Head-Fi.org › Forums › Equipment Forums › Sound Science › does SPDIF allow lossless and no processing from computer>
New Posts  All Forums:Forum Nav:

does SPDIF allow lossless and no processing from computer>

post #1 of 17
Thread Starter 
i currently use my computer as my media station.. the problem is i dont want it to "process the sound" i have a balanced amp and a balanced dac that i want to use to process it.. the problem is how do i get the signal to the dac without it being altered or processed by the computer. I was thinkin about using an optical cable but my question is will a sound card or usb external thing with a...
    ...SPDIF optical out process the sound? or digital is digital? since im only using it for this purpose can a cheap digital SPDIF card be bought with no problems? basically do not want the computer to convert anything.... go from computer to dac..
post #2 of 17

Yes, but depending on the computers OS you may need to tweak some settings or install additional software.

Basically you want your audio stream to pass on bit-perfect through the OS. Of course at the same bit depth and sampling rate as the played back audio stream.

post #3 of 17

the bit-perfect issue is a bit yesterday now...the real matter these days is jitter, and there's no question that jitter is very high on toslink. Some say it to be +800ps, when a short coax can be as low as 50ps...it makes the sound far more focused and tighter, mostly because S/PDIF is a terrible to begin w/: http://www.gearslutz.com/board/so-much-gear-so-little-time/172143-spdif-vs-word-clock-question.html

 

"S/PDIF is a horrendously poorly designed interface. This is because it combines the clock and audio coding onto the same signal. The receiver is supposed to recover the clock from this signal as well as extract the audio data. This turns out to be a non-trivial task, and one that almost always leaves the recovered clock contaminated with signal correlated jitter artefacts."

 

The less work your receiver has to do to recover the clock, the higher the SQ.


Edited by leeperry - 8/22/10 at 6:47am
post #4 of 17

Any modern PC soundcard, even many onboard adaptors, will work fine as a transport. Just use WASAPI or equivalent in the software to ensure no interference from the software and you're in business.

 

Jitter is a non issue with any DAC that's not essentially broken.

 

Co-ax is more reliable over long cable lengths and everybody loves big chunky RCA connectors, but otherwise there's not much to optical vs co-ax. Use USB if your DAC supports it as that will bypass all the above issues, real or imaginary.

post #5 of 17
Thread Starter 

my dac supports AES, coxial, optical..

 

the problem is i have a laptop and i need a USB or expresscard soundcard to plug into the dac... (dac has no usb...) what is the best thing i shuld use.. THe dac and the computer are like 3 feet away i have a 3 foot toslink cable ... i bought this thing

 

http://www.newegg.com/Product/Product.aspx?Item=N82E16829270009

 

 

i can still return it but im wondering am i bottlenecking there? Like i said i would like to keep the bottleneck at the dac and keep a pure stream from the computer to the dac (what is WASAPI???)

 

will this toslink work at the short distances.. or should i get a USB/expresscard coaxal thing?? please let me know i really am confused on the route to go..!

post #6 of 17
Thread Starter 

is windows or linux better for that pure audio stream... i have both installed.. which one should i listen to music on? software matters?

 

thanks so much for your help!

 

i was told to use foobar2000 ?

 

what about on linux... is there a reason why  these programs are better?


Edited by ambush276 - 8/22/10 at 9:49am
post #7 of 17

also, most cheap transport use vastly inaccurate clocks...you'll never hear your music at the right pitch: http://www.m2tech.biz/public/pdf/White%20Paper%20on%20hiFace.pdf

post #8 of 17

http://www.m2tech.biz/public/pdf/White%20Paper%20on%20hiFace.pdf - someone doesn't know the 1st thing about PLL - they are integrating feedback loops in the frequency domain - they do not have "2-5%" frequency error once locked up 

 

also its hard to buy crystal oscillators with worse than 50 ppm frequency tolerance - even those with asolute/"perfect" pitch won't notice that


Edited by jcx - 8/24/10 at 10:17am
post #9 of 17

Bit perfect from modern CDs has issues of it's own due to the effect of inter sample peaks.

