3 Mar
2008
3 Mar
'08
2:54 p.m.
On Mon, Mar 3, 2008 at 8:58 AM, Kevin Cosgrove <kevinc(a)cosgroves.us> wrote:
On 2 March 2008 at 10:23, Joe Hartley <jh(a)brainiac.com> wrote:
At first clush it sounded OK to me also, but after a minute I decided
that it isn't likely to find single bit failures. If there is a slow
drift between the two cards what might happen over time would be one
card (the master) would record all the data but the slave card might
skip a byte.
What's the best way to ensure that the cards
are actually in sync?
I've recorded an hour of a metronome in Ardour with the signal split
to an input on each card and saw no drift between the tracks. I then
played them back and heard no drift even after 30 minutes, but I have
a feeling that this isn't necessarily a valid way of testing.
Your test method seems to be closest to what you'll do in actual use.
In that light it seems pretty valid to me.
Another way to test the system is to record a sine
wave, maybe 1kHz
but other frequencies are good too. When you're done recording, then
run an FFT on the result and look for frequencies other than 1kHz.
If you use a large signal size, then you might end up seeing
distortion products in the frequency spectrum. Certainly you'll see
some noise that's broadly distributed. Some of that noise will be
due to clock jitter. Of that noise, some could be due to an increase
in clock jitter when using an external sync versus using the internal
clock. You could do a comparison.
I think this sounds like a better solution. I would personally look
for a way to treat the audio information as data, calculate the the
expected data from each card and then do a comparison of input and
output data mathematically. I would not personally want to trust
either my eyes or ears in this area. Let the operation of the hardware
speak for itself.
In all likelihood, there are going to be small phase shift differences
between the cards. The inputs won't be perfectly in sync even if they
are coming from the same sine wave generator. I would probably doing
something like a standard deviation comparison between the input and
two outputs. If the data looks identical, within some some deviation
amount, then most likely everything is working OK. On the other hand
if there is a single byte getting dropped or added once in awhile then
it will show up doing this sort of analysis very quickly.
If you don't want to do it in C or some programming language then you
can likely do it in Open Office or Excel.
Hope this idea helps,
Mark
Anyway, I think your system is probably sound. But, you could test
it more if it makes you happy.
Do let us know what you find out please.
Thanks....
--
Kevin
_______________________________________________
Linux-audio-user mailing list
Linux-audio-user(a)lists.linuxaudio.org
http://lists.linuxaudio.org/mailman/listinfo/linux-audio-user