Computer audio - what am I missing??
- Thread starter gorwig
- Start date
sq225917
Bit of this, bit of that
Jitter, with reference to digital audio, is errors in the timing of successive samples of audio such that a sample arrives either before of after it is expected to arrive. it can be caused by many things, poor eye pattern, impedance mismatching, power supply noise, poor clock stability, buffer under and over runs, all sorts of stuff.
If a sample arrives too soon then the rate of change between it and the previous sample is too high, and if it arrives too late then the rate of change between samples is decreased. This gives rise to frequency response errors in the analogue output.
The above case represents random jitter, correlated jitter is a slightly different case where the signal itself is somehow convolved into the output waveform as frequency response error.
None digital audio engineers may call whatever they like jitter, but the above description is pretty much as Julian Dunn, he who discovered jitter, described it in his research paper.
and it's audible, very much so.
If a sample arrives too soon then the rate of change between it and the previous sample is too high, and if it arrives too late then the rate of change between samples is decreased. This gives rise to frequency response errors in the analogue output.
The above case represents random jitter, correlated jitter is a slightly different case where the signal itself is somehow convolved into the output waveform as frequency response error.
None digital audio engineers may call whatever they like jitter, but the above description is pretty much as Julian Dunn, he who discovered jitter, described it in his research paper.
and it's audible, very much so.
hughjampton
pfm Member
Jitter?
Elephantears
Trunkated Aesthete
I didn't know that 'convolved' was a word, so thanks for this. I'm determined to use it in something I write today.
Let the lessons continue.
If you'd like to be a jitter bug,
First thing you must do is get a jug,
Put whiskey, wine and gin within,
And shake it all up and then begin.
Grab a cup and start to toss,
You are drinking jitter sauce!
Don't you worry, you just mug,
And then you'll be a jitter bug!
This is actually MUCH more important than it sounds. I was amazed at the difference it made when my music was being resampled by my PC to 48k before going into my Lavry DAC and ATC speakers. A setting was changed in the PC without me realising it and I couldn't understand why the music was so flat. Then after 5 minutes I took a closer look at the front panel display of the Lavry and realised that it was showing a lock to a sample rate of 48k rather than 44.1k...
Andrew
sonddek
Trade: SUPATRAC
This would be my first guess as to what is wrong too. The whole point of having a good DAC is that it does the upsampling in an intelligent way. If you let your Mac upsample the 44.1 source data to a 24/196 data stream you may as well not have bothered with the fancy DAC because any contribution its logic stage might make is well beyond the audible spectrum.
So try setting the Mac's TOSLink output to 44.1kHz/16bit using "/Applications/Utilities/Audio Midi Setup". Then if your DAC's any good it will clock and upsample the data stream to a higher rez in an ear-friendly way. That's what it's for.
P.S. don't use USB, and don't bother with all those Terminal tweaks and hi-fi hocus pocus: your hard drive does not suffer from acoustic feedback, and 44.1/16 data output will not tax any recent Mac.
Cereal Killer
432
Have a read of the wiki pages benchmark media have written about how to set up a MAC or PC for itunes and computer audio settings. It's very informative and states why leaving the MIDI output settings as high as possible has no issues. The MIDI driver doesn't up sample IIRC.
sonddek
Trade: SUPATRAC
OK - thanks - I stand corrected. For some reason I recall seeing my DAC switch on the 96k light when I set "Audio Midi Settings" that way, but perhaps I got confused by using hi-rez source material. I will verify.
'Lo.. Could you please kindly point me in the right direction for this article?
I can't find it
Cheers.
Purité Audio
Trade: Purite Audio
I find optical the least satisfactory of the Macs outputs , async firewire or Async USB transfer protocols are technically the best way of transferring timed data.
Keith.
SteveS1
I heard that, pardon?
+1. Optical can produce good results, but async Firewire/USB is a very useful improvement on that. There should be no problem outperforming CD replay even among highly regarded top end players. A properly set up PC/Mac and a decent DAC represents hugely better value than some of those players.
Hi Spacey, nice to hear from you again, even though you have abandoned the One True Path that is ATC
You are slightly wrong on this one. First of all it has nothing to do with "midi" or "midi drivers", it's just that the mac dialog box happens to have two tabs, one of which is "audio" and one of which is "midi". We're talking audio here even though people across the internet who don't quite know what they are talking about like to refer to it as "audio midi" (/rant).
