Solving Soundstage Collapse with Playback Devices

[StephenJK]

Some of you may recall the development of the Computer Audiophile Pocket Server, or CAPS.  I think development has been suspended because there was nothing more that could be done to improve performance?

 

I thought of building a CAPS, especially one of the v4 variants but for a number of reasons never got around to it.

 

We play a lot of music at home with a traditional two channel audiophile setup.  There isn’t a man cave, both my wife and I enjoy sitting at the supper table sipping on a glass of wine, listening to music and talking about our day.  If you want to watch TV, you have to go to the family room and put on a movie. 

 

To make a long story short, once we started on digital playback in the early ‘00’s the first thing was to rip CDs.  Next, how to record over 1,500 LPs to preserve the quality of the recording.  That ended up being with a Korg MR-2000S digital recorder that I eventually upgraded to a much larger hard drive.

 

We rarely sit in the sweet spot to listen to music, with our previous log home and an open concept the Klipschorn speakers could fill the house with music.  The clarity and detail was always there, but if you did sit in the sweet spot digital playback was not the same as the source material – in this case the LP.  We found that with playback from the laptop, the soundstage would collapse. 

What was once a wide and deep landscape with terrific dimensional imaging and accurate placement of instruments became restricted to something much smaller and with a loss of focus.

After a lot of testing with LP playback, playback directly from the recorder and then that same recorded file from the laptop we realized that was the culprit – the laptop.

 

Which bring us back to the CAPS.  I don’t remember Chris’s original justification for the CAPS other than trying to limit or eliminate noise.  Our solution was eventually to buy a Bryston BDP-2 digital player.  That’s just a Linux computer, but one, like the CAPS project, optimized for the best sound possible with minimal processing. 

 

The latest development is that the BDA-2 is now scheduled for replacement with a PS Audio Directstream Jr. that I expect it to arrive later this week.  I was hoping that with the new design and network connectivity options of the PS DSJ that I might be able to get rid of the BDP-2 player as well and go back to the laptop and DAC only.  It's a lot more convenient to use and gives more options for playback software.

 

The home that we’re in now after downsizing doesn’t have the footage of the previous, and the stereo now sits at the end of the dining room.  Essentially, we sit in the sweet spot all the time and can hear very subtle difference in music playback, especially now with the KEF Blade speakers.

 

I did a bunch of testing this weekend and yup, the soundstage collapse with the laptop playback is still there.  I’ve uploaded three different recordings of two test files to DropBox if anyone is interested.  Remember that the issue isn’t with clarity or detail, but with the developed soundstage – meaning you’ll need a good set of headphones or a system capable of presenting a decent soundstage to do a comparison.

 

I’ll test again once the new DAC is in place.  If anyone has any thoughts on how to optimize playback with the PS DSJ I would be interested in hearing them.  The unit does have the Bridge II card and can accept either a digital input or an Ethernet CAT5 connection.  It seems that PS Audio prefer USB as an input.  I’ll try that first, I’m curious if there will be the same issue as with the Bryston DAC but hope not.

 

The two test files I used are both from Mobile Fidelity Sound Labs LPs.  So Far Away from Dire Straits – Brothers in Arms (MFSL2-441 45 RPM) and Kiko And The Lavender Moon from Kiko (MFSL1-371).  For test purposes I believe the recording of these files falls under fair use. 

 

DropBox link;

https://www.dropbox.com/sh/tg445l0tpzuri1m/AAB-5HeLXV28HvR2D6nsfk7za?dl=0

 

There is a text file titled Notes on Files that explains what each recording represents.

 

Stephen

[ipeverywhere]

I recently replaced a MBP -> USB -> DAC with an Allo Digione Signature -> Coax -> DAC. This change did positively impact exactly what you're describing. The digital transport does make a difference and removing a "good" transport, your BDP-2, for the laptop is getting you the expected results. I'd focus back on removing the laptop from being directly connected to the DAC since you've proven it is not a great transport.

 

The DSJ can be fed directly over the network using various methods. It may very well be worth trying that. Remove the laptop from near the equipment. I'm assuming you've got your music directly plugged into the laptop via 1 or more HDDs? You can install Roon, UPnP controller, etc. (I don't know all the formats the Bridge supports) directly onto the laptop and then feed the audio over the network to the DSJr. Since this will cost you nothing but time I'd start there. Trying the USB from laptop to DSJr is certainly worth testing as well but, again, I think you're going to be unhappy simply because the laptop is a poor digital transport and your system is good enough to reveal that.

