Audiowise Opto-USB, SRC-DX and early experiments with the Chord Qutest before moving to TT2

[BlueDL]

2 hours ago, Middy said:

The only other option I have considered and a little cheaper was the Singxer SU-2... An optional World clock interface and a chance to mod with an LPS... Like most of this hobby..in your system is the proof of the pudding...

Hi Middy,

 

Is the logic there to feed the Qutest with an optical cable?

I tried something similar using a Matrix X-SPDIF2 into my TT2, half expecting to hear ‘RF busting benefits’ from it. But there was little if any difference versus going USB direct into TT2, and that was using a QED Reference glass optical cable too (which is great value given its spec).

 

No idea how the Qutest’s USB input fairs against it’s optical input or compares to the TT2 implementation (no experience of Qutest), so others may be able to advise or ‘chime in’ here.

[BlueDL]

6 hours ago, edwardsean said:

From the response, I'm moving to purchase the SRC-DX again. However, I remember that I wasn't too happy with the quality of the included BNC cables. So, I'm looking into better cables. 

 

Please let me know some suggestions. I don't think I want to spend more than the price of the SRC-DX for the pair, so around $500.

 

Thanks!

That should get you a pair of Oyaide DB-510, which I’ve found to perform well with M-Scaler, and now SRC-DX.

[BlueDL]

Nice one, Dave.......I mean......Middy.

 

I was wondering if a USB-powered decrapifier between source and SRC-DX might be drawing too much current from the source. If you don’t mind me asking, what decrapifier are you using? Thanks.

[BlueDL]

13 minutes ago, Middy said:

Hi @edwardseanreally glad its improved your setup. I am still not sure of whats so special about the DX. 

Would any old DDC do the job? But sounds so much better than USB alone.

My experiments are paying off on my little set up. Shielding, RAM cables, USB connectors on my pc worked but adding EMI absorber much more so..

3M AB5030.

 

The xilinx in my intona some ICs and USB connector shells.. a touch more SQ with 3m

Custom shielded supra cable shortened...Tick...

Windows debloater and scripts very nice..opening cores ect no power saving throttling for low latency... Big tick....

The main drive being EMI RFI the DX designer bases his products and engineering principles on....

 

The Qutest will be the last to be stripped and shielded.... i never knew about noise being reflect about the case not just noisy FPGAs and clocks transmitting RF emi..  pure luck but its working out and better SQ. Fingers crossed for the dac...

 

Lastly i just got this... from his website.

The intona ipurifier should block DC but it seems to help... subtle YMMV ECT

One thing you'll need is an adapter as its  a male/female in-line filter..

From a DX you need a female to female..

 

"DC•BLOCK Leakage Current Isolation – AudioWise" https://audiowise-canada.myshopify.com/products/gnd-zro-signal-ground-isolation

 

https://www.ebay.co.uk/itm/BNC-MALE-to-MALE-COUPLER-ADAPTER-FOR-TV-CCTV-CAMERA-CABLE-JOINER-CONNECTOR-/332086047040?mkcid=16&mkevt=1&_trksid=p2349624.m46890.l49286&mkrid=710-127635-2958-0

 

If you asked me 2 weeks my system was ok.  Eye's closed i could be tricked into believing its a good streamer TT2.

 

Finally a real step up in sound quality not perceived mini gains or marginal at best...

 

So happy..

 

Dave...

At this rate, I’ll be hiring you to help one of my team out with an EMC problem on a bit of aerospace kit. 😆

Two bites at the cherry (formal test campaigns) so far and both times we’ve not been able to fall within the stringent dB/frequency spectrum envelope! 🤔🙄

[BlueDL]

35 minutes ago, Middy said:

Here's what it looks like in situ..

Hi Dave. Is that the DC-Block filter from Audiowise that you linked to?

 

Has it had a tangible impact similar to the SRC-DX?

[BlueDL]

Middy, remind me, do you use an aftermarket power supply with your Qutest?

This is a tweak that many seem to swear by.

[BlueDL]

I am surprised that the ifi DC iPurifier made a difference given what was already in the chain.

 

I know what you mean about being careful with the spends. Getting a Farad Super3 to power my M-Scaler was somewhat extravagant by my standards.

 

Not sure what the score is with the Paul Hynes business? The Sean Jacobs DC3 was the go-to ‘affordable’ boutique power supply until it was discontinued recently. I personally think that was an own goal, because it struck me as the performance-value sweet spot. [I cringe at using a football analogy after last night’s farce 🤦🏻‍♂️]

[BlueDL]

“EMC seems like voodoo to some and a necessary evil you live with.”

This resonates with me due to an issue with a product in development at my work. The unit in question needs to be able to cope with a high bay ambient temperature at a ridiculous altitude (well above civil airliners), so thermal considerations mean we can’t use the likes of your techniques! I guess it’s all about Chokes/Ferrites, feed-through filters on connectors, braided screens, etc.

 

My TT2 is still in warranty, so I’ll not have it apart just yet!

[BlueDL]

31 minutes ago, MarkusBarkus said:

the only error I see is in the order of PK takers yesterday. Or too late with Grealish. Ugh.

