Hi,

I recently bought a Behringer ADA8000 8-channel preamp with ADAT out and an
M-Audio FW1814 interface with 8 analog ins and 8 ADAT ins. The aim was to
have 16 channels of analog input (to record full bands). In the past I've
had problems with phasing and sync when I split the drum kit over two
interfaces - eg overheads using one interface and snare on the other.

I thought that one interface with both analog inputs and ADAT inputs would
solve this problem, but today I ran an experiment (to cut it short, I ran a
split, ie identical, mono signal into both analog input 1 and ADAT input 1
and recorded the results into two parallel mono wav files) which suggested
otherwise. When I normalised both files to the same level with their RMS,
reversed the phase of one file and mixed it with the other, the signal
didn't cancel out as it should - instead I was just left with a phasey,
horrible sounding track. The average amplitude was about the same as it was
before.

I noticed that one of the tracks was behind the other by about 5ms so I
corrected this by lopping off the first 5ms and tried again to invert and
mix. This time I still got a phasey track, but it was a little quieter, and
with much less low end. More like what I should've got, but it was still
peaking at a good -8 dB or so.

What am I doing wrong? How can I sync up the ADAT inputs and the analog
inputs so that I know I can record a drum kit using both without worrying
about phasing problems, if I need more than 8 channels? I was using the
clock on the ADA8000 as master, which I've heard is fairly poor, so would I
improve performance by finding a cable and slaving to ADA8000 to use the
M-Audio's word clock instead? I don't have the budget to buy an external
clock.

Thanks,

TJ

In article writes:

I recently bought a Behringer ADA8000 8-channel preamp with ADAT out and an
M-Audio FW1814 interface with 8 analog ins and 8 ADAT ins. The aim was to
have 16 channels of analog input (to record full bands). In the past I've
had problems with phasing and sync when I split the drum kit over two
interfaces - eg overheads using one interface and snare on the other.

Well, then don't do that. The reason will become clear.

I ran a
split, ie identical, mono signal into both analog input 1 and ADAT input 1
and recorded the results

I noticed that one of the tracks was behind the other by about 5ms so I
corrected this by lopping off the first 5ms and tried again to invert and
mix. This time I still got a phasey track, but it was a little quieter, and
with much less low end. More like what I should've got, but it was still
peaking at a good -8 dB or so.

You're on the right track. All things digital have a certain
"throughput delay" and it's never the same for two devices. You just
have to get them lined up accurately. Also, they need to be running
from the same word clock so that once you get the absolute start time
lined up, they don't drift apart in time due to the independent
clocks. That's what gives you the "phasey" sound when you mix them
together.

You need to choose one to be the word clock master and the other to be
the word clock slave, and connect the master's word clock output to
the slave's word clock input (and set it to use the incoming word
clock signal rather than its internal clock source). Hopefully one of
your devices has a word clock output and the other has a word clock
input. You'll need to experiemnet to see if one makes a better master
than the other, but I suspect that if it works at all (some clock
outputs won't properly drive some clock inputs) both ways will be
about the same.

This is never simple. You just have to keep messing with it.


--
I'm really Mike Rivers )
However, until the spam goes away or Hell freezes over,
lots of IP addresses are blocked from this system. If
you e-mail me and it bounces, use your secret decoder ring
and reach me he double-m-eleven-double-zero at yahoo

"TJ Hertz" wrote in message



today I ran an
experiment (to cut it short, I ran a split, ie identical,
mono signal into both analog input 1 and ADAT input 1 and
recorded the results into two parallel mono wav files)
which suggested otherwise. When I normalised both files
to the same level with their RMS, reversed the phase of
one file and mixed it with the other, the signal didn't
cancel out as it should - instead I was just left with a
phasey, horrible sounding track. The average amplitude
was about the same as it was before.

This is consistent with what you found next - that the
tracks were not in perfect synch.

I noticed that one of the tracks was behind the other by
about 5ms so I corrected this by lopping off the first
5ms and tried again to invert and mix.

If you want those tracks to be in perfect synch, you're
going to have to learn how to do better. You're also going
to have to be able to tolerate less than perfect timing
between tracks.

This time I still
got a phasey track, but it was a little quieter, and with
much less low end. More like what I should've got, but it
was still peaking at a good -8 dB or so.

What am I doing wrong?

Probably nothing, at least as far as clocking goes.

How can I sync up the ADAT inputs
and the analog inputs so that I know I can record a drum
kit using both without worrying about phasing problems,
if I need more than 8 channels?

The usual rule of thumb is that an audio interface has just
one clock, and I'm going to assume that this is the case
with the FW1814. Given that your source was a mulitchannel
source, there's plenty of evidence that you were in some
sense clocking it properly, because if its clock was in any
sense running free, your interface would not be able to
demultiplex the channels.

I was using the clock on the ADA8000 as master, which
I've heard is fairly poor,
so would I improve performance by finding a cable and
slaving to ADA8000 to use the M-Audio's word clock
instead?

I don't think that you're going to fix your situation with
better clocking. The interface seems to be in about as good
synch as its ever going to get.

I don't have the budget to buy an external clock.

I don't think you need one. If you had a clocking problem
that you could fix, you could probably just run a second
lightpipe to the ADA8000's output side and set it up to
synch from the source. Set up the FW1814 to be the master,
and that would be that.

Rather, I think that Mike got on the right track when he
started talking about latency. I think you're in a situation
where your FW1814 analog input channels have a certain
amount of latency and your ADA8000 analog input channels
have a different amount of latency. Other than running both
interfaces at a higher clock (which in your case isn't that
much faster of a clock so forget it) you can't change your
audio interface's latency.

Some recording software (like Audition/CE) allows you to
compensate for the latency of each audio interface or
channel or channel pair seperately. The ADAT I/O channels on
the FW1814 may be independent enough from the analog inputs
that they would be treated as separate items by your
recording software. I don't know for sure if this feature
is available to you. You should look for it.

If you can adjust for the differences in interface latency
in your software, you're probably going to be able to adjust
it within 100ths of a mSec or better, which is pretty close
to perfect.

Sound travels about a foot in a millisecond, so you may be
able to correct your inter-channel synch to be about the
same as having two mics different distances from the source
by a small fraction of an inch. This is probably going to be
good enough.

Depending on your recording software, you can synch tracks
by hand much closer than 1 mSec. By keeping track of what
you are doing when you hand synch you can figure out what
the right latency settings are for your software if this
feature is available to you.

With a recording of a sharp click I can hand-synch two
tracks in Audition within a few samples. That's different
by only a tenth of an inch or less. With better synch you
can move the comb filter effects up so high that they aren't
much of a problem. How carefully do you and can you position
mics?