• staki
  • ADAT user
  • Offline
  • Registered: 2009-05-11
  • Posts: 13

Topic: HDSP 9632 and wordclock

Hi all,

I'm using a HDSP 9632 card (stereo analog outputs) on a pc for high quality playback on my hifi system, and I would like to know if adding a high precision external clock (like Antelope isochrone or others) would improve the sound quality.

If I read correctly, the internal jitter on the 9632 card is about 800 picoseconds (!!). Is that true ? If yes, this is far too much jitter compared to good external D/A converters which often have less than a few tens of picoseconds !

What do you think about it ?

Regards, staki

  • vinark
  • DAW Mastermind
  • Offline
  • Registered: 2009-06-05
  • Posts: 2,658

Re: HDSP 9632 and wordclock

AFAIK 800pS is very good (0,8nS). Mostly jitter is given in nanoseconds.
But of course there are better DA converters then a ~500€ sound card that also performs lots of other functions.
And no, clocking externally will never decrease jitter, no matter how good the wordclock (see Sound on Sound article on this subject)!
External clocking is about synchronising multiple digital devices to one clock, not about improving audio quality.
And again AFAIK there has only been one device which audio quality improved with external clocking (one of the first digidesign interfaces) and that one had a really bad clock (maybe even design flaw), this caused the myth for better audio by external clocking.

Vincent, Amsterdam
https://soundcloud.com/thesecretworld
BFpro fs, 2X HDSP9652 ADI-8AE, 2X HDSP9632
  • staki
  • ADAT user
  • Offline
  • Registered: 2009-05-11
  • Posts: 13

Re: HDSP 9632 and wordclock

Thank you for your answer vinark.

In fact, I find the sound of the 9632 very good. But I just wanted to know if it was possible to get a still better sound with this card.....

I am sorry to insist, but if externally clocking never decreases jitter, why do some companies (DCS, Antelope, Esoteric, and some others) sell very expensive clocks, even atomic clocks, and affirm that these devices increase the sound quality ?
Is that all only snake oil ?

And what about decreasing the jitter internally, by using better quality oscillators for example ?

  • vinark
  • DAW Mastermind
  • Offline
  • Registered: 2009-06-05
  • Posts: 2,658

Re: HDSP 9632 and wordclock

Read the article here http://www.soundonsound.com/sos/jun10/a … clocks.htm
And let me know what YOU think.

I have a question too. At what level does jitter become inaudible/insignificant. If you could have a zero jitter system and you could dial in jitter, at what level would you perceive degradation? Would going from 800pS to 400pS still be audible? I have no idea...Anyone?

Replacing the crystal (if even possible), I wouldn´t try it without the possibility to measure jitter before and after. Even the slightest electronic mismatch would increase jitter enormously, even if it was a better crystal and we might then even persuade ourself that the difference due to the increased jitter was an improvement.
Reversed snake oil HeadScratch

Vincent, Amsterdam
https://soundcloud.com/thesecretworld
BFpro fs, 2X HDSP9652 ADI-8AE, 2X HDSP9632

5 Reply by MC

  • MC
  • MC
  • Administrator
  • Offline
  • From: Germany/Thailand
  • Registered: 2007-04-06
  • Posts: 24,877

Re: HDSP 9632 and wordclock

staki wrote:

If I read correctly, the internal jitter on the 9632 card is about 800 picoseconds (!!). Is that true ? If yes, this is far too much jitter compared to good external D/A converters which often have less than a few tens of picoseconds !

You confuse real world clock jitter (which you can measure with any DSO) with the sample jitter, which is measured as change in audio via audio spectrum analyzers, and then re-calculated into a clock jitter like value. The ADC/DAC provides additional jitter damping, different per model, and the method of recalculation into a clock like jitter value is questionable. Therefore we only use real clock jitter values, others often use the much lower (nicer to read) sample jitter values.

Time machine on:
http://www.rme-audio.de/old/english/techinfo/adi8.htm

Scroll down to 'Jitter'.

Regards
Matthias Carstens
RME

MC's Website

  • staki
  • ADAT user
  • Offline
  • Registered: 2009-05-11
  • Posts: 13

Re: HDSP 9632 and wordclock

@ vinark and MC : thank you for opening my eyes. For a non specialist, it is always difficult to find REAL technical information, and not only marketing stuff. Thanks to your links, I believe that I start to better understand what is jitter.

Many thanks, regards
staki

7 Reply by +4dBu (edited by +4dBu 2011-06-01 14:58:44)

  • +4dBu
  • New member
  • Offline
  • From: France
  • Registered: 2010-09-28
  • Posts: 7

Re: HDSP 9632 and wordclock

staki wrote:

Hi all,

I'm using a HDSP 9632 card (stereo analog outputs) on a pc for high quality playback on my hifi system, and I would like to know if adding a high precision external clock (like Antelope isochrone or others) would improve the sound quality.

