Isochronous or Asynchronous?
Printed From: Graham Slee Hifi System Components
Category: Digital Audio
Forum Name: CD, DVD Audio, DACs, ADCs and Digitizing
Forum Description: The existing (and obsolete?) digital formats
URL: https://www.hifisystemcomponents.com/forum/forum_posts.asp?TID=1887
Printed Date: 27 Mar 2026 at 4:04am
Software Version: Web Wiz Forums 12.01 - http://www.webwizforums.com
Topic: Isochronous or Asynchronous?
Posted By: Graham Slee
Subject: Isochronous or Asynchronous?
Date Posted: 03 Jun 2013 at 9:39pm
|
Isochronous or Asynchronous DACs? There are four USB transfer modes Isochronous (guaranteed timing) Interrupt Control Bulk Three of which are used in USB audio: Isochronous, Interrupt and Bulk - the latter two are often refered to as Asynchronous. The Isochronous transfer mode was developed for time sensitive information which music is, and is intended for audio and video. It has guaranteed timing (important for playback). Asynchronous which can be of the Interrupt or Bulk transfer mode (there is no such named transfer mode as Asynchronous) has guaranteed data (important for recording). So why is Asynchronous supposed to be so good as a DAC which is replay, and Isochronous torn to shreds by what seems to be all and sundry? At this moment in time a number of manufacturers have developed their own USB playback streaming solutions based around Application Specific Integrated Circuits (ASICs) which are basically a computer and memory on a chip. These are required to make a USB to serial UART interface work. UART means Universal Asynchronous Receiver/Transmitter. All this is needed to extract the audio data which is then presented to the digital to analogue converter (DAC) for playback. It is rather a bulky way of converting USB audio to analogue audio and because of its many computational proceses is power hungry often requiring exclusive use of the computer's USB power supply. Often it can cause other programs running on the same computer or the operating system itself to freeze. These streaming solutions have to use Interrupt or Bulk USB transfer modes because the UART is Asynchronous (Universal Asynchronous Receiver/Transmitter). They cannot use the Isochronous USB transfer mode intended for audio streaming. They are not guaranteed timing - they are guaranteed data. Asynchronous (whether Interrupt or Bulk transfer mode) is always advertised as being better than the audio (and video) specific Isochronous transfer mode. As all these "home grown" devices are Asynchronous those who market them would obviously have to "shout louder" and rubbish the Isochronous mode. One way of rubbishing Isochronous is to claim that the clock signal from the computer will not be properly recovered by an Isochronous USB DAC, and this will result in high levels of jitter... quote John Atkinson: "this process is inherently prone to jitter because control of the dataflow is handled by the host PC". In subjective listening our customers who are quite aware of what jitter sounds like report the complete opposite - our USB DAC (the Bitzie) is unashamedly Isochronous! I find the article from which the above quote was taken to be suspect since elsewhere it is also written (quote): "the DAC simply truncates the 24-bit data to 16, which introduces distortion and quantizing errors" - this is not true. The truth is that truncating 24 bit data which is 20 bits of audio data (the other four being non-audio data) removes the four least significant bits and the only difference there is the dynamic range (or noise floor). For an explanation on this see: http://www.xiph.org/video/vid2.shtml - http://www.xiph.org/video/vid2.shtml chapter selection "bit depth". The only explanation I can offer for there being such (often heated) rows between the two transfer modes is there is much money at stake in digital audio, and in time honored fashion politics has entered into a subject it knows nothing about, to control the money flow in the direction the politics dictates. ------------- That none should be able to park up and enjoy the view without a smartphone and the knowledge in how to use apps |
Replies:
Posted By: Graham Slee
Date Posted: 04 Jun 2013 at 1:23am
|
In my article above I discussed Isochronous transfer mode used for audio. I mentioned Mr Atkinson's comment about Isochronous being prone to jitter because control of the data flow is handled by the host computer. Is this actually the case? Do we know another type of digital audio transmission where the data flow is handled by a host? Yes we do. It's S/PDIF. How does this gain acceptance as a high end method of audio transmission? Would it be because the embedded clock which syncs the data is detected and locked onto by a thing called a phase locked loop (PLL)? The much acclaimed Wolfson WM8804 receiver depends on its phase locked loop (which you can see in the block diagram in its data sheet) stated as being: "Advanced jitter attenuating PLL with low intrinsic period jitter of 50 ps RMS. " All the disparaging articles I've read regarding Isochronous transfer mode have the reader believing that all such equipped DACs rely purely on a perfect clock signal being received from the host and no more... How misleading is that? All such articles completely miss out the use of a PLL in the receiver - how convenient? There is a missing term however, that is "adaptive". Once you search for adaptive isochronous you may dig up the odd slightly more positive article such as this one: http://www.soundstagehifi.com/index.php/feature-articles/sound-reasoning/324-understanding-digital-music-usb-audio-transfer-methods - http://www.soundstagehifi.com/index.php/feature-articles/sound-reasoning/324-understanding-digital-music-usb-audio-transfer-methods I quote from the article: " Over a number of bursts, the sample rate will average out to the required 44.1 kilohertz, 48 kilohertz, or