The other error (in my opinion) is the formation that England started with. I understand that the Italian centre halves  are very experienced players, but age impacts one’s pace, and I’d have liked England to have gone with a more attacking 4-2-3-1 formation from the start. Having said that, well done to Italy (and their goalkeeper is already clearly world class yet very young).

 

If these EMI tweaks improve the DAVE, it will pique my interest even further! 🤔

[BlueDL]

49 minutes ago, OldBigEars said:

I use an OpticalRendu that's limited to Roon - so HQP is not an option.

Hi. I use an OpticalRendu as an NAA with HQP embedded.

You probably meant that your Server can’t host HQP for some reason, but if it can, it’s a worthwhile investment IMHO. The developer releases regular updates, so it’s not standing still either. 

[BlueDL]

Ah.......maybe your OR is the ‘Roon Edition’........in which case, ignore the above. Personally speaking, I think the SRC-DX beats USB regardless of what you send it (even with ‘only’ Roon).

[BlueDL]

3 minutes ago, davide256 said:

I dabbled with HQPlayer but did not find any compelling reason before PGGB to use it vs Euphony Stylus at native PCM rates. Since using PGGB I have more appreciation for it and do use Stylus to HQP output without upsampling, it does resolve better at PGGB upsampled rates. Not a fan of "massive computing"

HQP servers as that makes much more difficult/costly building an electrically quiet machine and matching power supply.

 

For taps I take the default of 512 at 7xx which is dictated by RAM quantity... would need to go from 32Gb ( ~ 5 minutes) to 128GB of ram 

to do 2 million (~20 minutes). You can override PGGB to do more but that means with less RAM it will use swap file much more, run slower for conversion.

What would you recommend if I were to be limited to the 16GB RAM of my 2012 Mac Mini for PGGB processing duties?

I have a Chord Hugo TT2 and M-Scaler and might consider replacing the latter with SRC-DX and PGGB.

[BlueDL]

1 hour ago, davide256 said:

 

I had problems using Win10 HQP Desktop with NAS, just very slow to scan/library refresh vs the others. So much prefer using another program as "librarian" for

NAS storage, feeding playlist to HQP embedded.

I currently use Roon feeding HQPe on a Linux based server. 
Is the SRC-DX unable to process 7XX/24 files then? Surely it’s purpose is to do exactly that to feed the dual BNC input of a modern Chord DAC?

Apologies if I’ve misunderstood somewhere.

So, I get the impression that my Mac won’t cut the mustard if I want a library of 7XX/24 files. Damn. 

[BlueDL]

I wonder if anyone could link me to a PC that has the necessary oomphf to process to 24/7XX at least 2 billion taps in a sensible amount of time. I get the gist that more cores and more RAM equals a faster processing time, but I wonder what’s reasonable without having to remortgage to pay for it.

 

By the way, I wish all on here a great Christmas or winter break.

[BlueDL]

3 hours ago, LowOrbit said:

Hi

 

It's not an easy thing to quantify. If you have a lot of lengthy files more RAM (min 32gb) is going to help. I went in stages from 16gb to 32gb then 64gb and the process got a bit quicker each time. SSD space to run a fast swap file also is a big factor (it is possible to run out of memory space during long file processing when you're aiming for 7xx output files). I bought a 500gb m.2 ssd to cover this (not hugely expensive) and this probably represents a better option than inflating ram. I already had a beefy i7 8-core pc on a good motherboard. RAM and hard drives are non-trivial expenses if you have an extensive library to convert. I use spinning discs for storage and copy an evening's listening to SSD on my replay machine (so I am not having hard drives spinning and chattering away during listening as my replay computer is close to where I sit for listening).

 

Also, just to be clear "billions of taps" is not typical as ultimate tap length is dependent on track length so varies every file (user can define a limit, but not specify a target bigger than the PGGB process will utilise). As a guide a typical five minute track is hitting maybe 250m taps. Longer tracks take proportionally longer to process because of this.

 

So - I think 4 cores, 32gb RAM and 500gb ssd for the paging file, plus output file storage represents the minimum. Another point worth clarifying is tap length does not correlate to output file size. Only the output sample rate (e.g 768k/s) and file format (WAV vs FLAC or WAVPAC) determine file size.

 

Compare the price of buying a pc to meet this spec with the $4k+ for an MScaler.

 

 

 

 

Thanks - this is very enlightening. I am an M-Scaler owner looking to ‘scratch the PGGB itch’.

The one file that a colleague processed in PGGB consistently ceased playback after around 30 seconds, so there may be an issue with my Sonictransporter i7 reading files this size from an external SSD caddy. I got a “file was too slow to load” message.

[BlueDL]

21 minutes ago, creativepart said:

 

I did suggest that he add more specific cautions about how his product can be an issue with specific devices like the SoTM streamer.

Yes, this should be made more clear.

I got lucky with my Sonore OpticalRendu as there is a simple fix in the Sonicorbiter OS settings where you just set the ‘DAC’ to max volume.

[BlueDL]

Hopefully they will come up with a hardware tweak for SRC-DX (a ‘Version 2’) in the near future so you get another bite at the cherry with your SotM. I share your dislike of messing with Linux......this hobby shouldn’t really  need a computer science degree! 🤣