ef37a
Joined: 29/05/06
Posts: 5622
Loc: northampton uk
|
usb sync
#995287 - 29/06/12 07:36 AM
|
|
|
|
Silly question but every now and then I wonder and forget to ask...
Does a usb
AI generate its own clock and the computer then lock to it or is the interface clocked by
the usb bus in the PC?
I do know that some AIs have at least a basic clock in
them because the Fast track pro for instance can run as a stand alone S/PDIF converter but
is this clock overidden when running as a usb interface? I know that if you run the Pro as
2 analogue ins and S/PDIF the latter has to be sync' but is that always the case?
Be gentle with me!
Dave.
|
Johnsy
member
Joined: 06/04/04
Posts: 109
|
Re: usb sync
[Re: ef37a]
#995317 - 29/06/12 09:03 AM
|
|
|
|
The USB audio spec allows no less than three methods of sync: Asynchronous, synchronous
and adaptive.
Adaptive = sample rate converted (ie, re-clocked); synchronous
= clock derived from 8kHz packet rate.
All decent USB i/f's I know of use the
contadictory-sounding asynchronous synchronization method (otherwise known as 'buffer
fill'), which involves continuous feedback from an audio function at the receiving end
controlling the rate of transmission.
The reference clock is provided by the
crystal oscillator in the device, either free-running, or synched (via pll or similar) to
a digital input.
|
ef37a
Joined: 29/05/06
Posts: 5622
Loc: northampton uk
|
Re: usb sync
[Re: Johnsy]
#995335 - 29/06/12 09:46 AM
|
|
|
Quote Johnsy:
The USB audio spec
allows no less than three methods of sync: Asynchronous, synchronous and adaptive.
Adaptive = sample rate converted (ie, re-clocked); synchronous = clock derived
from 8kHz packet rate.
All decent USB i/f's I know of use the
contadictory-sounding asynchronous synchronization method (otherwise known as 'buffer
fill'), which involves continuous feedback from an audio function at the receiving end
controlling the rate of transmission.
The reference clock is provided by the
crystal oscillator in the device, either free-running, or synched (via pll or similar) to
a digital input.
....OoooooKayyyyy? So, if I connect two (let us say identical AIs so as to
K.I.S.S.)to my pc do they each do their own thing or does the pc "take charge"? I have in
fact run two (dissimilar) interfaces on one PC in Samplitude some time ago and they seemed
to keep in step for several minutes at least. I must repeat the exercise.
Dave.
|
Johnsy
member
Joined: 06/04/04
Posts: 109
|
Re: usb sync
[Re: ef37a]
#995363 - 29/06/12 11:10 AM
|
|
|
Quote ef37a:
....OoooooKayyyyy?
So, if I connect two (let us say identical AIs so as to K.I.S.S.)to my pc do they each do
their own thing or does the pc "take charge"?
For this you'd need two things:
a)Multi-device driver
support. b)The ability to select external sync, via wordclock or digital i/o.
Quote ef37a:
I have in
fact run two (dissimilar) interfaces on one PC in Samplitude some time ago and they seemed
to keep in step for several minutes at least.
Without b), each is in sync with the host, but not with each
other.
|
ef37a
Joined: 29/05/06
Posts: 5622
Loc: northampton uk
|
Re: usb sync
[Re: Johnsy]
#995368 - 29/06/12 11:39 AM
|
|
|
|
Ok again John. Beats me how two things can each be in sync with a third but not with
each other but that is just me! (Euclid or something?).
But the point about
S/PDIF is well taken. I shall try this out with a Fast track pro and a NI Ka6 as soon as I
am able. I guess the Ka6 would best be left as "master"?
Dave.
|