A novel way to massively improve the SQ of computer audio streaming

[romaz]

Hi Romaz, based on your "last" clock should be the "best" clock finding, it will be interesting to hear the ISO REGEN between the Adnaco USB and my DAC here. I am already injecting LPS-1 power into the DAC via an AQVOX cable, so using the ISO REGEN as substitute is straightforward.

 

Also, I have to admit to being rather perplexed about the last clock rule. It seems to me that there are at least two independent​ clocks in the USB receiver and DA converters in our asynch DACs. Leakage currents aside, I don't know why these "last" clocks don't suffice to deliver great SQ. Improved SQ from the enhancement of the upstream clocks with cleaner power and better clock PPM suggests the signal is degraded in an irrecoverable fashion in these earlier stages.

 

If this is true, then it makes sense to reduce the number of clocks in the chain, as well as delivering the cleanest power and best PPM clock at each processing stage. To my mind this is the key argument against the 3 machine nas->upsampling->renderer model.

I agree, the Iso Regen could potentially change the music server landscape once again and I think we are all eager to see how well it performs. In my mind, it may come down to how good its clock is but I'm sure there are other factors to consider.

 

It has been eye-opening to realize just how much clocking and reclocking occurs in a typical audiophile setup. Anything that passes through a buffer requires a reclock. Any transfer from one bus to another (ie M.2 PCIE to SATA or system bus to USB) requires a reclock. Any CPU processing, rendering, conversion, upsampling, etc involves clocking. Anything that enters and leaves RAM is clocked. I think the ideal system would have equivalent good clocks from beginning to end and this is what I am aiming for with my upcoming build but again, it's simply not possible to replace any subclocks (DPLLs) that may be in the path. If it isn't possible to remove the bad clocks in your chain, my listening experiences have suggested that if you have one exceptional clock to use, that you use it as close to your DAC as possible. This boils down to the concept of signal integrity and wanting to provide your DAC the highest SI possible.

 

As to upstream signal degradation resulting in irrecoverable damage, I believe this is probably is true. At some point, the damage becomes so engrained into the signal that it becomes part of the signal and no "signal decrapifier" device made will be able to distinguish what is artifact and what is original. Here is what Paul Hynes told me once regarding the importance of applying clean power to every component in your system:

 

"All circuits require a reference to operate and react with each other. This is typically via a ground (0V) system. Some circuit sections also require a voltage reference above (or below) ground to provide precise operation and this voltage reference is usually connected to ground for its own reference. If the power supply is not clean and free from noise and transient disturbances, it will pollute the ground reference and anything connected to it creating uncertainty of reference voltage. This noise and transient disturbance can be passed on from stage to stage once embedded in the signal and it is quite capable of causing timing errors in data streams."

[romaz]

What I want to know are the MEASURABLE differences or based on listening only? Just asking.

As an objectivist at heart, I, too, like to see measurements and to be able to understand the "physiology" behind a dynamic process. The problem is in knowing what to measure and having instruments sensitive and precise enough to measure them.

 

Digital playback seems to have so many moving parts, some of them unknown or not well understood, that it's not just difficult to control variables, it's difficult to know what all the variables are. At this point, I still don't understand all that is going on with this "direct connection." Exactly what makes having the router in the path so deleterious?

 

The idea of why a good clock can make a difference is easier to grasp and SOtM has already published the impact of their prior clocks with respect to jitter. Hopefully, I'll get to see data on their newest clock soon but regardless of the measurements, in my system, it's impact is quite clear. It's been a while since I've experienced step changes this large for so little money.

[gadgetman]

What I want to know are the MEASURABLE differences or based on listening only? Just asking.

The result is based on my own hearing and my preferences. There's no scientific explanation of how I hear it. My environment and my setup is very likely different from others. So YMMV.

[greenleo]

Thanks Roy for initiating this project. I was thinking that it must take something quite special to lure you away from Head-fi!

 

We previously discussed via PM that I was holding off on improving the digital side of things given how fast things were moving and that it was daft to spend big bucks on the server side of things. Recently, I found some spare time (and courage) to read through all 922 posts in this thread. I have compiled a word document of the comparatively significant findings (highlighting different types of findings with colour code).

...

 

What a courage. Would you mind to share the doc?

