jitter in the local clock. Some great strides have been made in the digital isolator space, significantly decreasing the added phase noise which over all helps, but now the phase noise from the input is a larger percentage, so changes to it are more obvious.
The result is that even digital isolators and reclocking don't completely block the phase noise contribution of the incoming data stream. It can help, but it doesn't get rid of it.
For USB (and Ethernet) it gets more complicated since the data is not a continuous stream, it comes in packets, thus this PG noise comes in bursts. This makes analysis of this in real systems much more difficult since most of the time it is not there. Thus any affects to an audio stream come and go. Thus just looking at a scope is not going to show anything since any distortion caused by this only happens when the data over the bus actually comes in. To look at anything with a scope will take synchronizing to the packet arrivals. Things like FFTs get problematic as well since what you are trying to measure is not constant . It will probably take something like wavelet analysis to see what is really happening.
The next step in my ongoing saga is to actually measure these effects on a DAC output. Again I have to build my own test equipment. The primary tool is going to be an ADC with a clock with lower phase noise than the changes which occur from the above. AND it needs to be 24 bits or so resolution. You just can't go out and buy these, they don't exist. So I build it myself.
I have done the design and have the boards and parts, but haven't had time to get them assembled yet. Then there is a ton of software to make this all work. Fortunately a large part already exists, designed to work with other systems but I can re-purpose it for this.
So it's not going to be right away, but hopefully not too off in the future I should be able to get to actually testing the end to end path of clock interactions all the way to DAC output.
Crespi on SMPSs
astron lpsu experience?
The Mytek Brooklyn DACs measure well—with their internal SMPS (another Mean Well model, similar to what Mutec uses in the MC3+ we were discussing). Yet Mytek acknowledges better performance when an external LPS is used, even going so far as to link to a couple on their web page.
We build about 250 per year of our big, choke-filtered, dual-output, 5-7.2A JS-2 linear supplies, and in 2017 about 50 of those went to Brooklyn DAC owners—who were, to a one, quite thrilled with the sonic result.
I even gave you a link to one of the Mutec engineers saying that for their new product they focused even more on the power supply and ditched the SMPS.
If you do not think power supply design matters to product performance that’s your choice. But if you acknowledge that it does matter, then you need to think of it all the way to the wall. If a preamp or DAC has its own complete AC>DC PS built in that’s fine—and hopefully the designers took care not just with the LDO regs for their various circuits, but also with the AC>DC rectification side to produce quality, low-noise, low-impedance, regulated gross DC for their network of lower voltage regs—while not also infusing them with common-mode AC leakage.
Unfortunately, the all too popular use of off-the-shelf, caged SMPS modules inside some gear (even the $10K Merging NADAC), is a hindrance to allowing otherwise very fine products to achieve the best SQ they are capable of.
So then why do manufacturers use SMPS bricks—either internally or externally? Two main reasons:
1) Cost. A SMPS wall-wart, brick, or caged module costs (in quantity) between $3 to $12. That’s a LOT less than the cost to design and build in even the most basic trans>diodes>caps>regulators LPS.
2) Certifications. The (typically Chinese) SMPS units already come with certificates and marks from sometimes as many as a dozen world certifying bodies. So the audio component manufacturer does not have worry about the hassle and expense of getting their AC-attached product’s PS certified.
You keep referring to “well designed hi-fi components.” Yet that is a spectrum, and is almost as vague as “well prepared food.” Yes, the food—or the hi-fi component may be “well done,” but that does not mean that it will taste or sound the best that it could be.
Jitter: The Digital DevilIndeed. But esldude has a point. I have also noticed that John S is happy to discuss the mechanics of jitter but generally steers clear of discussing its audibility. A wise man.
I can hear affects of jitter, but I have not been able to quantify exactly what aspect of the jitter correlates to what I hear. I know this is just "anecdotal evidence" that would not convince anybody since it was not garnered from a 5 year study with 10,000 people etc.
For me as a person who has been working on DAC designs for many years I'm not willing to spend the time and money to run every possible change through a 5 year study. It's just too slow as part of the design process.
My usual approach is to come up with a hypothesis and come up with a way to test it, design a circuit that I think might make a change, build a small circuit that I hope is just testing that one hypothesis (not always true) and listen to it and do a number of measurements. Most of the time these tests make no difference, or make it worse, but occasionally they make it sound better. If it does sound better I will send a copy of the circuit to a friend or two to see what they hear. If everybody in the test set agrees I then run a ton of measurements trying to come up with some correlation to what is being heard. I will usually go through 5 or 6 trials per year. Every year to year and a half I will then make several copies of a complete new DAC using the best of the old and the few new mechanisms which seems to make things better. These go out to a number of people for extended listening tests to see how successful I was.
