[bfreedma]

   
Disclaimer: I haven't read the article yet.
 
That method sounds a bit flawed though. The primary problem is you are counting on the other sound card's ADC to be just as good as the main soundcard's DAC. This is not always the case, especially for consumer equipment.

 
One relatively simple way to do this would be to run each machine as a Virtual Machine on VMware - it would be easy to create identical resource pools for the two VMs and assign them to a processor not used by any other function.  Of course, people would then blame VMware for not being sufficient to identify the differences....
 
Somewhat embarassingly, I do have a VMware server at home.  If I have a few slow work days, I may give it a go.  Perhaps someone more familiar with accurately comparing the output can log in remotely and take on that part of the testing.

 

[prot]

Disclaimer: I haven't read the article yet.

That method sounds a bit flawed though. The primary problem is you are counting on the other sound card's ADC to be just as good as the main soundcard's DAC. This is not always the case, especially for consumer equipment.

Noone said the test was easy .. or perfect. And I guess it's a good idea to use an ADC that is better than the DAC used for playing. But as long as the only variable is the different software used for playing, the test should be ok. You are only testing for differences, not absolute quality.

 

[fastfourier]

Noone said the test was easy .. or perfect. And I guess it's a good idea to use an ADC that is better than the DAC used for playing. But as long as the only variable is the different software used for playing, the test should be ok. You are only testing for differences, not absolute quality.

 
Right, but the point is to not introduce more error into the measurement via a potentially noisy/not up to spec ADC.
 
The best way I can think of to do this would be with a good oscilloscope, since that's what it's designed to do. Two channels, and you can run a diff right on the scope. The ADCs in oscilloscopes are designed just for this sort of thing, with a tremendously higher sampling rate than the audio stream, so you're bound to get a much more accurate result from it. 
 
Of course, that requires an oscilloscope, but it wouldn't surprise me if many folks here have access to one. (Hell, I have one on my bench, they're not terribly expensive these days)

 

[prot]

Right, but the point is to not introduce more error into the measurement via a potentially noisy/not up to spec ADC.

The best way I can think of to do this would be with a good oscilloscope, since that's what it's designed to do. Two channels, and you can run a diff right on the scope. The ADCs in oscilloscopes are designed just for this sort of thing, with a tremendously higher sampling rate than the audio stream, so you're bound to get a much more accurate result from it. 

Of course, that requires an oscilloscope, but it wouldn't surprise me if many folks here have access to one. (Hell, I have one on my bench, they're not terribly expensive these days)

Sounds good. Just do it

 

[coli]

I achieved major improvement with jplay. WIndow audio is just horrible. 

 

[Yoga]

  I achieved major improvement with jplay. WIndow audio is just horrible. 

 
If you're running OSX, Aurdivana Plus (2.0.9) is amazing. Buffers, auto optimises system, INT mode, BP etc. Best I've heard so far.

 

[dglow]

   
If you're running OSX, Aurdivana Plus (2.0.9) is amazing. Buffers, auto optimises system, INT mode, BP etc. Best I've heard so far.

++ to that. For me, it's the relatively recent introduction of Direct Mode in A+ that provides the big, final step to complete transparency for USB audio delivery on Macs. 
 
Before moving to Audirvana, Bitperfect was my friend. Yet something always disturbed me: that employing different sample rate converters would change the sonic signature. How could this be? I was not oversampling. I ran in integer mode. Bitperfect auto-switched my DAC's sample rate to match the source material. Why were these software SRCs in the audio path at all?
 
OS X's Core Audio was the answer. Direct Mode circumvents this layer entirely, and I can't recommend the sound (or lack thereof!) enough.

 

[Music Alchemist]

  I achieved major improvement with jplay. WIndow audio is just horrible. 

  If you're running OSX, Aurdivana Plus (2.0.9) is amazing. Buffers, auto optimises system, INT mode, BP etc. Best I've heard so far.

 
HQPlayer is the only player I've found that gives a very obvious night and day improvement. With the old JPLAY, I wasn't sure I was hearing a real difference, but with this one, there is no doubt. Unfortunately, the interface is atrocious.

 

[ZoNtO]

I take it you'd say that Audirvana is a better product than Amarra HiFi now? I tested maybe 1.5-2 years ago and remember Audirvana sounding more detailed, whereas Amarra sounded a bit warmer (something welcome, but missed the detail). 

 

[dglow]

I'm saying none of these players should be coloring the audio at all. With the arrival of USB and integer mode I expected every last bit would be shuttled to my DAC fully and transparently. Yet I (and others) discovered this was not the case.
 
Bitperfect, Fidelia, Decibel, Amarra – I've tried these and more. They all output through Core Audio, and hence employ either the system's native SRC (at various quality settings), or an open source (SoX) or third-party (iZotope) alternative.
 
Your observation that Amarra sounds warmer is likely the SRC at play – and Audirvana's greater detail results from the absence of that processing layer.
 
If every player could equally bypass Core Audio, one would expect all of them to sound identical. At which point choosing between them would become solely a question of UI preference. Were that the case, I'm not certain I'd stick with Audirvana, but for now it's the most transparent transport on the Mac.

 

[coli]

   
HQPlayer is the only player I've found that gives a very obvious night and day improvement. With the old JPLAY, I wasn't sure I was hearing a real difference, but with this one, there is no doubt. Unfortunately, the interface is atrocious.

Jplay 6 is just out, definitely try it out again. I got a couple live recordings that are just unbelievable now. It's like a drug, lol.

 

[Music Alchemist]

  Jplay 6 is just out, definitely try it out again. I got a couple live recordings that are just unbelievable now. It's like a drug, lol.

 
Were you able to listen to HQPlayer as well? I have JPLAY 6, but haven't been able to configure the settings to work on my system yet.

 

[coli]

   
Were you able to listen to HQPlayer as well? I have JPLAY 6, but haven't been able to configure the settings to work on my system yet.

Nope, I really need my foobar playlists...

 

[Music Alchemist]

  Nope, I really need my foobar playlists...

 
I highly recommend trying it out just to experience it. Unless JPLAY 6 is a giant leap forward in sound quality compared to the old version, HQPlayer has it beat. But yeah, HQP's interface is really awful. When I use JPLAY, I use JPLAYmini. Some people say it sounds better that way as opposed to integrating it with another player. But right now, I'm just using foobar2000. With everything, I manually select files from folders, since the only software I have all my music organized in is iTunes, which I don't use anymore.

 

[coli]

   
I highly recommend trying it out just to experience it. Unless JPLAY 6 is a giant leap forward in sound quality compared to the old version, HQPlayer has it beat. But yeah, HQP's interface is really awful. When I use JPLAY, I use JPLAYmini. Some people say it sounds better that way as opposed to integrating it with another player. But right now, I'm just using foobar2000. With everything, I manually select files from folders, since the only software I have all my music organized in is iTunes, which I don't use anymore.

K, let me try it. (I got way too much music files, most in rar/zip, so I heavily depend on foobar playlist, I delete stuff from the list very aggressively)