Topic: Syncing a V3 Word Clock output with a DMIC-20 AES REF input

[Brian Skalinder]

Chuck -- you should talk to Scott on the CO Crew.  He got a Mytek Stereo192 A/D with a word-in and got his V3 reverse modded back to the original (not the first to have it done according to Grace).  He's been using the Mytek and V3 with word out to record 4channels and mix them in post.  With all 4 channels using the same clock, syncing is a breeze.

Hey Scott - you out there?  Just curious as I dream about future 4-ch options...

What recorders are you using behind the V3 / Mytek?
What pre in front of the Mytek?
Mics?

[gratefulphish]

I took a look at the DMIC-24 manual, and it seems to indicate that the input is to ensure that the DMIC samples at the same rate as the other device.  From the manual:

Connecting AES Reference
The DMIC-24 can be synchronized to an alternate output frequency (from the default frequency)
by connecting a valid AES reference signal to the AES REF BNC connector. The EXT REF LED
on the front panel lights when a valid reference signal is applied.
To phase lock the unit’s sample frequency to a reference, connect a reference signal (between 30
KHz and 50 KHz sampling rate) to the AES REF BNC connector. The AES signal output will now
phase-lock to the incoming reference — independent of the DMIC’s default reference frequency.

It does not seem to be a wordclock in/out.

That's funny, I read that a whole different way. It sounds to me like by runing AES out of the V3 to the AES in on the DMIC-20 they will share the same clock. That's what I want.

It refers repeatedly to the sample frequency, which must be between 30-50KHz (like 32, 44.1 or 48), and then seems to indicate that this will automatically cause the DMIC to sample at the same frequency.  My R-4 Pro has an indicator, so that if I have a different sample rate set on my V3, than on the R-4 Pro, it lets me know, so that I can reset one or the other.  This connector just seems to do the job for you, preventing two different devices from sampling at different rates.  I might e-mail GP for an answer.  I would like to know myself.  This is just my interpretation of the manual, and we all know how well those are all written.

[Todd R]

Brian -- it might be worth dropping Scottsch3 a line in the CO crew thread, I don't know how much he trolls the rest of ts.com.

He uses at least a MT for one of his recorders, but I seem to remember he has some sort of multi-track recorder.  The mytek side is V2>mytek, plus he has the V3.  He also archives for Brothers Past, so I think he often is getting the channels from FOH or the monitor board.  But for mics he has DPA 402x, DPA subcards, and DPA 4061's (notice a trend?).

On the AES Ref issue, I don't think it'd be easy to implement what Gratefulphish is saying -- you'd need to somehow have the ability for the DMic to generate any sampling frequency between 30khz and 50khz, which would be no easy feat.  It seems much more likely to simply strip this clocking frequency information off the AES input.  Of course, checking with GP is a good idea, but I don't think the DMic would be using the AES signal to simply decide what to clock at -- it'd then still need to somehow generate that clock.

[gratefulphish]

Maybe this helps clear things up.  From the front of the DMIC manual:

Two versions of the DMIC-24 are available:

• 44.1 KHz Model
This model defaults to 44.1 KHz, via the internal crystal. However, if an external
reference is applied to the AES REF connector, the output frequency will lock to the
reference.

• 48 KHz Model
This model defaults to 48 KHz, via the internal crystal. However, if an external reference
is applied to the AES REF connector, the output frequency will lock to the reference.

NOTE The printed legend on the rear panel indicates the default frequency.

This is the "default" frequency referred to in the other section that I quoted.  In other words the frequency is fixed, depending on the model you purchase, unless you go in through the AES reference input, in which case, it will lock to that frequency, overriding the default.  Note that it appears that you can only get 32, 44.1 or 48 under any circumstances. 

[Chuck]

Brian -- it might be worth dropping Scottsch3 a line in the CO crew thread, I don't know how much he trolls the rest of ts.com.

He uses at least a MT for one of his recorders, but I seem to remember he has some sort of multi-track recorder.  The mytek side is V2>mytek, plus he has the V3.  He also archives for Brothers Past, so I think he often is getting the channels from FOH or the monitor board.  But for mics he has DPA 402x, DPA subcards, and DPA 4061's (notice a trend?).

On the AES Ref issue, I don't think it'd be easy to implement what Gratefulphish is saying -- you'd need to somehow have the ability for the DMic to generate any sampling frequency between 30khz and 50khz, which would be no easy feat.  It seems much more likely to simply strip this clocking frequency information off the AES input.  Of course, checking with GP is a good idea, but I don't think the DMic would be using the AES signal to simply decide what to clock at -- it'd then still need to somehow generate that clock.

I called GP, and the guy I need to talk to is out for the rest of the day.
I e-mailed him and will call if I don't hear something by late tomorrow.

Todd, when you say Scott had his opti-mod V3 reversed, that sounds weird, because when I talked to Neal over at Grace, he said the sync should work fine from either the SPDIF (RCA) or AES (XLR) outputs. Essentially making the extra work to re-add the clock output un-neccessary.

I have been working on trying to figure out how to get coax SPIDIF input rigged into my iRiver iHP-120 just in case I really do need the clock out jack on the V3 back to make this work...    

Of course, if this is all possible, I have a few INA217's around here and will probably want to gain pot mod the DMIC too.

I am mod crazy!  

[Todd R]

Grateful -- I'm not sure where you're getting that the Dmic will only do 32k, 44.1k, and 48k -- it seems quite clear from the manual that is will lock to any external source between 30k and 50k.  Chuck's got the call in, that should clear up exactly how it works.