[greenleo]

What I want to know are the MEASURABLE differences or based on listening only? Just asking.

That's what I've learnt from a quote: "If it sounds good and measures bad, you've measured the wrong things."

 

Did not answer your question but I find it quite true most of the time. Somehow CAS is not an exact science like Physics(that in general the evolution of a system is pretty much predictable) but something like Biology that too many factors affect a living thing. As a side note, I guess why vinyl or CD player(at least few years ago) sound so good when compare to CAS may be explained by reading this thread--the signal path is much simpler and the parts that needs improvement is much clear cut.

 

Regards

[botrytis]

That's what I've learnt from a quote: "If it sounds good and measures bad, you've measured the wrong things."

 

Did not answer your question but I find it quite true most of the time. Somehow CAS is not an exact science like Physics(that in general the evolution of a system is pretty much predictable) but something like Biology that too many factors affect a living thing. As a side note, I guess why vinyl or CD player(at least few years ago) sound so good when compare to CAS may be explained by reading this thread--the signal path is much simpler and the parts that needs improvement is much clear cut.

 

Regards

 

Well CD players are not that simple - they are similar to Computer Audio. They have a digital source, a DAC and have to output to you pre-, integrated, etc. Computer Audio can be just more distributed is all. I am not sure all this tweaking is doing anything - reclocking and reclocking an reclocking? I can see once, maybe. The trick is the computer needs a good PS and a name brand MB with good components. I would like to see a blind test on this.

[greenleo]

Well CD players are not that simple - they are similar to Computer Audio. They have a digital source, a DAC and have to output to you pre-, integrated, etc. Computer Audio can be just more distributed is all. I am not sure all this tweaking is doing anything - reclocking and reclocking an reclocking? I can see once, maybe. The trick is the computer needs a good PS and a name brand MB with good components. I would like to see a blind test on this.

 

 

CD as transport is relatively simple:

Rotate the disc in constant speed, read it through the laser head, and then output to the DAC. No error correction (real time). Hence the name CD transport, and why a good CD sounds better than some music files stored in CDs with less quality.

Of course in this case a DAC is needed. In the old days that Sony just launched their Discman, the DAC was integrated into the CD player and the term CD transport or CD as an transport carries very little meaning.

 

 

For the blind test, I encourage you to go ahead. There are situations that blind test are even not needed because the differences were (and are) so obvious.

 

Romaz, among many other CA members, has explained things happening in the signal path eloquently and I won't repeat here. A computer with good PS and quality components is just the beginning (let alone the definition of good and quality), not an end.

 

Enjoy.☺

[romaz]

I am not sure all this tweaking is doing anything - reclocking and reclocking an reclocking? I can see once, maybe. The trick is the computer needs a good PS and a name brand MB with good components. I would like to see a blind test on this.

Your viewpoint is fine and I respect it. If all it takes to make you happy is a good PS and a name brand MB, then I'm totally cool with that.

 

Just to be clear, I have no hidden agenda with this thread and I have tried to be as transparent with my process as possible. I'm fine if you think I don't have a clue or am somehow incompetent but I'm not fine if you think I have something to sell or promote because it's not true. As a matter of full disclosure, no one has put words in my mouth or money in my pocket. What I'm reporting are my experiences in the context of my biases and so if they are helpful to others, that's great, and if not, then fortunately you have the freedom to look elsewhere.

[Superdad]

CD as transport is relatively simple:

Rotate the disc in constant speed, read it through the laser head, and then output to the DAC. No error correction (real time).

 

That is incorrect. Real-time CD playback does have the disc rotating at a constant speed, and there is error correction--but it is based on interpolation.

 

You may want read up on Reed-Solomon encoding and eight-to-fourteen modulation.

 

In fact the single largest factor in "computer audio" (where CDs are ripped, stored, and played back) being superior to conventional CD transports is the utilization of CD ROM drive data mode extraction--instead of realtime, variable-speed, Reed-Solomon error correction, etc.

 

 

Sent from my iPad using Computer Audiophile

[romaz]

With that aside and on the subject of "name brand" motherboards, as I'm waiting for other components to come in with my build, I got a chance to examine my Gigabyte GA-H110TN motherboard ($110) in closer detail today and this is really an excellent motherboard. I don't know where this experiment will lead and maybe it will lead nowhere further but if replacing key clocks turns out to lead to significant enough SQ improvements to warrant the purchase of SOtM's new clock board, I can now see that this Gigabyte motherboard could be used for a fairly inexpensive "single box" A/V server.

