People who find this thread unhelpful may find this helpful. Also this.
My conclusion is that provided you are using asynchronous mode (and you should be if you want high quality audio) then USB cable jitter is irrelevant; any USB compliant cable will do. Further, if you are using some other mode (as many devices do) then USB cable jitter is likely to be negligible when compared with the timing generated by the PC, so you can still use any USB compliant cable. Sync mode is best avoided; adaptive mode may be OK if properly implemented.
I may get shouted at now.
My conclusion is that provided you are using asynchronous mode (and you should be if you want high quality audio) then USB cable jitter is irrelevant; any USB compliant cable will do. Further, if you are using some other mode (as many devices do) then USB cable jitter is likely to be negligible when compared with the timing generated by the PC, so you can still use any USB compliant cable. Sync mode is best avoided; adaptive mode may be OK if properly implemented.
I may get shouted at now.
USB is everywhere I have never had any problems with the plugs...
Actually the HDMI interface is better than DVI... I use HDMI for my monitors.
Again if the plugs are inserted properly and there is no excessive strain on them there is no problem especially with USB.
OK...
I will admit that asynchronous USB is not necessarily perfect in sound quality, and that paying attention to certain details can improve the sound quality of USB audio interfaces, both on the sending side, and the receiving side. I would not say this is "jitter", perhaps it is just due to noise transmission to the analog side of the DAC circuitry, or ground plane noise affecting the audio clocks, or... Well isolated USB interfaces with separate supplies help when well implemented.
Attention to details in these interfaces, which many respected engineers feel do not matter, often result in improved sound quality.
What has recently come to my attention, is that using a very low phase noise clock for the USB processor even improves sound quality, yes I know this sounds totally ridiculous, and I have no technical explanation to offer.
Sorry, just some of my experiences reported here.
I will admit that asynchronous USB is not necessarily perfect in sound quality, and that paying attention to certain details can improve the sound quality of USB audio interfaces, both on the sending side, and the receiving side. I would not say this is "jitter", perhaps it is just due to noise transmission to the analog side of the DAC circuitry, or ground plane noise affecting the audio clocks, or... Well isolated USB interfaces with separate supplies help when well implemented.
Attention to details in these interfaces, which many respected engineers feel do not matter, often result in improved sound quality.
What has recently come to my attention, is that using a very low phase noise clock for the USB processor even improves sound quality, yes I know this sounds totally ridiculous, and I have no technical explanation to offer.
Sorry, just some of my experiences reported here.
In case it isn't clear to everyone, the USB clock isn't the same clock as the audio DAC clock in the USB dac.....! The USB stream gets the data to the box but doesn't time the edges on which the digital audio turns into analog audio, there's an internal oscillator in the DAC box for that. Sort of like how it doesn't matter too much how the postman delivers your CDs from Amazon nor how erratic his path is, as long as you get the disk intact. Your CD player doesn't respond to the postman's meandering and coffee breaks provided he gets the disk to you both intact and in time for you to play it.
That said, regular USB audio isn't guaranteed bit-perfect. It is "isochronous" which means guaranteed stream rate and delivery time *IF* the signal gets delivered -- but no guarantee that it will get delivered at all! In particular, in Windows (if you use that as your audio source), gaps and interruptions are not all that rare, audio isn't the PC system's highest hardware priority and it is possible for things to get too busy or distracted for the processors to keep up with all needed packets when needed. And if the data gets there in error, the DAC will usually recognize it (and respond to it in some way, usually via a dropput) but can't go back and get a corrected packet... things just keep chugging along.
(BTW, I've written some USB drivers and has also had to try to keep USB audio converters behaving for use in Windows based audio measurement equipment, so not just parroting something read online).
That said, regular USB audio isn't guaranteed bit-perfect. It is "isochronous" which means guaranteed stream rate and delivery time *IF* the signal gets delivered -- but no guarantee that it will get delivered at all! In particular, in Windows (if you use that as your audio source), gaps and interruptions are not all that rare, audio isn't the PC system's highest hardware priority and it is possible for things to get too busy or distracted for the processors to keep up with all needed packets when needed. And if the data gets there in error, the DAC will usually recognize it (and respond to it in some way, usually via a dropput) but can't go back and get a corrected packet... things just keep chugging along.
(BTW, I've written some USB drivers and has also had to try to keep USB audio converters behaving for use in Windows based audio measurement equipment, so not just parroting something read online).
Yes, clicks, drops or even silences can ocasionaly be heard on audio pcs, with usb and not only with usb. It must belong to some kind PC audio fate, these machines must bevery busy with all kind of other jobs, but in my case it does not bother me any more since i have confessed to my creator that i am not bit perfect either...
Yes...
Certainly I understand the above. This is why it is astonishing to me that the phase noise of the USB clock matters. And by no means do I believe that data corruption is an issue with Async USB interfaces, as you mention, a dropped bit/packet will result in a dropout, and not a change in sound quality. I suspect that the (perhaps we can call them "second order") sound quality changes are due to noise issues, coupling either to the audio clock or to the analog side of the DAC.
Certainly I understand the above. This is why it is astonishing to me that the phase noise of the USB clock matters. And by no means do I believe that data corruption is an issue with Async USB interfaces, as you mention, a dropped bit/packet will result in a dropout, and not a change in sound quality. I suspect that the (perhaps we can call them "second order") sound quality changes are due to noise issues, coupling either to the audio clock or to the analog side of the DAC.
More likely due to the fact that the PC and the DAC system share a ground in the cable and another in their power supplies (which might be on the same cable as well). a.k.a, ground loop. Noise on the cable ground can appear on the audio ground path out of the DAC and end up in series with the analog audio. A good argument for optical.
Last edited:
I have only once had trouble with USB drop outs.
I noticed when internet explorer got new pages the audio would sometimes drop out.
However I recently had to reinstall windows 10 and the problem went away.
I don't know if USB uses it but I have used error checking in RS232 serial data before.
If the data isn't received correctly then the serial port software asks for a resend.
I can only go by my own experience and that is USB is incredibly reliable.
My USB DAC sounds very good all the time.
I noticed when internet explorer got new pages the audio would sometimes drop out.
However I recently had to reinstall windows 10 and the problem went away.
I don't know if USB uses it but I have used error checking in RS232 serial data before.
If the data isn't received correctly then the serial port software asks for a resend.
I can only go by my own experience and that is USB is incredibly reliable.
My USB DAC sounds very good all the time.
Not in isochronous (the usual) USB audio. There are no resends, when a packet gets missed, that ship has sailed. The error is detected, but not much can be done other than avoid sending a big spike out the analog audio connections.
Not like when writing data to a USB flashdrive, where it will retry till it gets it right. The flashdrive uses bulk transfer, which is guaranteed delivery but not guaranteed rate or time.
CDs delivered via UPS sound better to me than ones delivered by USPS. Swear to God!, true story!
Wonder how many here are old enough to remember when motherboard integrated audio first became a thing.
No hardware acceleration or buffering, audio playback took significant CPU %, and if you try playing midi it will never be played at a constant speed.
And of course drop outs. It was normal at that time.
Years later, when onboard sound finally had the same capability as a Sound Blaster AWE64, people moved to PCM2702 USB just to suffer the same thing again. I opted to use Audigy SPDIF out instead.
No hardware acceleration or buffering, audio playback took significant CPU %, and if you try playing midi it will never be played at a constant speed.
And of course drop outs. It was normal at that time.
Years later, when onboard sound finally had the same capability as a Sound Blaster AWE64, people moved to PCM2702 USB just to suffer the same thing again. I opted to use Audigy SPDIF out instead.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.