After 10 years of doing this I'm getting some pretty good DACs and learned a LOT along the way.
Through all this I have found that decreasing jitter frequently does make significant improvements in sound. But strangely enough not always. There have been a few cases where decreasing jitter did nothing or made it worse. In one of these cases I did determine that the mechanism used for decreasing the jitter was increasing distortion through a not originally thought of mechanism.
I still remember one of my first jitter experiments many years ago. I had given up on trying to get get S/PDIF to work really well, adaptive USB was just not good enough and there was no way to do async at the time with out a HUGE expense of time. I found that the squeeze boxes were perfect for what I wanted to experiment on. Local fixed oscillators which directly controled the rate coming out of the buffer, a perfect architecture. So I bought a couple SBs and listened to them for quite some time as is to get a base line. Then I hacked into the I2S lines, sent them out to a board with Tent clocks on it, an FPGA to convert I2S to a pair of 1704s, reclocked that data with the Tent clocks and sent the clocks back into the SB instead of it's local oscillators. It sounded way better than what was coming out of the SB analog outs.
I wasn't using any good method for getting the clocks to the SB, just a simple ribbon cable, I didn't care if it picked up all kinds of jitter on the way since it was just for syncing the SB. After listening to this for a while I decided to listen to the analog outs from the SB. I fully expected it to sound horrible since I was sure the jitter being fed to it over the ribbon cable was worse than what was originally there. When I listened I was astonished to find it sounded quite a bit better than the "factory" configuration.
Later I tested the jitter on the clock in the SB and even with the not great signal integrity coming over that cable, it was still quite a bit lower than the factory configuration, and that really did make a significant difference. It was then and there that I realized that there really was something to this jitter stuff and I should look into it more.
Signature Series Rendu SPDIF/i2s - Discussion and Experiences5 hours ago, Cooler said:
Yes, they are USB to i2s, and there is no ethernet to i2s converter on the market or i didnt hear about that device. Now if you want get out all from the modern dac (with i2s input) you need ethernet to usb device and usb to i2s device, thats not only really expensive, but also to many conversions and connections, 2 PSU, additional cables etc. Why not to create new microRendu ethernet to i2s device, small and simple, it could be a new product and with global trend it should be very popular, especially if you would not leave $1000 price range.
I understand, that you have clearer picture, what people want and how to do your business, but just check the recent dacs market and the potential to sell that new device.
ps. i would definitely buy ethernet to i2s mR (with Roon Ready and HQ NAA of course)
There are some major technical issues with this. The processor in the micr/ultraRendu was chosen because it lets the USB subsystem be powered and clocked separately from the rest of the processor. That particular processor can only do I2S up to 192, and CANNOT to DSD over the I2S wires (other than DoP). Note that the the I2S spec has absolutely nothing to do with DSD. There are a few DACs and DDCs that multiplex in the DSD signals onto the same wires used for I2S, but it is NOT I2S.
If you are willing to live with the 192 maximum and no DSD, then it COULD output I2S, but then you will degrade the performance of the USB output. If you optimize USB, you get less than great I2S, if you optimize I2S you get less than gret USB (or you can get less than great both).
There are some ways to do both USB and I2S and do both very well, but they are neither simple nor cheap. There would be long development time and the end result would be expensive. Even just a Ethernet to VERY good I2S is not easy at all. None of the processors have an I2S block that will do it, it will take something like a processor interfaced to an FPGA to do it properly, with a hole bunch of other stuff in there to keep the processor from messing up the I2S timing. Again complex and expensive.
The reason you see it in USB input DDCs is that XMOS has USB audio code that does I2s and DSD over I2S wires. That makes it easy to do. There is nothing equivalent for something that can do Ethernet and support the various different Ethernet audio Protocols.
At this point the best way to do this is Ethernet to USB, to DDC to I2S. Anything else "simpler" is going to take a long time to come to fruition and cost a lot.
SuperclocksOn 4/25/2017 at 4:11 PM, Hammer said:
Are these clocks different than a rubidium clock from say Stanford Reseach Systems? I picked one up on the cheap off eBay and had been meaning to purchase a DAC such as a Mytek which accepts clock input to play around, but have not had the time. Has anyone tried this with good result? Thanks, hammer
Rubidium clocks are usually very bad to use for audio. They have very good long term stability, but high phase noise. The long term stability has nothing to do with audio but the close in phase noise is what is important. So a rubidium is exactly the wrong oscillator to use.