Secondly, as the wiki correctly states, by setting the audio in OSX up for the highest sample rate that your DAC will support (eg 96k) rather than the actual sample rate of the digital audio you are playing, you are telling OSX to resample everything that isn't actually at 96k (or whatever you've set there). This was, as the wiki states, disastrous in older versions of itunes. The wiki now says that the resampling abilities of iTunes have greatly improved from version 7, so that it is not harmful or rather "will not add significant distortion to the audio". Not exactly a ringing endorsement...
I expect that iTunes is better than the ancient Creative Audio resampling on my PC so perhaps the effect is less marked but it is still a sub-optimal way of going about things. I will have a listen to see how it sounds when I get a moment. You should also try a compare and contrast exercise when you get a chance.
The fact is that changing the sample rate in this way to force a conversion will guarantee that you do NOT have bit perfect digital information flow from your source media to your DAC. Doesn't sound like a great idea. Others would agree with me, eg the manufacturers of the DACport:
http://www.centrance.com/products/dacmini/blog/2010/09/28/resampling-exposed/
I've also just found this interesting guide to computer audio published by DCS (the co that makes very very high end CD players and DACs).
http://www.dcsltd.co.uk/assets/dCS_Guide_to_Computer_Audio.pdf
It was updated last month and recommends (on page 20) that when using iTunes and Mac OS the sample rate is set in that Audio dialog box to match the exact sample rate of the tracks to be played. There are also similar instructions for the PC world - all with the goal of maintaining bit perfect audio...
Andrew
Cereal Killer
432
'Lo.. Could you please kindly point me in the right direction for this article?
I can't find it
Cheers.
http://www.benchmarkmedia.com/wiki/index.php/Main_Page
http://www.benchmarkmedia.com/wiki/index.php/ITunes-QuickTime_for_Mac_-_Setup_Guide
http://www.benchmarkmedia.com/wiki/index.php/ITunes-QuickTime_for_Windows_-_Setup_Guide
Cereal Killer
432
Hi Spacey, nice to hear from you again, even though you have abandoned the One True Path that is ATC
You are slightly wrong on this one. First of all it has nothing to do with "midi" or "midi drivers", it's just that the mac dialog box happens to have two tabs, one of which is "audio" and one of which is "midi". We're talking audio here even though people across the internet who don't quite know what they are talking about like to refer to it as "audio midi" (/rant).
Secondly, as the wiki correctly states, by setting the audio in OSX up for the highest sample rate that your DAC will support (eg 96k) rather than the actual sample rate of the digital audio you are playing, you are telling OSX to resample everything that isn't actually at 96k (or whatever you've set there). This was, as the wiki states, disastrous in older versions of itunes. The wiki now says that the resampling abilities of iTunes have greatly improved from version 7, so that it is not harmful or rather "will not add significant distortion to the audio". Not exactly a ringing endorsement...
I expect that iTunes is better than the ancient Creative Audio resampling on my PC so perhaps the effect is less marked but it is still a sub-optimal way of going about things. I will have a listen to see how it sounds when I get a moment. You should also try a compare and contrast exercise when you get a chance.
The fact is that changing the sample rate in this way to force a conversion will guarantee that you do NOT have bit perfect digital information flow from your source media to your DAC. Doesn't sound like a great idea. Others would agree with me, eg the manufacturers of the DACport:
http://www.centrance.com/products/dacmini/blog/2010/09/28/resampling-exposed/
I've also just found this interesting guide to computer audio published by DCS (the co that makes very very high end CD players and DACs).
http://www.dcsltd.co.uk/assets/dCS_Guide_to_Computer_Audio.pdf
It was updated last month and recommends (on page 20) that when using iTunes and Mac OS the sample rate is set in that Audio dialog box to match the exact sample rate of the tracks to be played. There are also similar instructions for the PC world - all with the goal of maintaining bit perfect audio...
Andrew
Hello Andrew, yeah what you say is correct, i just put it all in a shorter sentence, in the subjective world it's much simpler - it has no effect IMO.
The one true path of ATC turned into an autobahn
Nik
pfm Member
Isn't it funny when someone proclaims with huge confidence something so obviously wrong?
Nik
pfm Member
I saw it simplified as the "digital equivalent of W&F". I think that still works.