 

See this article.

 

http://www.gearslutz.com/board/tips-techniques/334385-intersample-peaks.html.

 

It might in fact be best to attenuatte the signal a little, say 3dB.

post #10 of 17

 

Originally Posted by jcx View Post

http://www.m2tech.biz/public/pdf/White%20Paper%20on%20hiFace.pdf - someone doesn't know the 1st thing about PLL - they are integrating feedback loops in the frequency domain - they do not have "2-5%" frequency error once locked up 

 

also its hard to buy crystal oscillators with worse than 50 ppm frequency tolerance - even those with asolute/"perfect" pitch won't notice that

 

Many companies sell 10ppm PLL's tbh: http://www.google.com/search?hl=en&q=14.318+crystal+10ppm

 

And that's what crappy clocking does IRL: http://hifiduino.wordpress.com/2010/04/08/reading-sample-rate/

 

That's in precision mode, in "fast" mode it's way worse: http://hifiduino.blogspot.com/2010/02/musiland-driver-1082-fast-vs-precision.html

 

I don't think your DAC will slow down time to play it at 44.1000000Hz.
 

Originally Posted by RonaldDumsfeld View Post

Bit perfect from modern CDs has issues of it's own due to the effect of inter sample peaks.

 

See this article.

 

http://www.gearslutz.com/board/tips-techniques/334385-intersample-peaks.html.

 

It might in fact be best to attenuatte the signal a little, say 3dB.


Interesting!

post #11 of 17

pitch is a psychoacoustic perceptual phenomena - largely determined by frequency (with frequency ratio being most important and “tuning”, ~= absolute frequency reference, having varied +/- single digit % over centuries of western music: 1% = 10000 ppm )

 

there are experimentally determined thresholds of pitch resolution – neither of the 1st 2 of the sample rate errors you show would give a perceptible pitch change vs a “perfect” clock rate in fast switching – which isn’t happening in typical home listening music playback so what you have at worst is an imperceptible change in tuning with Relative Pitch being presented as recorded - ie musical intervals, timbre, chords, melody are all preserved as they are determined by relative frequency and timing ratios

 

Rossing , “Springer Handbook of Acoustics “ section on pitch perception is online in Google Books preview

fig 17.13, p479 gives peak pitch discrimination of ~600 ppm at ~500 Hz if you know your erbs from your barks ( http://www.ling.su.se/staff/hartmut/bark.htm )

 

If you consider Wikipedia articles “definitive” http://en.wikipedia.org/wiki/Pitch_(music) then 4.3 cent jnd gives 245 ppm frequency ratio difference – even this smaller number is not a huge challenge for xtal oscillators of even moderate quality and properly working plls

 

The larger i2s frequency measurement deviation in the last example could easily come from limited resolution/wrong choice of divisor in the “fast mode” software – much more information is needed to understand what is really being shown

And even then while someone may be able to distinguish a tuning difference with the last example’s sample rate error assuming the tighter 4.3 cent jnd resolution estimate (requiring switching to a reference to make the comparison, except possibly for the <0.1% of the population with "perfect pitch") it would not affect the Relative Pitch of the music


Edited by jcx - 8/25/10 at 2:41pm
post #12 of 17

ok, point taken! but jitter is also a problem on the Musiland...it's been discussed here: http://hifiduino.wordpress.com/2010/03/29/summary-of-musiland-01-mini-mods/

 

and their newer cards use a FPGA w/ far more gates and less DCM to overcome this problem...anyway, the theoritical 5ppm figure on the Hiface is nice, but crystals age around 3/5ppm per year.

 

I've compared the fast/precision modes on the Musiland 01USD and 02US, I'm willing to believe that it's more due to a fluctuating high jitter than to their respective ppm resolution, but the fast mode has always sounded very poor in comparison...most ppl will agree, and that surely didn't sound like placebo to my ears.

 

I like the Firestone Bravo, it's using a 0PPM PLL1707 masterclock and reclocks all the inputs through WM8804(50ps jitter).