Another problem is that the rubidium is probably NOT going to be outputting a frequency that can be used directly by audio circuitry, so some for of frequency synthesizer is going to have to be used, and these ALWAYS increase the phase noise.
A rubidium is great for an actual clock (you can read the time) that you want to be accurate to the microsecond over years of run time, but not so good for audio.
ISO REGEN launch thread! (product web page up; photos, etc.)16 hours ago, rickca said:
Alex, you said that when you and John evaluated the Crystek CCHD-575, you quickly decided it was well worth using in the ISO REGEN.
Have you experimented with more expensive clocks? I'm trying to understand whether there's a point of diminishing returns even if you had no cost constraints.
No, we have not tried better clocks than the 575, the next step up in lower phase noise needs an OCXO. Note that inexpensive OCXOs do NOT have lower phase noise than the 575, you have to go to very expensive OCXOs to better it. And it is not just the cost of the OCXO itself. An OCXO takes a lot of power to run the oven and other circuitry, this will also add cost to the system. With the right OCXO we can probably still use the LPS-1 to power the board, but then you would have a very hard time powering anything else from the same one.
Both Alex and I are very much interested in producing items that are very high performance but still low enough in cost that a fairly large number of people can afford them. This is one of those cases where my gut feeling here is that spending the money on a better clock will give better results in a DAC rather than in an upstream device.
CLOCKS, what should we look for in next generation
I've been thinking about writing a primer on crystal oscillators and digital audio and this looks like the perfect place to put it. I promise I will leave out all the complex math that most articles are filled with. I'm NOT going to go into how it all works, since most people don't care, just what makes them different and how that matters for audio.
A crystal oscillator is a combination of a special piece of quartz crystal and an electronic circuit, the combination produces periodic signal at a specific frequency, several things can change this frequency:
Thickness of the quartz piece, this is the primary determining factor in the frequency
Temperature of the crystal, this is called the temperature coefficient (TEMPCO for short), it is the change in frequency for a small change in temperature. It is not constant but changes with temperature, this is the TEMPCO curve. All TEMPCO curves have a temperature where the TEMPCO is zero, this is called the inflection point. If you run the crystal at this temperature, small changes in temperature produce no change in frequency, THIS point is where you want to run a crystal oscillator. If the temperature is far away from this point a small change in temperature makes a big change in frequency, you do not want to be here.
Capacitance across the crystal, all crystal oscillators need some capacitance across the crystal to work, changing that capacitance changes the frequency.
Power flowing through the crystal. The oscillator circuit works by running power (in the form of an AC signal) through the crystal, changing the power changes the frequency.
TEMPCO is THE most important characteristic besides the thickness, so a lot of crystal oscillator design has to do with this.
Now on to "cut", this is how a slice of crystal is cut out of a block of quartz. This is all very complicated so I won't go into the details, just to say there are many ways to do this and the exact cut determines the properties of the oscillator.
The most common cut (BY FAR) is called the AT cut. Almost all the oscillators in your electronics devices use the AT cut. The primary reason for this is that the inflection point of its TEMPCO curve is at 25-35C, right around "normal" room temperature, especially in a box where the electronics warm it up slightly. With this cut you usually do not need to apply any temperature stabilization since it is at a point where a change in temperature makes a very small change in frequency.
The other cut we need to talk about is the SC cut, this is used in OCXOs, I'll talk about that later. This cut has much higher Q than the AT cut, which means much lower phase noise, BUT in order to get that the inflection point of the TEMPCO curve is at 95C. THIS is why an oven is needed, not so much to stabalize the temperature but to get the crystal to the inflection point where a change in temperature makes an extremely small change in frequency. The slope of the TEMPCO cut around the inflection point is much shallower than the AT cut, so a given change in temperature makes a much smaller change in frequency, IF it is at 95C, outside of that and it is worse than an AT. So you ONLY want to use an SC cut in an oven.
So what aspect of this is really important for digital audio? Most oscillator spec sheets spend a lot of time talking about their long term stability. It turns out crystals will change frequency over time (called aging). Some applications need this, digital audio does not. A 1 part per million change in frequency over years time is completely irrelevant. Another spec that is important for some application is the TEMPCO, how much the frequency is going to change as the heater turns on and off. Again, irrelevant to digital audio. What DOES matter is phase noise. I'm not going to go into any detail on this but that is what matters. It is not a single number but a graph, you have to see the graph to really get an idea of what it is.