 

Here is a schematic of this motherboard:

 

 

If you look carefully, you will see 5 raised oval shaped silver metal objects on the motherboard. If you notice the PCIEx4 slot on the top left of the photo, one of these oval silver objects is located just north of the slot. These represent the motherboard clocks. Some motherboards only have 1-2 clocks that you can replace (meaning the subsystems are all clocked by DPLL which is less ideal) and so it is pretty special that this motherboard has this many clocks that you can replace (and power with a single high quality PSU). Anyway, as I inventory the clocks, one is a system clock that is responsible for the most important timings (either directly or indirectly) including the CPU, USB, PCIE and SATA buses. Two of the clocks are for the two Intel gigabit LAN ports. One clock is for the audio subsystem. The last clock is for the video subsystem.

 

Maybe you can see where I'm going with this but as SOtM's clock board has 4 clocks with independent frequencies that can be utilized, as an alternate build to my own, one could be used to replace the system clock, one could be used to replace a clock for one of the LAN ports, one could be used to replace the audio clock and the last one could be used for the video clock.

 

The benefits of replacing the system clock should be more clear. Less OS errors with greater precision and efficiency of flow, fewer traffic jams, less latency, less noise injected back into the ground plane and less jitter. Those who have reported replacing their motherboard's system clock have reported much quicker boot times and more responsive computing, exactly what you would expect and the SQ improvement you might get will probably be similar to what you would experience with Audiophile Optimizer + Process Lasso. Those who have followed the development of the SGM 2015 know that they have gone to great lengths to find just the right replacement clock for their motherboard.

 

The benefits of replacing the audio clock could be just as important for those looking for a really good one box solution. As we know, Toslink has the advantage of completely blocking RF noise. The main problem with Toslink is that it is very jitter prone although with its clock replaced by the SOtM clock, provided your DAC has an optical input, this could prove to sound wonderful without the need for anything extra.

 

The benefits of replacing the video clock could prove to result in improved video output. The latest generation of Intel CPUs with their integrated low power GPUs all do a fantastic job of decoding all the high-def video codecs including 4K and so it remains to be seen what replacing the stock clock will result in but I'm willing to bet that video is as susceptible to good clocking as audio is.

 

Unfortunately, there aren't 2 clocks left to replace the clocks on both LAN ports although replacing only one of the clocks should suffice if you decide to use the Toslink output. My thought would be not to incorporate any internal storage on this build and instead use a "direct connection" to a NAS where music and movies can be stored with the added benefit that the stream from the NAS will be reclocked by the SOtM clock. The other LAN port (the one with the stock clock) could then be used to connect to the router for the purposes of DHCP as well as streaming services.

 

The SOtM clock board with 4 clocks would be the most expensive component at $1100 but the rest of the components to complete the build should only cost a few hundred dollars total. Of course, you will want to power it with the best 12V PSU you can afford but that should go without saying.

[greenleo]

That is incorrect. Real-time CD playback does have the disc rotating at a constant speed, and there is error correction--but it is based on interpolation.

 

You may want read up on Reed-Solomon encoding and eight-to-fourteen modulation.

 

In fact the single largest factor in "computer audio" (where CDs are ripped, stored, and played back) being superior to conventional CD transports is the utilization of CD ROM drive database mode extraction--instead of realtime, variable-speed, Reed-Solomon error correction, etc.

 

 

Sent from my iPad using Computer Audiophile

 

Hi Superdad,

 

If the interpolation is taken as the true error correction, then I agree with you. But, the interpolation is the best guess if the data is allowed to read only once. This attempt to do correctioin doesn't guarantee exactly the same results or physically poor quality CD would sound the same as that of good physical quality.

 

My point is that the actual signal path and the actual processing took place affects the SQ. I am a CAS user (or I'll not be here☺) rather than a CD player promoter. However, the simple signal path of a good CD player that contributes to the SQ may not be ignored.

 

For the data extraction mode of CDs, it has been done for many years. However, the same music file sounds differently in different players (say foobar, JPlaymini), and two bit-by-bit-wise identical files sound differently on the same music players tells that CAS still has a long way to go because there are simply too many factors affecting the SQ. I feel that this journey contributes to the fun of being a CA. Lol.

