diyAudio (
-   Digital Line Level (
-   -   Does USB->I2S make sense ? (

percy 1st February 2009 08:20 PM

Does USB->I2S make sense ?
In USB audio, I know of no way of 'slaving' the host computer's transmission rate other than using the Asynchronous mode of feedback/synchronization in the Isochronous mode of operation.

As of now none of the mass market (the ubiquitous PCM27xxx) usb chips offer that kind of synchronization. I believe they all just do the Adaptive mode of synchronization, which is nothing but adjusting the DACs clock to match the PC's transmission rate. Basically the DAC follows the PC (insert chilling shiver here).

So what's the point in doing an I2S conversion from USB ? There no reverse feedback possible anyway!

In fact in those terms it would be just like spdif. What would one acheive by 'converting' spdif to I2S ?

analog_sa 1st February 2009 08:33 PM

Re: Does USB->I2S make sense ?

Originally posted by percy
So what's the point in doing an I2S conversion from USB ?
As opposed to converting to what? Spdif? This will only make jitter worse. I2S is a native format for most dac chips.

percy 2nd February 2009 12:03 AM

ok I should have been more clear - when I said I2S I was referring to the use of I2S for the purpose of slaving the source as well.

I agree that I2S would still be better than spdif any day, but when converting from USB all that it will do is provide a better interface, but still no advantage in terms of using your own clock.

sangram 2nd February 2009 04:20 AM

I see three ways of looking at it:

Your DAC may not accept SPDIF input, and you may need a I2S input for it (TDA1543, PCM1792, etc).

If it does, converting USB (1.1) to I2S should not provide any benefit at all, over USB to SPDIF.

If it offers both, there may be differences in the way it handles SPDIF versus I2S, and then it may or may not make sense to use USB->I2S.

You cannot reclock a PC source with a USB 1.1 connection, not with a mass-market product (if at all). Since it is a mass-market product, reclocking is probably an extravagance, anyway. However, two-way communication does happen on USB 1.1 - since thumb drives can do it, there may be a way to write some kind of a driver for a USB controller. It's still a lot of work.

IMO nothing beats Firewire for getting a signal out of the PC today, as long as you have drivers and the right device. Zero latency, external clock and everything is outside the PC environment.

audiosteve 2nd February 2009 06:18 AM

Is there any such thing as a sound card with a word clock input?? If so could you have a converter with Is2 on your end that feeds its word clock back to the sound card and forces it to transmit (whether USB or SPDIF) in sync with your local clock? Just thinking out loud here. I'm not sure if such a card exists for consumer PC's but, in the pro audio world, digital audio gear that syncs to an external clock is pretty common.

sangram 2nd February 2009 06:26 AM

Plenty soundcards have a word clock input, and some even have word clock outputs.

The E-mu1212m, for example, interfaces with an add-on card with BNC connectors for word clock sync.

But then the SPDIF output of that card is pretty good.

Soundcards normally don't have USB connectors for passthrough, only SPDIF. Normally the controllers on the cards determine the clock of the DACs on them so even though the chips are separate it's a closed system, difficult to get in to.

lucpes 2nd February 2009 07:14 AM

Hi there,

I'm using a M-audio transit to get I2s from a TAS1020A chip (USB) and it sound quite good into a TDA1541A. Much better than an EMU 1212m internal card. Get one of those, follow the datasheets and you get the pinout.

Make sure that the USB card has the proper grounding with the DAC board, other than that, happy listening.



percy 2nd February 2009 04:55 PM

I have come across pci soundcards with a word clock input. But I haven't found any 'external' (or atleast sanely priced external) soundcards that have word clock input.

Even having just a word clock input isn't enough, unless you want to stick with the DACs and opamps of the soundcard itself. But to do what we mostly do around here, it would only make sense for a soundcard to have a word clock input AND an I2S output(but beware, read below). Or atleast bidirectional I2S that will allow you to slave the soundcard to an external clock. But I suppose if a card has a word clock input it most likely has an I2S output as well.

It's possible that some firewire or ethernet soundcards could exist that would have a word clock input but I would be very surprised to see a USB soundcard. And if you find a USB soundcard you might want to be careful how in the world would they do that because of what I said earlier about the whole Asynchronous synchronization/feedback (or)deal in USB.
I don't know much in detail about Firewire and Ethernet but it sounds like it has the feedback built into the nature of the protocol.

Key 2nd February 2009 07:35 PM

There are firewire with BNC word clock inputs. My Mackie 400f has them and I am pretty sure the focusrite and presonus cards had them too when I was looking.

I only know of the Black Lion external clock that is intended to be used to improve jitter or clock performance.

peufeu 3rd February 2009 05:56 PM

It's part of the USB audio standard. Unless you use asynchronous isochronous mode (which needs special hardware ie not PCM27xx) your clock will be derived from USB ticks.

USB audio standard 2.0 should correct this.

I'm building a FPGA based USB soundcard which will do it the right way (ie clock in the DAC)...

All times are GMT. The time now is 05:13 PM.

Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 18.75%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Copyright ©1999-2018 diyAudio