The manufacturers are starting to realize this and are now making some fairly inexpensive AT cut crystals with extremely low phase noise. They don't have great aging or great TEMPCO but they DO have great phase noise.
There are three common crystal oscillator configurations you will come across in digital audio:
XO - basic simple crystal oscillator, always uses an AT cut crystal, susceptible to the ambient temperature (remember that 25-35C) changes a fair amount over the years, has a huge range of phase noise from one model to the next. Anywhere from $0.35 to $25.
TCXO- Temperature Compensated Crystal Oscillator. Standard AT crystal with a temperature sensor that feeds a voltage variable capacitor across the crystal. In order to have a large enough "pull range" to handle large changes in temperature the crystal is modified so the frequency changes a lot with a given capacitance change. Unfortunately this radically increases the phase noise of the crystal. Thus TCXOs are about the worst clock you can use for digital audio. You get much better temperature stability, which you don't care about, in exchange for much worse phase noise which you DO care about. A very bad trade off. So if you see a digital audio device with a TCXO, stay away.
OCXO Oven Compensated Crystal Oscillator. The oscillator sits in an oven that brings its temperature to 95C. Most writing you find on the net will say this is to stabilize the temperature, but the real reason is to bring an SC cut crystal up to 95C where its built in TEMPCO is zero. This gives extremely low frequency change with temperature, but the SC also has MUCH lower aging than the AT AND much lower phase noise than the AT. Thus the OCXO is great for both systems that require extremely low drift but also systems that require extremely low phase noise.
The problem is that OCXOs are not cheap, $100 and up (WAY UP). The cheapest OCXOs have about the same phase noise as the best AT cut XOs, for about 4 times the price. So for digital audio at least a low end OCXO is not particularly useful. You have to get in the $300 range to get OCXOs with significantly lower phase noise. As you go up from there you can get WAY better phase noise, but you really have to pay for it. So when looking at OCXO specs, all you need look at is the phase noise, all the stuff in PPB etc is irreverent. Don't waste money on getting the best in those specs. If a manufacturer just shows the PPB numbers and doesn't give phase noise, stay away.
Another thing that has been talked about is "atomic clocks". The "inexpensive" ones (less than $10k) are rubidium. These have EXTREMELY low long term drift, but very bad phase noise. There is NO reason to get one of these for digital audio. Sometimes a rubidium oscillator is paired with an OCXO, the rubidium "disciplines" the OCXO, this gives the best of both worlds, but if you spent the same amount of money on just the OCXO you could get much lower phase noise which is what matters.
In the next installment I'll go into frequency synthesizers and how recent changes are changing the landscape of clocks for digital audio.
Isolation & Reclocking
Are all Asynchronous USB chips/implementations created equal??Thanks for your explanations.
Would an opto-isolator between USB receiver chip and a separate, cleanly powered area with the master clock plus DAC chip, not prevent any noise/jitter from passing on from the computer? If the opto-isolator is powered and grounded through the USB connection will it matter? This certainly would make it easier and cheaper to connect the computer to the DAC.
As Alex mentioned I hate opto-isolators, there are much better isolators out there, I use the GMRs exclusively. Unfortunately they are pretty expensive.
Even using isolators does not completely block jitter. I'll try and get this across without pictures. A signal goes into the USB side of the isolator, current flows from the driver, through the input side of the isolator (whatever it is) and back through the plane to the driver chip. That signal passes through the isolator some how (light, magnetic field, radio waves, whatever) (yep one of the isolator technologies actually sends radio waves between the sides) and causes the receiver side to do something, which changes the signal on it's output. That output then drives the DAC chip or reclocking flop, which then sends the current back to the isolator output on the groundplane. The current ALWAYS goes in loops, thus the signal going to the DAC chip creates noise on the DAC side groundplane.
Thus any jitter on the signal crossing the isolation barrier is added to the inherent jitter of the isolator and that shows up as noise on the DAC side ground plane, even with the isolator! What the isolator does is prevent OTHER GPN such as being produced by the USB receiver itself from getting into the DAC groundplane. It's definitely worth it, but you still have to deal with the jitter on the I2S signals themselves which cross the barrier.
