Mac Mini

[planet10]

Originally posted by dukeoddorkdom
while I read it elsewhere, firewire is more jitter prone.

where?

This post is informative -- the whole thread is worthwhile.

http://www.diyaudio.com/forums/showt...872#post984872

dave

[error401]

The short answer is that Firewire is probably the best (read the thread though, it's interesting). This is because the DAC clock is the master in all cases - the transport induces no jitter at all because it doesn't carry the clock. Data is sent to the audio interface when it requests it; the DAC is fully in control of the data stream. In the USB Audio Class and S/PDIF, data is sent at a constant rate and the clock is recovered from the data timing, which is where jitter becomes important.

Since the USB timing is generated indirectly from the main CPU clock (where frequency stability is not critical), and the USB frames themselves pass through a complicated path before they're transmitted, there's a ton of potential for jitter there. USB can be pressed into service to work as Firewire does, and most professional interfaces do, but not using the generic USB audio driver included with your OS.

I'm quite sure that USB doesn't guarantee data integrity (without special protocols anyway), and I don't think Firewire does either, though it may have a CRC field, I can't recall right now. It's moot for audio anyway though because there's no time for retransmission in most cases. You either ignore the corrupt sample or play it (though it would arguably be better to just replay the previous sample than a corrupt one, it's not a huge difference).

I would say Firewire = USB (DAC master)> S/PDIF > USB Audio Class.

[phofman]

Originally posted by error401

...Since the USB timing is generated indirectly from the main CPU clock...

Hi,

I have been looking for the actual implementation of USB 1ms frame clock, and so far have been able to uncover only http://www.diyaudio.com/forums/showt...64#post1690264 . This specific implementation does not use SW-driven clock.

[error401]

Originally posted by phofman


Hi,

I have been looking for the actual implementation of USB 1ms frame clock, and so far have been able to uncover only http://www.diyaudio.com/forums/showt...64#post1690264 . This specific implementation does not use SW-driven clock.

Interesting discovery, I never really thought to look into it too deeply, but obviously it's not totally software driven or it'd never really work well. I had assumed it was driven by a CPU interrupt, which is only sort of software driven (I never said SW!). Still, it's generated from the same source as the CPU clock (and almost every other clock in the PC), and with USB's low timing requirements it's probably not treated particularly critically in the USB section of the ICH. Though you're right, I don't see how this is any worse than using the S/PDIF output which is clocked from the same source - at a guess I'd think low jitter recovery is probably considerably more difficult to achieve than in S/PDIF, however.

[phofman]

True, the USB controller clock is very likely coming from the same clock infrastructure as most other clocks in the PC. On the other hand, the diy-acclaimed beagleboard has its I2S clock generated in the same manner.

BTW, the USB audio chain can be pretty resilient, see http://www.diyaudio.com/forums/showt...44#post1719044

[Zoran]

Try FireWhave from Griffin Tech.
about 30USD...
sniffed that could be of help and odrdered few months ago...
I was amased when I opened (read crashed...)
but without damaging the pcb...
*
oxford FW chip with I2S
wolfson dac
*
And next FW product is LaCie firewire speakers, also have ozford FW chip to interface
digital from firewire...

[Zoran]

also mini has hidden another firewire maybe even I2S bus at the connector inside...
google for I2S or FW for mac or something, I loose the link... sorry

[ionomolo]

Originally posted by planet10
The best way is FireWire. As to USB or Toslink, it will depend more on the DAC. One of the (few) Asyncronous USB DACs would likely work best, but you are well into the price range of FireWire DACs.

(BTW, it is Mac Mini -- i've updated the thread title for you)

This is the one i'm looking at getting next

http://www.rolandus.com/products/pro...1&ParentId=114



I have 1 of the cheaper USB ones (cost $75 at swap-n-shop) and it is quite decent, but still doesn't trump our better CD players, and is limited to 96k one way. The Firewire will allow simultaneous in/out at 192k, which makes digitizing records worthwhile, and will allow the playback of scarce yet, but available hi-rez downloads.

dave

I'm interested in the pro stuff as it is remarkable for being both high-quality and myth-free.

I have read some bad reviews about m-audio equipment and since there is not a big price differential between them and the roland you linked to i might buy this firewire audio interface (Much cheaper than a benchmark dac-1, which isn't mentioned often here). Is there any way to trade channels and portability for quality?

There are also the PCMXXXX + paper in snake oil caps interfaces that sell for arround $150 but i don't want to go that way.

[BionicSniper]

you can also go crazy and get a sound card that will spit out an aes/ebu signal and then build a dac to take that. Thats what I'm doing

[planet10]

Originally posted by BionicSniper
you can also go crazy and get a sound card that will spit out an aes/ebu signal and then build a dac to take that. Thats what I'm doing

No place to put a sound card in a MacMini... it already has SPDIF out anyway.

dave