Topic: Sound Devices 744t inverted phase issues

[fsulloway]

PS- I'm in NYC right now and I PROMISE I'll send those Tishamingo DAT's as soon as I get home! 

Sure you will.

You might want to give Charles a call as he had this problem too. I'm pretty sure he figured it out though.

[wboswell]

Charles bumped my thread over there and here was the response:

New postPosted: Tue Dec 13, 2005 1:30 pm    Post subject:     Reply with quote
There are differences in arrival times between analog inputs and digital inputs with ALL devices.

There is latency inherent in every analog-to-digital conversion. Take a look at your digital preamp's specifications and look for it's "group delay" or "latency" specification and its polarity specification. A complete specification will list these, and what the group delay is at different sampling rates (it is SR dependent). You can compensate for it (latency) in the 744T.

Our input delay is adjustable in microsecond increments precisely for this issue. Each input can be pushed back in time, so if the arrival time of your digital signals is after the analog signals through the 744T, then you can push back the analog signals to match the digital signals. Looking at recorded waveforms after recording is the easiest way to find out the group delay differences. Again, these will usually be different for each sampling rate.

When using only analog sources, the two identical two-channel A/D's on the 744T are clocked at the same rate. With a mix of A/D converters (744T and some other) you will get some phase difference that needs to be compensated for if you are to maintain absolute phase. Again, this is inherent to all analog-to-digital converters.

One point of clarity, while phase and polarity have a direct relationship, if a circuit "reverses the phase" it is inverting signal polarity. A specification would be written that inputs are "inverting" or non-inverting". While absolute polarity is not as important (arguable point), having all inputs and outputs in the same relative polarity generally is, especially with multiple sources and in sound reinforcement applications.
_________________
Jon Tatooles
Sound Devices
Report Bugs to Support

I'm sure he knows what he's talking about, but I have my doubts that latency is the issue here.  One source is 180* out of phase with the other with the end result being all the low end cancelling out. 

[Dr.FOB]

I'm going to take a leap here and forecast that it is the digi out of the V3 that is causing the problem.
I first noticed this long before I got my 744 when I was running the MOTU Traveller and the V3 together but couldn't eliminate the cables from the equation.

I too believe that my Neumann actives are wired in phase and am certain that the cables I made Friday are wired correctly.

Friday night I ran 4 channels for Jazz is Dead.
I ran split 170s with my new silver clad cables straight into 1/2.  I ran the USM69i into the V3>digi>3/4.
It was difficult to see the inversion with this particular combination due to the physical separation but I could hear it in the mix.

I'll make some comps later tonight and post the results.

[wboswell]

I'm going to take a leap here and forecast that it is the digi out of the V3 that is causing the problem.
I first noticed this long before I got my 744 when I was running the MOTU Traveller and the V3 together but couldn't eliminate the cables from the equation.

I too believe that my Neumann actives are wired in phase and am certain that the cables I made Friday are wired correctly.

Friday night I ran 4 channels for Jazz is Dead.
I ran split 170s with my new silver clad cables straight into 1/2.  I ran the USM69i into the V3>digi>3/4.
It was difficult to see the inversion with this particular combination due to the physical separation but I could hear it in the mix.

I'll make some comps later tonight and post the results.

hmm no.  There is no way Grace could ever do anything wrong!

Seriously, I'm with you.  Its not a minor difference.  We're talking major lack of bass response.  May be worth an email to Grace by someone who cares.

another thing to do would be to route an identical analog signal to all inputs on the 744, then compare the .wavs in DAW.  Then add the V3 and see what happens.  That should tell you something

[it-goes-to-eleven]

Interesting response from SD..  This isn't a problem I was aware of but I can see how it could happen.

If true, this isn't inversion but time offset between channels. I think of inversion as flipping the wires in a cable or inverting the wave in the DAW. 

If this is what SD describes.. The V3 buffers a certain number of samples before starting to stream them down the AES output, that would create a 'delay' on that channel. That delay is simply different then the delay in the 744.  I don't see that as a v3 problem or 744 problem..

Looks like you need to do some good testing with test tones and careful data gathering.

[pilgrims622]

So far, I havent had any phase issues running 4 channels with the v3 and 744....knock on wood

2 different sources so far 4 times....

run for 2 shows
(Neumann ak40+ak20 > lc3 > km100 > Lunatec V3(digi out)) + (Neumann km143) > Sound Devices 744t @ 24/48
&
run for 2 shows
(Neumann ak40+ak20 > lc3 > km100 > Lunatec V3(analog out)) + (Schoeps cmc621xt > Lunatec V3(digi out)) > Sound Devices 744t @ 24/96

I am recording in mono and using the 1 gig split if that helps...Cant think of what else I changed on the 744, but Charles did walk me thru each step like a baby

[cleantone]

For what it is worth, I can say that I have been running AES/EBU out of my V3 for years into my MOTU rig and have NEVER had the polarity reversed. Countless times.

another thing to do would be to route an identical analog signal to all inputs on the 744, then compare the .wavs in DAW.  Then add the V3 and see what happens.  That should tell you something

This is the best way to verify this right quickly.

[wboswell]

Interesting response from SD..  This isn't a problem I was aware of but I can see how it could happen.