higher. Since the DAC has no control over the rate at which the computer sends the data, its clock is slaved to the timing of the buffer’s output. A phase-locked loop (PLL) keeps the DAC’s clock synchronized to the computer’s clock. The implementation of the PLL determines to what extent the DAC is subject to jitter on the incoming datastream." Now that sounds very much the same to me as how S/PDIF is handled. The key to understanding a phase locked loop is to realise it will not lock instantaneously. It requires a number of clock periods to lock on to the host's clock - each period in USB audio being 1 milli-second (0.001S) your ears won't detect the start-up. Referenced to a good crystal oscillator the PLL will soon lock and the quality of the lock decides the jitter content. If somebody measured the jitter within the initial lock period - before the PLL locks which digital test gear can capture it is bound to be poor. Then again the lock quality and resulting jitter doesn't simply rely on how good the crystal is or how good the PLL is, it also relies on the circuit layout and power quality - bad grounding can wreck an otherwise good phase locked loop or crystal oscillator, or both. In Mr Atkinson's article one particular USB DAC was measured. Unfortunately Mr Atkinson seems to infer all Isochronous DACs as being bad from this one test device. And when somebody as "important" in the industry as the editor of America's most influential hi-fi magazine says something like that it gives an excuse to the torch and pitchfork brigade to go all out and damn anybody for going the Isochronous route. I said the Bitzie is unashamedly Isochronous. I will qualify that statement by saying it is Adaptive Isochronous - it features an excellent crystal oscillator and phase locked loop, but more than that, the PCB is tracked properly and its PLL is, along with the rest of the analogue circuit, supplied by a high quality on-board voltage regulator. The moral of all this is stop simply believing the figures and misinformation and give the designer - a good designer - some credit for knowing his job. ------------- That none should be able to park up and enjoy the view without a smartphone and the knowledge in how to use apps |
Posted By: Humboldt
Date Posted: 04 Jun 2013 at 3:30pm
|
Very interesting posts. There is an web-zine called The Audio Critic http://www.theaudiocritic.com/cwo/Back_Issues/ I don´t think it is active anymore, but the back issues are available on their website. Anyway, The Audio Critic has an very critical and skeptical approach to the audio industry and all myths feed by the industry and all the reviewers writing for HiFi magazines. In issue 24 they actually have put up a list of what they think is bad and god guys in the industry. When reading Grahams post I remembered they wrote about John Atkinson as one of the bad guys in the industry. At page 14 (pdf15) for those who wants to read. http://www.theaudiocritic.com/back_issues/The_Audio_Critic_24_r.pdf |
Posted By: musicdude
Date Posted: 08 Jun 2013 at 4:52am
|
Ahhh, Mr. Atkinson is full of it ..... Maybe he should get some advise from a honest manufacturer before he writes an article were he is not familiar with the subject. I'm always amazed how much knowledge you have and how you take the time to educate us! Hats off Graham!  ------------- Andy ProJect Xtension 10 with Clearaudio Charisma V2, GA2 SE, Majestic DAC, CuSat50, Yamaha CD-S1000, Yamaha M-80, Revel F-52. |
Posted By: musicdude
Date Posted: 08 Jun 2013 at 6:42am
Very interesting read. Thanks for that! ------------- Andy ProJect Xtension 10 with Clearaudio Charisma V2, GA2 SE, Majestic DAC, CuSat50, Yamaha CD-S1000, Yamaha M-80, Revel F-52. |
Humboldt wrote:Posted By: Fatmangolf
Date Posted: 08 Jun 2013 at 8:35am
|
I would like to add my thanks for this article Graham. ------------- Jon Open mind and ears whilst owning GSP Genera, Accession M, Accession MC, Elevator EXP, Solo ULDE, Proprius amps, Cusat50 cables, Lautus digital cable, Spatia cables and links, and a Majestic DAC. |
Posted By: morris_minor
Date Posted: 08 Jun 2013 at 9:16am
That Audio Critic issue provided an evenings good read last night! ( - with headphones on to drown out The Voice on TV  .) Definitely an "audio objectivist" publication!While I felt very comfortable with most of the opinions (forgiving the lack of faith in LP's future, back in 1997), the dismissal of "cables" as a key component was somewhat wide of my experience with the Spatia and CuSat50. So this begs the question "what else" might not be all it seems. The comparison of audio technology to video/photographic technology, however, was an interesting one. Using our visual sense to determine quality is taken for granted. You can point to something wrong; when you see it, it's hard to refute. However hearing a problem, or hearing differences in sound reproduction is easily dismissed (by some). "What?! You can't hear how my new room tuning brick has totally transformed my hifi? You need your ears testing....". Ok, a bit of an exaggeration, but this is how a great deal of hifi seems to be sold (you just buy the "brick" to prove you have golden ears!). ------------- Bob Majestic DAC/pre-amp Accession MC/Enigma, Accession MM, Reflex M, Elevator EXP, Era Gold V Solo ULDE, Novo, Lautus USB and digital, Libran balanced, CuSat50 2 x Proprius + Spatia/Spatia Links |
Posted By: Fatmangolf
Date Posted: 08 Jun 2013 at 10:57am
|
Agreed! It's been said here before that "try before you buy" or "full refund if not satisfied" shows the supplier believes in their product enough to let you make your own mind up. GSP's home trial scheme is a great example of this. ------------- Jon Open mind and ears whilst owning GSP Genera, Accession M, Accession MC, Elevator EXP, Solo ULDE, Proprius amps, Cusat50 cables, Lautus digital cable, Spatia cables and links, and a Majestic DAC. |