[Theobetley]

With that aside and on the subject of "name brand" motherboards, as I'm waiting for other components to come in with my build, I got a chance to examine my Gigabyte GA-H110TN motherboard ($110) in closer detail today and this is really an excellent motherboard. I don't know where this experiment will lead and maybe it will lead nowhere further but if replacing key clocks turns out to lead to significant enough SQ improvements to warrant the purchase of SOtM's new clock board, I can now see that this Gigabyte motherboard could be used for a fairly inexpensive "single box" A/V server.

 

Here is a schematic of this motherboard:

 

[ATTACH=CONFIG]33721[/ATTACH]

 

If you look carefully, you will see 5 raised oval shaped silver metal objects on the motherboard. If you notice the PCIEx4 slot on the top left of the photo, one of these oval silver objects is located just north of the slot. These represent the motherboard clocks. Some motherboards only have 1-2 clocks that you can replace (meaning the subsystems are all clocked by DPLL which is less ideal) and so it is pretty special that this motherboard has this many clocks that you can replace (and power with a single high quality PSU). Anyway, as I inventory the clocks, one is a system clock that is responsible for the most important timings (either directly or indirectly) including the CPU, USB, PCIE and SATA buses. Two of the clocks are for the two Intel gigabit LAN ports. One clock is for the audio subsystem. The last clock is for the video subsystem.

 

Maybe you can see where I'm going with this but as SOtM's clock board has 4 clocks with independent frequencies that can be utilized, as an alternate build to my own, one could be used to replace the system clock, one could be used to replace a clock for one of the LAN ports, one could be used to replace the audio clock and the last one could be used for the video clock.

 

The benefits of replacing the system clock should be more clear. Less OS errors with greater precision and efficiency of flow, fewer traffic jams, less latency, less noise injected back into the ground plane and less jitter. Those who have reported replacing their motherboard's system clock have reported much quicker boot times and more responsive computing, exactly what you would expect and the SQ improvement you might get will probably be similar to what you would experience with Audiophile Optimizer + Process Lasso. Those who have followed the development of the SGM 2015 know that they have gone to great lengths to find just the right replacement clock for their motherboard.

 

The benefits of replacing the audio clock could be just as important for those looking for a really good one box solution. As we know, Toslink has the advantage of completely blocking RF noise. The main problem with Toslink is that it is very jitter prone although with its clock replaced by the SOtM clock, provided your DAC has an optical input, this could prove to sound wonderful without the need for anything extra.

 

The benefits of replacing the video clock could prove to result in improved video output. The latest generation of Intel CPUs with their integrated low power GPUs all do a fantastic job of decoding all the high-def video codecs including 4K and so it remains to be seen what replacing the stock clock will result in but I'm willing to bet that video is as susceptible to good clocking as audio is.

 

Unfortunately, there aren't 2 clocks left to replace the clocks on both LAN ports although replacing only one of the clocks should suffice if you decide to use the Toslink output. My thought would be not to incorporate any internal storage on this build and instead use a "direct connection" to a NAS where music and movies can be stored with the added benefit that the stream from the NAS will be reclocked by the SOtM clock. The other LAN port (the one with the stock clock) could then be used to connect to the router for the purposes of DHCP as well as streaming services.

 

The SOtM clock board with 4 clocks would be the most expensive component at $1100 but the rest of the components to complete the build should only cost a few hundred dollars total. Of course, you will want to power it with the best 12V PSU you can afford but that should go without saying.

 

All your work on clocks brings back some recollection of when I had a Buffalo 32 Diy dac. I loved that dac it was my 1st real life diy project. It wasn't that complicated . All I had to do was assemble the power supplies and mount everything on a board. I never put the module in a case, although I bought a case. So I left all the dac components fully exposed including the clocks. Anyway I remember going through a phase wherein I tried different levels of sound deadening material pressed onto the clocks. Darn if that didn't make an improved difference as it should. So today I fixed some sound deadening material (mortite) on top of the last clock prior to my dac (the microRendu). I think it makes a difference. So Romaz have you tried this? Or said another way could you try?

[gstew]

That is incorrect. Real-time CD playback does have the disc rotating at a constant speed, and there is error correction--but it is based on interpolation.