Because the I2S signals are fairly jittery after the isolators, you usually should reclock them before sending them to the DAC chip. Why do you need to do this? Isn't the jitter on the clock the only signal that matters? Because GPN also happens INSIDE the DAC chip. Jittery input signals generate noise on the ground traces in the DAC chip, which change how the clock signal is received. You can have an extremely low jitter clock going into the DAC chip, but if the I2S signals are very jittery, the GPN inside the chip will cause that ultra low jitter clock to look MUCH worse.
So you still have to look at the jitter on the I2S signals, even with a perfect clock.
There are ways to cut down on these issues by careful board layout, but you have to include these as part of the overall design from day 1 to make sure they will be effective. But even with this, some influence still gets through.
Some info on I2S and especially I2S between boxes.
I2S is very simple, no packets or overhead, one wire with serial data, alternating between left and right channels on the same wire, an LRCLK signal that says when it is right data and when it is left data and a bit clock, that specifies when to read the serial data line. In addition sometimes a "master clock" is sent along as well, this is an integer multiple of the bit clock.
The timing of the signal is on these wires. If the DAC chip is connected directly to the wires, there is no other timing, so whatever is generating the I2S signal IS determining the timing and jitter going into the DAC chip. Thus if the I2S signal is coming from another box/board, THAT is now directly determining the jitter going into the DAC chip. How the I2S signal gets between the boxes has a lot to do with how good that timing is really going to be.
It IS possible to use a local clock that will reclock the incoming I2S signals, but in order for this to work that local clock has to be sent to the source of the I2S signals so it can synchronize the signals to the local clock in the DAC. This requires a DAC that sends the local clock OUT and a source the synchronizes itself to the local clock. These are few and far between, and the few that do, do not always use the same clock signal or pins on the interface cable.
So lets take the more common case of an I2S source and DAC that do not synchronize to each other, the clock in the SOURCE is in charge. If the source has a REALLY good clock and the circuitry used to drive the link between boxes is REALLY low jitter, then this configuration MAY sound better than another interface using a local clock in the DAC. If the DAC does not have a particularly good clock AND the I2S source component DOES have a really good clock, then the I2S connection may sound significantly better. If the I2S source does NOT have a really good clock then the I2S connection is probably not going to be much better and may be worse.
Another aspect to this is that all of the box to box I2S implementations out there do NOT block leakage loops between the source power supply and the DAC power supply. IF a setup is using the approach where a clock is fed back from the DAC to the source it is possible to implement isolation on the I2S signals, but I don't know of anyone that has actually done that.
Now on to the details of different I2S implementations. Most I2S signaling is done as CMOS level digital signals on PC boards between chips. This type of signal is only good for a few inches on a PC board. ANYTHING else, especially between boxes needs something different.
The early implementations used a single ended line driver chip to drive 75ohm coax, the most popular implementation used a DIN connector, the same one use by the S_Video standard. There were several companies that used this.
Recently most implementations have shifted to a differential method (LVDS - Low Voltage Differential Signaling) sent over HDMI cables, primarily because they already exist and the cables have just the right number of wires to make this work. This LVDS over HDMI gives significantly better signal integrity than the earlier single ended implementation over S-Video connectors. Unfortunately not all CMOS <-> LVDS converters are really low jitter, so there can be higher jitter levels in the DAC than there should be.
With a well done LVDS interface on both the source and DAC the jitter in the DAC is going to be primarily determined by the source, if it does a good job, the jitter at the DAC chip will be good, if it does NOT do such a good job, then the jitter at the DAC chip will be higher.
So as with just about anything else in audio there are no absolutes here, it depends on the implementation in the DAC and in the I2S source component.
Digital Data Transmission Protocols
HDMI=ISSUES, USB=ISSUES, TOSLINK=ISSUES, WHAT ABOUT DLNA or Network?
DLNA etc are complex protocols riding on top of Ethernet, which in itself is a fairly complex protocol. This is going the wrong direction. The more complex the protocol the more work has to be done at the DAC which means more noise generated in the DAC to deal with those protocols.
I can come up with three requirements:
1) master clock is in DAC, right next to DAC chip(s).
2) protocol is very simple, preferably not bursty packet based.
3) full galvanic isolation
S/PDIF coax, I2S, HDMI don't meet #1 or #3
S/PDIF optical meets #2 and #3 but not #1
USB async meets #1 but not #2 or #3
Ethernet solutions meet #1 and #3 but not #2
So none of the common interfaces in use today meet all three.
So how do you you get something that gets all three?