 

[fas42]

Soundstage collapse is the eternal problem with digital playback, and IME requires diagnosis that's specific to the rig - what happens is that some type of interference mechanism, or time dependent degradation of circuitry optimal functioning occurs, and low level detail is too poorly presented. Result is loss of "big picture" aspects in the sound.

 

As your first reply points out, rounds of experimenting will likely be necessary - there is at least one weak link, and it can be extremely frustrating to circumvent them.

[StephenJK]

16 hours ago, ipeverywhere said:

snip

The DSJ can be fed directly over the network using various methods. It may very well be worth trying that. Remove the laptop from near the equipment. I'm assuming you've got your music directly plugged into the laptop via 1 or more HDDs? You can install Roon, UPnP controller, etc. (I don't know all the formats the Bridge supports) directly onto the laptop and then feed the audio over the network to the DSJr. Since this will cost you nothing but time I'd start there. Trying the USB from laptop to DSJr is certainly worth testing as well but, again, I think you're going to be unhappy simply because the laptop is a poor digital transport and your system is good enough to reveal that.

I will, of course, test the new DSJr. DAC with the existing Bryston player as well as with a network configuration using Roon or some other type of connectivity.  My concern is that even with the configuration that PS Audio suggest there's still a laptop or computer of some type to pull the files from local or NAS devices and then feed it to the DAC as an endpoint.  But, if I continue to use the digital player as a control device that's certainly an equitable solution.

16 hours ago, ipeverywhere said:

 

 

[StephenJK]

12 hours ago, fas42 said:

Soundstage collapse is the eternal problem with digital playback, and IME requires diagnosis that's specific to the rig - what happens is that some type of interference mechanism, or time dependent degradation of circuitry optimal functioning occurs, and low level detail is too poorly presented. Result is loss of "big picture" aspects in the sound.

 

As your first reply points out, rounds of experimenting will likely be necessary - there is at least one weak link, and it can be extremely frustrating to circumvent them.

I don't believe that soundstage collapse with digital playback is well known or understood.  We know it isn't with all computers - are MACs OK, and PCs not?  I'm not going to go much further than what I already have - my interest was in being able to get rid of the Bryston BDP or not.  It's starting to look a lot like the latter, and I'm fine with that.  

 

But, as people argue over whether Foobar or JRiver is the best playback software, I wonder if there's a greater issue they aren't aware of.  I know that some people say soundstage isn't everything and some people don't care, but suspect that's from people who perhaps have never experienced what a great soundstage can do for any music.  You don't need a big fancy stereo, you just need a better than decent pair of headphones.

 

If you don't have a good stereo soundstage, is it just double mono?

[One and a half]

With frequent experimental changes in USB cables, I found the sound stage changes from sublime to dramatic. Depending on the cable and the source, the stereo sweet spot would drop from 2 feet above the speaker line and almost mono with distortion of the signal to boot.

 

The source was either, computer, Mac Mini, laptop, or a USB extender from a Xeon server. The common factor is either USB transmission, and the interaction of the USB cable to the source & DAC. Since the setup at home eliminating the USB portion of the signal chain has occurred, and using either S/PDIF or AES3 to the DAC, the sound stage has remained stable, usually 2 feet above the speaker line, and depending on the material, about the same at the right and left boundaries of the speaker, super wide, so to speak. For some classical recordings, instruments are spaced in 'air' at a fixed point in space, rather a nice moment, but doesn't happen too often for rock/pop recordings.

 

There is plenty of channel separation for digital components, that's not the problem, but the effect of noise and how it changes the sound stage is very real. IMHO, the noise from the computer to the DAC affects the timing between left and right channels, remove the noise and soundstage is restored. Removal is not that simple for USB, with many fixers vying for which one does best, but none really work 100%. Use a different method, eg, Ethernet -> DAC -> Amp or Ethernet -> Amp, removing USB altogether.

[StephenJK]

8 hours ago, One and a half said:

With frequent experimental changes in USB cables, I found the sound stage changes from sublime to dramatic. Depending on the cable and the source, the stereo sweet spot would drop from 2 feet above the speaker line and almost mono with distortion of the signal to boot.