If I read correctly, the internal jitter on the 9632 card is about 800 picoseconds (!!). Is that true ? If yes, this is far too much jitter compared to good external D/A converters which often have less than a few tens of picoseconds !

What do you think about it ?

Regards, staki

Hi,

I agree with all that has been said.

If you need better sound quality from your HDSP 9632 card then the only way you are likely to obtain this is to run its AES/EBU XLR digital output to a very high quality external DA converter/clock with AES/EBU XLR input. This means finding one that is "better sounding" than the HDSP 9632's own converter/clock which is not going to be easy or come cheap! It is obviously possible to find one though. The external converter clock (master) should be connected either to the HDSP 9632 clock (slave), via the : Word Clock Module (WCM) or you could just set the HDSP 9632 to "slave" mode in software. In general, the converter being used for final signal, gives better sound quality when using its own clock.

I have no idea whether the WCM connection would give improved sound quality over the software option, but it would cost you 90€ to find out!

This is one way to upgrade your system in the future.

Lots of listening tests in perspective... :-)

[EDIT] BTW the 3.24 driver update made a dramatic improvement to the sound of the HDSP 9632 when connected (before and after update) to my hifi system. Anyone else found this to be the case? Improved kernel streaming?

8 Reply by staki (edited by staki 2011-06-01 23:56:42)

  • staki
  • ADAT user
  • Offline
  • Registered: 2009-05-11
  • Posts: 13

Re: HDSP 9632 and wordclock

@ +4dBu: I understand what you mean, but such an external device would cost at least ten times the cost of the HDSP 9632 ! And even so, it is not sure that it will give a better sound than the HDSP alone !! The integrated Analog Devices 1852 DACs are SOOOO good that it will be difficult to beat them !

For now I prefer to explore an other way, I will try to have a less noisy environment, for example buying (or building) a very good linear power supply for the motherboard instead of the normal switching power supply.
I don't know if it will make any difference, but it is a good listening test in perspective......  ;-)

  • +4dBu
  • New member
  • Offline
  • From: France
  • Registered: 2010-09-28
  • Posts: 7

Re: HDSP 9632 and wordclock

staki wrote:

@ +4dBu: I understand what you mean, but such an external device would cost at least ten times the cost of the HDSP 9632 !

Yes, I think it probably would :-)

The HDSP 9632 does sound good out of the box! Computer Audiophile really likes this card.

http://www.computeraudiophile.com/conte … dware-List

Searching for a noise free environment is cheaper but only worthwhile if you are building a dedicated music server (the HDSP 9632 is ideal for this: Windows XP, an old Pentium 4 and 1Go RAM is more than enough). Go for a single rail passive power supply, a passive CPU cooler and fan-less video card. Keep case fans to an absolute minimum, better if none at all and use an SSD for the operating system/music files. On the motherboard, put the sound card as far away from everything else as is possible.

You can effectively shield the sound card by using an unused PCI bracket, to which you need to screw a thin steel plate that resembles a full size extension card. Be careful it does not touch other components though or you will short circuit something.

If you are using your desktop for other things or are only using the digital output of the 9632 (into an external DAC) then it probably isn't worth the trouble...

Don't let the quest for "audio nirvana" spoil the music cool

10 Reply by staki (edited by staki 2011-06-01 18:33:20)

  • staki
  • ADAT user
  • Offline
  • Registered: 2009-05-11
  • Posts: 13

Re: HDSP 9632 and wordclock

Yes, it is exactly what I want to build: a dedicated music server.

I will take a mini-itx motherboard with Atom CPU (Intel D510MO, or Jetway JNC96FL) with 12 V DC input, so it is easy to build an external linear power supply. No fan at all on the motherboard, and no videocard with fan, the videocard is integrated on the motherboard. And no fan in the case, I use the "natural chimney effect" (mounting the motherboard vertically).

The OS wil be on a small SSD, and the music files on a SATA or external USB (900 Gigabytes of files in FLAC !!!).

For the soundcard I use a PCI risercard, so the soundcard will be outside of the motherboard. And furthermore I can shield it without problems.

The only negative point is that I have to use Windows XP. Hundred of back services which use CPU activity for nothing !! It is possible to tweak it a little, but not as much as Linux !
I wanted to use Linux, but with the 9632 card, under Linux, it is impossible to switch automatically the sample rate on the fly when listening to files with different sample rates, if you want to make bitperfect playback (with MPD for example).
Sad that the RME team doesn't want to deal with Linux !!

Anyway, with Windows and Foobar 2000 (with Kernel Streaming or ASIO) I hope the sound quality will be good enough..........