Edited by leeperry - 8/25/10 at 3:44pm
post #13 of 17
Quote:
Originally Posted by leeperry View Post

the bit-perfect issue is a bit yesterday now...the real matter these days is jitter, and there's no question that jitter is very high on toslink. Some say it to be +800ps, when a short coax can be as low as 50ps...it makes the sound far more focused and tighter, mostly because S/PDIF is a terrible to begin w/: http://www.gearslutz.com/board/so-much-gear-so-little-time/172143-spdif-vs-word-clock-question.html

 

 

A listening test have been done in France with a jitter generator : http://www.homecinema-fr.com/forum/viewtopic.php?f=1037&t=29915550&p=172887105#p172887105

 

The source was a Marantz CD5400 CD player.

The output was the coaxial one.

A QSC hardware ABX switch was used to compare the Marantz directly feeding a DAC, or feeding it through the jitter generator.

The DACs were a DAC Zero (a low end DAC), a Zhaolu D3, and an iRiver H140 pocket player with an optical input.

Amplifier : Kora explorer 90 SII

Speakers : Triangle Antal 202

 

None of the listeners present could hear any difference below 100 ns of jitter.

Around 100 ns, the Zhaolu DAC produces clicks like a badly worn vinyl.

The zero first show some kind of intermodulation distortion, audible on medium / high frequencies, like female voices or piano, at a jitter level just below the apparition of clicks.

The iRiver player was the most resistant to jitter.

 

One of the listeners tried to ABX the Zhaolu DAC with and without 75 ns of jitter in the digital input, but gave up after three trials.

 

I insist that we are talking nanoseconds here, that is 100 to 1000 times the amount of jitter produced by any digital audio source, usually measured in picoseconds.

 

post #14 of 17

 

Originally Posted by Pio2001 View Post

 

A listening test have been done in France with a jitter generator : http://www.homecinema-fr.com

HCFR is such a great forum, ideal to cheer up [:mistersid3]


Edited by leeperry - 8/25/10 at 7:33pm
post #15 of 17
Quote:
Originally Posted by Pio2001 View Post



 

A listening test have been done in France with a jitter generator : http://www.homecinema-fr.com/forum/viewtopic.php?f=1037&t=29915550&p=172887105#p172887105

 

The source was a Marantz CD5400 CD player.

The output was the coaxial one.

A QSC hardware ABX switch was used to compare the Marantz directly feeding a DAC, or feeding it through the jitter generator.

The DACs were a DAC Zero (a low end DAC), a Zhaolu D3, and an iRiver H140 pocket player with an optical input.

Amplifier : Kora explorer 90 SII

Speakers : Triangle Antal 202

 

None of the listeners present could hear any difference below 100 ns of jitter.

Around 100 ns, the Zhaolu DAC produces clicks like a badly worn vinyl.

The zero first show some kind of intermodulation distortion, audible on medium / high frequencies, like female voices or piano, at a jitter level just below the apparition of clicks.

The iRiver player was the most resistant to jitter.

 

One of the listeners tried to ABX the Zhaolu DAC with and without 75 ns of jitter in the digital input, but gave up after three trials.

 

I insist that we are talking nanoseconds here, that is 100 to 1000 times the amount of jitter produced by any digital audio source, usually measured in picoseconds.

 



Thanks for that. 100ns is a lot of jitter ! - I have only heard of one digital device (the McIntosh music server)  that does worse than 10ns, even the lowly Oppo 970 tops out at just above 4ns.

 

You are aware that jitter-worriers will just say that you only adding jitter to existing jitter and that the 5400 is so high in jitter already that it is not a meaningful test as the damage has already been done

 

Steve (not only can I hear 2ps of jitter but I can hear the differences between different 2ps jitter spectra) Nugent may take you to task on this...

http://www.head-fi.org/forum/thread/407522/external-hd-airport-extreme-mbp-airport-express-esoteric-dac/15#post_5397633

New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Sound Science
Head-Fi.org › Forums › Equipment Forums › Sound Science › does SPDIF allow lossless and no processing from computer>