 

You may want read up on Reed-Solomon encoding and eight-to-fourteen modulation.

 

In fact the single largest factor in "computer audio" (where CDs are ripped, stored, and played back) being superior to conventional CD transports is the utilization of CD ROM drive data mode extraction--instead of realtime, variable-speed, Reed-Solomon error correction, etc.

 

 

Sent from my iPad using Computer Audiophile

 

Alex,

 

I think you mixed up speed and rate here. Audio CD playback reads the data off the discs at a constant rate, but the rotation speed varies from 480rpm - 210rpm due to the circular track length varying from inner to outer circumferences.

 

From: https://en.wikipedia.org/wiki/Revolutions_per_minute

 

"Audio CD players read their discs at a precise, constant rate (4.3218 Mbit/s of raw physical data for 1.4112 Mbit/s (176.4 kB/s) of usable audio data) and thus must vary the disc's rotational speed from 8 Hz (480 rpm) when reading at the innermost edge, to 3.5 Hz (210 rpm) at the outer edge.[1]"

 

Greg in Mississippi

 

P.S. I've made/used several DIY'd CD Players where the disc spun out in the open air... I was always amazed at how much difference there was in the speed from beginning to end.

[romaz]

All your work on clocks brings back some recollection of when I had a Buffalo 32 Diy dac. I loved that dac it was my 1st real life diy project. It wasn't that complicated . All I had to do was assemble the power supplies and mount everything on a board. I never put the module in a case, although I bought a case. So I left all the dac components fully exposed including the clocks. Anyway I remember going through a phase wherein I tried different levels of sound deadening material pressed onto the clocks. Darn if that didn't make an improved difference as it should. So today I fixed some sound deadening material (mortite) on top of the last clock prior to my dac (the microRendu). I think it makes a difference. So Romaz have you tried this? Or said another way could you try?

This is interesting what you've done. I didn't even know what mortite was until now when I looked it up. Is the intention vibration control or mitigation of RF?

[scan80269]

Alex,

 

I think you mixed up speed and rate here. Audio CD playback reads the data off the discs at a constant rate, but the rotation speed varies from 480rpm - 210rpm due to the circular track length varying from inner to outer circumferences.

 

From: https://en.wikipedia.org/wiki/Revolutions_per_minute

 

"Audio CD players read their discs at a precise, constant rate (4.3218 Mbit/s of raw physical data for 1.4112 Mbit/s (176.4 kB/s) of usable audio data) and thus must vary the disc's rotational speed from 8 Hz (480 rpm) when reading at the innermost edge, to 3.5 Hz (210 rpm) at the outer edge.[1]"

 

Greg in Mississippi

 

P.S. I've made/used several DIY'd CD Players where the disc spun out in the open air... I was always amazed at how much difference there was in the speed from beginning to end.

 

Greg,

 

CD players typically read the disc with CLV (constant linear velocity) for a constant bit rate as you said, but many computer CD/DVD drives used to rip CDs will spin the disc at high speed (can be 8000RPM or higher) in digital audio extraction mode. This is because no one has the patience for a disc to be ripped at 1X speed: over an hour depending on content total duration. In DAE mode, there is no constant bit rate requirement, and in fact the faster the better.

 

C1 errors of CDs are completely correctable using Reed-Solomon code. C2 errors are uncorrectable and will trigger CD players to do interpolation to fill in gaps where the data went bad, or mute briefly at the error location resulting in clicks. For CD/DVD drives, C1 errors are fully corrected as with CD players, but C2 errors are usually just reported (assuming the drive is capable) and it's up to the ripping software to decide how to deal with these errors. For example, dBpoweramp gives options on how many times to re-read sectors with C2 errors. A CD with lots of C2 errors (e.g. badly scratched) will typically cause the ripping progress to slow way down with re-read attempts, unless the SW is configured to ignore the C2 errors and move on (resulting in some bad data within the ripped stream).

[Theobetley]

This is interesting what you've done. I didn't even know what mortite was until now when I looked it up. Is the intention vibration control or mitigation of RF?

Vibration control.

[lmitche]

Vibration control.