The easiest way is to use optical and do two fibers, one going each direction, one sending the data from the computer to the DAC and one going from the DAC to the computer carrying the clock. If you do this right it works very well.
Isolation & Reclocking
USB Isolator advice neededWhat about Chord's upcoming 2Qute and Hugo TT which offer "galvanically isolated" USB 2.0 port? Did they come up with their own design, or is it marketing fluff?
You don't have to isolate BEFORE the USB receiver, you can isolate AFTER the USB receiver. The USB receiver chip is directly connected to the USB bus, but all the signals coming out of and into the receiver go through isolators. As long as you either have two power supplies or power the receiver (and input side of the isolators) from VBUS you have full galvanic isolation. Of course this has to be built in to the DAC, it's not a separate box that you can add to any USB DAC out there.
This quite easy to do, but you have to be careful of the implementation. ALL digital isolators add a lot of jitter to the signal, so you need to reclock the signals after the isolators. This means the low jitter master clock has to be on the DAC chip side, and that same clock gets fed back through an isolator into the "dirty" side and on to the USB receiver.
It's surprising how many designs get this wrong, they don't reclock or they put the master clock on the dirty side, both of which add a lot of jitter to the clock.
Uptone Audio RegenI don't recall John or Alex ever focusing on the REGEN as a USB reclocker. I think that's a mischaracterization of their technical explanations of its effectiveness.
As with many things it is not a simple yes or no. The term "reclocker" has several different meanings and some of them have a lot of baggage associated with them by audiophiles.
The basic engineering definition of reclocking is running a digital signal through a flip-flop. This will constrain the edges of the input signal to only change at the active edge (negative or positive) of the clock fed to the flip-flop. There a couple reasons to do this, one is to reduce jitter in the input signal (as long as the FF clock has lower jitter than the input signal), another is to synchronize the timing on the input signal to a different "domain".
Reclocking can be either synchronous or asynchronous. In synchronous reclocking the clock fed to the FF is exactly the same frequency as the clock used to generate the input stream (or an integer multiple). This can be because the to clocks were derived some the same clock, OR one of the clocks used a PLL to synchronize it to the other.
If the clocks are NOT the same this is called asynchronous reclocking, this can cause weird things to happen and can even result in bits getting lost. In some circumstances it actually does work, but you have to be very careful about it. Note this is completely different than asynchronous sample rate conversion. Don't get these confused. ASRC generates new bits to deal with the difference in clocks, asynchronous reclocking uses the same bits, but at different times.
For SPDIF there have been boxes called "reclockers" for quite some time, most of these employ some sort of PLL to generate a local clock that is synchronized to the timing of the incoming data. The incoming stream is then run through a FF being clocked by this local clock. IF the clock coming from the PLL has lower jitter then the result will be a lower jitter S/PDIF signal. Note that this is a very simple operation, the reclocker does not have to know anything about S/PDIF protocol, it just moves the edges to line up with the local clock. The primary reason for these boxes was to decrease jitter, but they also had the side benefit that they could also clean up the edges and if the designer did it right, provide a signal that more closely matched the impedance spec.
Now on to USB.
First off you cannot use the simple reclocker model as above with USB. It is a bidirectional bus, data goes both ways over the same wire, in order to do simple reclocking you need to know which way the data is going at any given time, but this is not easy with USB. There is no separate wire that says which way the bus is going. The ONLY way to do it is to actually decode all the bus transactions in order to figure which way the data is going. This takes a full blown USB protocol engine.
The easiest way to do this is a USB hub chip, it has a built in protocol engine and data buffer, a packet comes in, the data goes into the buffer, then it builds a new packet with the same data and sends it out the other end. The transmission is done with a local clock, so in a sense it IS being reclocked, but it is not the simple reclocking that is done in an S/PDIF "reclocker". An interesting aspect of this is that it is asynchronous reclocking, it uses a local clock that is not synchronized in any way to the computer clock, but this doesn't cause a problem because the data comes (and goes) in packets. If the local clock is a little slower than the computer clock the outgoing packets will take a little longer to transmit, but this doesn't cause a problem because there is a lot of dead time in between packets. What matters is the average rate of data, the local clock doesn't change this even if the rate of the bits speed up or slow down a little bit.
So yes the REGEN is a reclocker, but not the same thing as used in the S/PDIF reclocker boxes. The primary purpose for the REGEN was not the fact that it reclocks but that it builds a new wave form with better signal integrity. The reclocking comes along for free.