If true, this isn't inversion but time offset between channels. I think of inversion as flipping the wires in a cable or inverting the wave in the DAW. 

If this is what SD describes.. The V3 buffers a certain number of samples before starting to stream them down the AES output, that would create a 'delay' on that channel. That delay is simply different then the delay in the 744.  I don't see that as a v3 problem or 744 problem..

Looks like you need to do some good testing with test tones and careful data gathering.

the latency mentioned above would be similar to that of taking a sbd recording and a mic recording made with two different machines and trying to synch them up in post.  I see these two issues as apples and oranges.

I guess I need to touch base with Charles as he had the same problem as I had and seems to have it fixed.

For what it is worth, I can say that I have been running AES/EBU out of my V3 for years into my MOTU rig and have NEVER had the polarity reversed. Countless times.

unless the MOTU has its polarity reversed also

[cleantone]

If true, this isn't inversion but time offset between channels. I think of inversion as flipping the wires in a cable or inverting the wave in the DAW.

If this is what SD describes.. The V3 buffers a certain number of samples before starting to stream them down the AES output, that would create a 'delay' on that channel. That delay is simply different then the delay in the 744.  I don't see that as a v3 problem or 744 problem..

This is why I was asking my initial questions. It seems obvious enough, so it is not likely the case. If your mixing multiple sources you need to time align everything before you mix it. This would be more of a true phase issue and not polarity. The pictures I saw from the linked thread (http://taperssection.com/index.php?topic=49528.0) definitly look like polarity (attatched)

I assume this is the same problem right? This photo clearly shows reversed polarity and not a time delay causeing phase issues. I think it is pretty safe to assume neither pairs of cables were wired backward, right? That would cause (and solve) this issue.

The definitive test would be to route an identical signal (not 4 different mics) to the inputs with the same setup. A sine wave would be usefull. I assume you could get the same results by sending a stereo signal, left channel 744 analog 1, right channel V3 channel 1 to see if the polarity is getting reversed or if there is a latency delay issue. You woldn't really need to run four channels to test it though it would work.

[Dr.FOB]

For what it is worth, I can say that I have been running AES/EBU out of my V3 for years into my MOTU rig and have NEVER had the polarity reversed. Countless times.

There is another thing to check.  Perhaps the outputs are independantly driven with the AES configured differently than the coax SPDIF.

[Brian Skalinder]

Perhaps the outputs are independantly driven with the AES configured differently than the coax SPDIF.

Depends on which AES output you're using on the V3.  AES1 is driven from the same transmitter as the coax S/PDIF output.  AES2 has its own transmitter.

[cleantone]

Dr.FOB,

So you really think it is the V3 sending reversed polarity out the SP/DIF? I REALLY doubt it. I would be VERY surprised if that was the case. I'm beyond curious with this now. I really wish I had the gear so I could figure it out right now. I have used all the outputs of the V3 all at once an never noticed any polarity issues. I have used the analog outs to monitor, one AES/EBU to my rig, one to a backup rig, and SP/DIF to DAT for two track back up all at once. I have never had a polarity issue that did not result from a cable being wired backward. I have also never used the 744. Since this problem has come about from two or more people on this board I have to assume it is something about the 744 unit.

PLEASE figure this out today... 

Depends on which AES output you're using on the V3.  AES1 is driven from the same transmitter as the coax S/PDIF output.  AES2 has its own transmitter.

Didn't know that. I have used either and sometimes both.

[Dr.FOB]

Depends on which AES output you're using on the V3.  AES1 is driven from the same transmitter as the coax S/PDIF output.  AES2 has its own transmitter.

I can also see that a difference might be noted if an XLR>rca connector was wired incorrectly.
But I have been using the coax output...

[wboswell]

FWIW, I used the s/pdif output and not the AES out.

I didn't notice that the two sources were out of phase until I authored a DVD.  On my computer it sounded pretty normal but when I played it back in my den, it was very obvious.  It may not jump out an grab you unless you're putting it on something capable of reproducing the low end.

[jamiex]

Hi all-

Jamie from Grace Design here.  I just received an email from Jim, which drew our attention to the thread.  Sorry there has been so much confusion regarding this.

The digital outputs (AES1, AES2, S/PDIF and TOSLINK) are 180 degrees out of phase.  As you know absolute polarity does not have an affect on the sound when the relative polarity between the channels is consistent.  It does however become an issue if you are summing multiple sources that have inverted polarities, which is the case here.  With the V3 digital outputs 180 degrees out of phase from the analog inputs on the 744, the low frequency response will be compromised if these signals are summed.

The fact that the V3's digital outputs are 180 degrees out of phase was only recently discovered when both the analog and digital outs were recorded simultaneously to a workstation.  Upon discovery we immediately corrected this issue in new production units.  Beginning with serial number V3497, all of the digital outputs are in phase with the input.  Any unit prior to this serial number will have inverted polarity on the digital outputs.

Again, when recording two channel audio this is not a problem, but for multichannel or matrix recordings that use the digital outputs, this should be corrected.  Obviously we are here for any of you who want/need to have the digital output in phase with the input and will provide this service at no charge under warranty.  To arrange for this service, please contact me at jamie@gracedesign.com.

Once again our apologies for the confusion and any trouble this may have caused.

Best regards from snowy Boulder, CO

- jamie