Isn't Moretite caulking? http://www.homedepot.com/p/Frost-King-E-O-90-ft-Grey-Weather-Strip-and-Caulking-Cord-B2/100152937

[Theobetley]

Isn't Moretite caulking? Frost King E/O 90 ft. Grey Weather-Strip and Caulking Cord-B2 - The Home Depot

Yes. I have used it on cone drivers and elsewhere to kill vibrations. Use of mortite goes back to the early days of TAS.

[Superdad]

Alex,

 

I think you mixed up speed and rate here. Audio CD playback reads the data off the discs at a constant rate, but the rotation speed varies from 480rpm - 210rpm due to the circular track length varying from inner to outer circumferences.

 

Hi Greg: I mixed it up only in writing, not in my mind. I entirely meant to say "Real-time CD playback does NOT have the disc rotating at a constant speed, and there is error correction--but it is based on interpolation." I blew the very correction I was trying to make to greenleo! Oh well, it was a long day...

[lmitche]

Yes. I have used it on cone drivers and elsewhere to kill vibrations. Use of mortite goes back to the early days of TAS.

I thought so, how do you apply it to a clock in circuit board?

[Theobetley]

I thought so, how do you apply it to a clock in circuit board?

 

 

Just push it on the outer case of the clock (if you have access to it) hard enough so that it sticks.

[romaz]

Just push it on the outer case of the clock (if you have access to it) hard enough so that it sticks.

Here is a photo of the clock board. The 4 clocks are in the left upper quadrant of the board. I presume this stuff comes off pretty easily?

 

[austinpop]

I agree, the Iso Regen could potentially change the music server landscape once again and I think we are all eager to see how well it performs. In my mind, it may come down to how good its clock is but I'm sure there are other factors to consider.

 

It has been eye-opening to realize just how much clocking and reclocking occurs in a typical audiophile setup. Anything that passes through a buffer requires a reclock. Any transfer from one bus to another (ie M.2 PCIE to SATA or system bus to USB) requires a reclock. Any CPU processing, rendering, conversion, upsampling, etc involves clocking. Anything that enters and leaves RAM is clocked. I think the ideal system would have equivalent good clocks from beginning to end and this is what I am aiming for with my upcoming build but again, it's simply not possible to replace any subclocks (DPLLs) that may be in the path. If it isn't possible to remove the bad clocks in your chain, my listening experiences have suggested that if you have one exceptional clock to use, that you use it as close to your DAC as possible. This boils down to the concept of signal integrity and wanting to provide your DAC the highest SI possible.

 

As to upstream signal degradation resulting in irrecoverable damage, I believe this is probably is true. At some point, the damage becomes so engrained into the signal that it becomes part of the signal and no "signal decrapifier" device made will be able to distinguish what is artifact and what is original. Here is what Paul Hynes told me once regarding the importance of applying clean power to every component in your system:

 

"All circuits require a reference to operate and react with each other. This is typically via a ground (0V) system. Some circuit sections also require a voltage reference above (or below) ground to provide precise operation and this voltage reference is usually connected to ground for its own reference. If the power supply is not clean and free from noise and transient disturbances, it will pollute the ground reference and anything connected to it creating uncertainty of reference voltage. This noise and transient disturbance can be passed on from stage to stage once embedded in the signal and it is quite capable of causing timing errors in data streams."

 

Roy,

 

In your relentless pursuit of the best clocks, do you wonder if you might still have leakage loops remaining?

 

This is where the Intona or ISO-Regen would normally add isolation, but at the cost of a bad clock (potentially) relatively close to the DAC.

 

I wonder if another option would be an EMO EN-70HD between your Mini and your modified Trendnet. This would give you another layer of galvanic isolation, but still have 3 "best" clocks between it and the DAC to recover any poor clocking.

 

Might be worth a try?

[Theobetley]

Here is a photo of the clock board. The 4 clocks are in the left upper quadrant of the board. I presume this stuff comes off pretty easily?

 

[ATTACH=CONFIG]33737[/ATTACH]

 

Yes I comes right off.

[gadgetman]

Here is a photo of the clock board. The 4 clocks are in the left upper quadrant of the board. I presume this stuff comes off pretty easily?

 

[ATTACH=CONFIG]33737[/ATTACH]

 

Hi Romaz,

 

On the lower left corner there's a rectangular silver top with golden edge, it looks like an OSC. Do you know what is it for? Thanks a lot for sharing it.