• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

Can two software players sound different from each other and why?

Sal1950

Major Contributor
The Chicago Crusher
Patreon Donor
Joined
Mar 1, 2016
Messages
6,227
Likes
4,070
Location
Central Fl
#43
While it's certainly possible to configure that way, the defaults for most linux systems don't do that. They usually use some combination of ALSA's 'dmix' and 'plug' plugins and PulseAudio to convert to a sample rate and sample format that the sound device supports, and allow multiple applications to play sound at the same time.
No different than Windows which does things similar. Using a number of players it is simple in the extreme to configure Linux for bit perfect playback. ;)
 
Last edited:

graz_lag

Major Contributor
Joined
Oct 13, 2018
Messages
1,232
Likes
1,265
Location
Le Mans, France
#44
Using a simple Linux system the player (mpd in my case) sends that bytes to ALSA and they go out via the USB. Nothing else is involved and nothing can sound different.

(Oh no that bolding thing is contagious)
Indeed, here the same configuration, the straightest line into the heavens
 

Ron Texas

Major Contributor
Joined
Jun 10, 2018
Messages
2,003
Likes
1,597
Location
Equidistant From Everywhere
#45
The DSP sections of two different players could possibly sound different. That might include conversion of DSD to PCM. Although they might measure differently, I doubt anyone could hear the difference in a fair blind test.
 

mitchco

Senior Member
Joined
May 24, 2016
Messages
307
Likes
826
#47
I appreciate what @Blumlein 88 is asking @maty but hasn't this one been done to death, no?

Aside from @mansr pointing out @Archimago tests, and here is part II, myself and others have compared s/w players, (plus file format containers and resampling algo's), comparing not only digital bitstreams but analog output as well.

JPlay versus JRiver

JPLay versus JRiver redux on HA

JRiver on Windows vs JRiver on Mac

As one can understand based on digital audio engineering principles, bitperfect is well... bitperfect - if the bits are the same then by default, and as the repeatable experiments above show, along with ABX listening tests, there is no audible difference and no measurable difference.

I did try another experiment to figure out how far away from bitperfect one can hear a change in sonic quality. One test involved adding broadband eq and the other reducing bitdepth. Both correlated to about -72 dBFS or about 12 bits of resolution is when I could start hearing a real audible difference. That is about ~1 dB of broad band eq variation and 12 bits when reducing bit depth in JRiver's bitdepth simulator: https://audiophilestyle.com/ca/bits-and-bytes/Fun-With-Digital-Audio-–-Bit-Perfect-Audibility-Testing/ Note the ABX comparisons for all of the tests above were performed at normal listening levels and not jacking the volume to ridiculous levels.

Perhaps Maty can provide us with "one" repeatable experiment so that we can verify his claim...?
 

Veri

Major Contributor
Patreon Donor
Joined
Feb 6, 2018
Messages
2,969
Likes
2,900
#49
What's the point of bitperfect playback anyway? Except treating paranoia of lusers, of course. Resampling can be done transparently with very little CPU, these days.
I use equalizerAPO so I could care less about bitperfect :) I much prefer the flexibility DSP offers, so DirectSound and enhancements it is.
 

Veri

Major Contributor
Patreon Donor
Joined
Feb 6, 2018
Messages
2,969
Likes
2,900
#52

maty

Major Contributor
Joined
Dec 12, 2017
Messages
2,026
Likes
895
Location
Tarragona (Spain)
#53
@mitchco

I have never complained about JRMC, with it I enjoy music. The only one that has given me problems with some tests has always been foobar2000, with JRMC it was always much easier until the difference between the audios was negligible.

Months ago, maybe one year, PMA in diyaudio.com proposed tests with different op-amps. I beat them very easily with JRMC, costing me a little more with foobar2000. When they got complicated, I only got it with certainty with JRMC. In one of the tests I even established a quality ranking of the op-amps and that I sent by message, waiting to make it public once the test was finished, it was so easy!

Those tests were unable to pass with foobar2000 + ABX plugin, but not with JRMC, something that bothered PMA. But such was my experience. And as I knew he would doubt, hence the private message with the ranking.

I tried also with Lacinato, win soft only for ABX and could not pass the same test.

With the embarrassment suffered by the foobar2000 + ASIO4ALL v23 with the cartridge test I do not want to know anything more about it for any type of test, just as a tool. "Una y no más, Santo Tomás".
 

daftcombo

Major Contributor
Patreon Donor
Joined
Feb 5, 2019
Messages
1,415
Likes
1,110
#55
I am not an expert in Windows audio (I got rid of it many years ago ...) as I was never sure the Windows mixer, interpolation, upsamplig etc. were out of the equation ...
Archimago' measurements on Windows 10 upsamplig:
http://archimago.blogspot.com/2015/11/measurements-windows-10-audio-stack.html
Looks a lot like the totaldac output!!

Archi wrote:
But wow... Isn't that cool?! Imagine you saw that exact Windows 10 impulse response printed in the pages of a glossy audiophile magazine for a US$15,000 DAC. I suspect many would be impressed, right? After all, no pre- or post-ringing! (May I suggest someone have a look at the quality of anti-aliasing and frequency response with the emm Labs DAC2X :)
 
Last edited:

maty

Major Contributor
Joined
Dec 12, 2017
Messages
2,026
Likes
895
Location
Tarragona (Spain)
#57

Jim777

Member
Patreon Donor
Joined
May 28, 2019
Messages
78
Likes
71
Location
Greater Boston
#59
What output mode are you using, and on what os version?
Have you reported this to the Devs?
I was using WASAPI, probably the one that supports DSD. But foobar2000 only uses 32bit floating point format internally, so not enough for 32 bit integers (once you get rid of the exponent part). So this behavior is expected, it's a limitation, not a bug. The only time I ever saw the RME pass the 32 bit test was with JRiver.
 
Top Bottom