Re: EDP Stereo-sync +brilliant undocumented feature.

[Matthias Grob <matthias@grob.org>]

>Kim wrote
>>  brothersync
>>   connection that is necessary for putting multiple echoplexes together 
>like
>  >  this. It does two things, one is to synchronize the actual system 
>clocks
>>   of the different units to operate at the same frequency. That keeps 
>them
>  >  from drifting over time.  The second thing is to provide a 
>precise indication
>>   of when the Record button is pressed.

Andy observes well:

>Er, I disagree. I don't think the exact moment of the Record button being
>pressed is transmitted through Brothersync, at least not in a way that 
>allows
>simultaneous start of record on 2 slaved EDPs. (It very obviously does 
>signal
>the exact length of the loop, of course).

right, we pass the exact moment, but not as an order to the brother.
In case of stereo connection, we use MIDI for the slaving, the ring 
with the loop start pulse is not used, the tip with the word clock is 
enough. (which does not mean you can use a mono cable that shorts the 
ring!)

>I don't presume to know more about the EDP than one of it's designers, 
>but I
>have 2 brand new perfect EDPs which behave as if there is a definate delay
>between the Master EDP starting to record, and the slave starting to 
>record.

correct: there is a delay of about 3ms through MIDI, which does not 
matter as long as its constant.

>I imagine that the 2 loop-start-points are at slightly different places, 
>but
>that this doesn't matter because brother sync keeps the 2 loops together  
>in
>that staggered relationship perfectly. i.e. Audio is recorded in sync, but
>record actually starts on the slave  machine after it does on the master.
>Thus even if you restart the stereo loop, the MIDI delay remains constant 
>and
>so the stereo image is preserved.
>If you hit reverse the stereo image is lost (out of sync by twice the MIDI
>delay)
>but returning to forward play resyncs perfectly, and this occurs even when
>Quantise=On. This is most noticeable if a mono signal is recorded on both
>EDPs.

Makes sense. I dont remember having noticed or discussed this...

>In practice, the only time I notice the difference in timing is when
>reversing a loop which has mono(on both machines) content.

right, because the MIDI delay is doubled instead of compensated.

>This produces a Haas effect on the
>reversed loop, which is quite pleasing, although this would be pehaps
>undesirable if you wanted to record and reverse a true stereo mix.

true. Maybe so far noone complained because a little different 
orientation is somehow expected from a reversed signal?

>I reckon that using MIDI to slave the controls of the 2 EDPs is always 
>going
>to be tricky, because of the MIDI delay, however the system as it is works
>excellently.
>
>Brilliant undocumented feature:-
>Record a stereo-loop, go to the next loop

leaving the previous at the same point, so the are together when you 
come back...

>and record 2 different length loops (totally unrelated times if you like),

unrelated to the previous loop or between the two units?

>whenever you NextLoop back to your original stereo loop it comes out 
>perfectly synced.

because both units simultaneously jump back to where they left the 
previous loop

>This happens (mysteriously) when SwitchQuant=On, although when 
>switching between 2 pairs of loops all having unrelated times the 2 
>EDPs will change loop (after NextLoop) at different times.

now this is mysterious really: With different loop times on the two 
units they dont switch simultaneously when quantized... or do they 
because they quantize to the brother sync instead of the internal 
loop start?
Do they jump together, although one is not at loop end?
If so, this feature may have gone in the new version... maybe we 
should dig it out?

>Seems like I can connect 2 EDPs and they know whether or not they've got a
>synced loop waiting to run. (and that's amazing)

hmm... we did a lot of magic, but this one may be yours... :-)
-- 


          ---> http://Matthias.Grob.org