 

The source was either, computer, Mac Mini, laptop, or a USB extender from a Xeon server. The common factor is either USB transmission, and the interaction of the USB cable to the source & DAC. Since the setup at home eliminating the USB portion of the signal chain has occurred, and using either S/PDIF or AES3 to the DAC, the sound stage has remained stable, usually 2 feet above the speaker line, and depending on the material, about the same at the right and left boundaries of the speaker, super wide, so to speak. For some classical recordings, instruments are spaced in 'air' at a fixed point in space, rather a nice moment, but doesn't happen too often for rock/pop recordings.

 

There is plenty of channel separation for digital components, that's not the problem, but the effect of noise and how it changes the sound stage is very real. IMHO, the noise from the computer to the DAC affects the timing between left and right channels, remove the noise and soundstage is restored. Removal is not that simple for USB, with many fixers vying for which one does best, but none really work 100%. Use a different method, eg, Ethernet -> DAC -> Amp or Ethernet -> Amp, removing USB altogether.

As an update, I've replaced the Bryston BDA-2 with a PS Audio Directstream Jr.  I can duplicate any test with the test files I recorded the first time around from LPs.

 

I installed the PS Audio DSJ using their USB drivers and JRiver ASIO out to the USB port of the DAC.  The other connection is with AES/EBU from the Bryston BDP-2 to the DAC.  Fortunately, the DAC with switch inputs automagically based on what's playing so that I can quickly switch from one to the other.

 

You know, it was close - very close.  With the updated USB drivers to the PS Audio DAC the sound was over 90% of the soundstage developed from the BDP.  I was pleasantly surprised, and can only presume that it isn't just USB, but likely a combination of factors, not the least being I'm still using a Dell laptop.  

 

I had another test that I could have done but didn't and may some day.  I have a Musical Fidelity V-Link 192 that I could install and have that output with SPDIF to the DAC rather than USB.  But for now I'll go with the existing configuration.  

 

I didn't bother setting up Roon and a network configuration because I have no interest (at this point) of investing in an NAS.  Playback for testing was from two identical USB 2 TB drives that hold copies of the master music library.  

 

I ended up staying with the Bryston player - it's a great unit and I don't have to invest any more in the stereo - at least not now.  

 

I have deleted the test files from DropBox that were linked in my first post.  I don't think anyone was really interested in my foolish little science experiment anyway.

[audiobomber]

Asynchronous USB is technically superior to S/PDIF, assuming a good quality USB implementation.

 

My Dell Inspiron 7000 laptop sounds very poor as a digital transport compared to my SMS-200. Have you tried the Microrendu with the DSJ? 

 

I believe your DAC is self-powered. Try blocking the 5V pin on the USB cable with electrical tape. That made a significant difference in my system.

[StephenJK]

Not that anyone is really interested, but in the interest of closing off this topic I believe I may have found a solution.  I installed my V-Link 192 USB/SPDIF converter.  I've had the V-Link for some time, where I used to test against the very issue I'm having now, loss of soundstage with a direct USB connection to a DAC.

 

The test setup is this, identical 2 TB external drives are connected to a laptop running JRiver and to a Bryston BDP-2 that I connect to with a web page.  The SPDIF from the converter and AES/EBU from the Bryston are connected to the inputs of the PS Audio DSJ DAC.  

 

Fortunately, the DAC will automatically switch to the live input.  I can control JRiver from my iPad with JRemote, and the Bryston with the web page or with an application called Soundirok which gives better volume control.  Mind you, for this test the files are played at 100%, I can control volume with the integrated amplifier remote.  

 

I cued up a number of high resolution titles, a couple from Dire Straits Brothers In Arms (MOFI 45 RPM), Los Lobos Light of the Moon (MOFI) and Dave Brubeck Take Five (Reissue).

 

Interestingly, the V-Link connection is slightly but very noticeably superior to the Bryston player.  There's more depth around the midrange, and light and quick sounds from piano notes or acoustic guitar tend to be more solid, more lifelike where with the Bryston there's a bit of shimmer or splash to it.  

 

It looks like playback from the laptop with not only the converter but that in combination with the new DAC presents more of a noise-free signal to the DAC, and that difference can definitely be heard.