Chuck -- I mentioned Scott simply since he has been doing 4ch mixes using two rigs with common clocks, thought you might be interested in how he does the post production mixes.  He needed the reverse mod done since the Mytek only takes a word clock in.  It isn't set up for stripping the clocking info off an AES stream like we're assuming the DMic does.  So you wouldn't need this reverse mod done.  (I'm just imagining that a work clock input needs just a word clock stream.  Is Neal at Grace saying any word clock in can just take an AES full stream automatically and strip off the audio to get the clocking info?  In which case, I don't know why you'd need a separate word clock out if you already have two AES outputs on the V3.)

[Chuck]

Grateful -- I'm not sure where you're getting that the Dmic will only do 32k, 44.1k, and 48k -- it seems quite clear from the manual that is will lock to any external source between 30k and 50k.  Chuck's got the call in, that should clear up exactly how it works.

Chuck -- I mentioned Scott simply since he has been doing 4ch mixes using two rigs with common clocks, thought you might be interested in how he does the post production mixes.  He needed the reverse mod done since the Mytek only takes a word clock in.  It isn't set up for stripping the clocking info off an AES stream like we're assuming the DMic does.  So you wouldn't need this reverse mod done.  (I'm just imagining that a work clock input needs just a word clock stream.  Is Neal at Grace saying any word clock in can just take an AES full stream automatically and strip off the audio to get the clocking info?  In which case, I don't know why you'd need a separate word clock out if you already have two AES outputs on the V3.)

That's what I understood Neal saying... at least for my needs. Maybe I won't have to undo the optical mod on my V3. That would be great!
If this works I'll be four track recording... on the cheap, soon. I recently made a deal to purchase another iRiver, the iHP-140. 

It's a wonder this question hasn't come up before, since both these pieces of gear have been around for a while.

[trajhip2000]

I don't believe a typical wordclock input will take a regular digital feed and pull the wordclock out - but I do believe that the DMIC-20 AES input will do that based on their specs. Lots of devices are set up to sync their wordclock to that extracted from a digital feed, I'm pretty sure my 744T does that.

Steve

[Todd R]

So, another iRiver, huh?  Looks like you're putting your days of DAT behind you.  Are you planning on modding the Dmic-20 for an optical out?  I've heard it's pretty easy, at least electronically -- getting a hole cut into the back of the Dmic might be the hardest part.  Get in touch with Richard/poorlyconditioned, I think it was him who did the optical mod.

[gratefulphish]

Grateful -- I'm not sure where you're getting that the Dmic will only do 32k, 44.1k, and 48k -- it seems quite clear from the manual that is will lock to any external source between 30k and 50k.  Chuck's got the call in, that should clear up exactly how it works.

What other frequencies have ever been used between 30 and 50K other than 32, 44.1 and 48?  It does say anything between those numbers, but my guess is that it was easier than saying, "between 32 and 48K"  Just sounds nicer to say 30-50, even if it doesn't mean anything.  As I also read the manual, it would also mean that you could not do 24/88.2, 96 or 192 under any circumstances.

Another issue is that the DMIC-20 is a 20 bit converter, and a V3 is either 16 or 24, creating other compatibility, synching and conversion issues.

[Todd R]

Ok Chuck, we're going to need a report back from those GP folks.    I'm still assuming they mean anywhere between 30k and 50k.  I'm not sure how often it came into play, but for some reason video people use 44,056 not 44.1k for sampling, just as a for instance.

You should be able to mix a 4ch mix of the V3 at either 16 bits or 24 bits with the 20bit recording from the Dmic -- that shouldn't cause a problem with syncing.  It does make me wonder on how to record off the DMic though -- it seems like truncating the 20bit signal to 16 bits would be bad.  Maybe there's enough/the right level of noise in the system to provide some natural dither to 16bits.

And I'm with you Chuck -- why haven't we thought of this before.  Seems like a nice way of doing 4ch mixes.  I may need to look into a used DMic.

[Chuck]

Grateful -- I'm not sure where you're getting that the Dmic will only do 32k, 44.1k, and 48k -- it seems quite clear from the manual that is will lock to any external source between 30k and 50k.  Chuck's got the call in, that should clear up exactly how it works.

What other frequencies have ever been used between 30 and 50K other than 32, 44.1 and 48?  It does say anything between those numbers, but my guess is that it was easier than saying, "between 32 and 48K"  Just sounds nicer to say 30-50, even if it doesn't mean anything.  As I also read the manual, it would also mean that you could not do 24/88.2, 96 or 192 under any circumstances.

Another issue is that the DMIC-20 is a 20 bit converter, and a V3 is either 16 or 24, creating other compatibility, synching and conversion issues.

Good points. I'll see what Graham Patten says tomorrow. I am hopeful, at least for now.

So, another iRiver, huh?  Looks like you're putting your days of DAT behind you.  Are you planning on modding the Dmic-20 for an optical out?  I've heard it's pretty easy, at least electronically -- getting a hole cut into the back of the Dmic might be the hardest part.  Get in touch with Richard/poorlyconditioned, I think it was him who did the optical mod.

I may do an optical mod on it, if all this stuff shakes out correctly. There is a thread around here somewhere. I think you are right, that Richard did mod one of his units. He's been having problems with his DMIC-20's lately I heard.

Todd, didn't you swap the instrumentation chips out of Josh's old DMIC-20 a while back? INA217's? I happen to have some I was going to use in a pre-amp project. If all this works right, I will do the gain potentiometer mod that allows adjustment between the various 5 db steps. The gain will be set just like the V3 in that respect...  I hope it's cool to run line-in on the DMIC at it lowest gain setting.

[Todd R]

Todd, didn't you swap the instrumentation chips out of Josh's old DMIC-20 a while back? INA217's?

Could be, I really don't remember at this point.  We talked about doing it, I just can't remember if we followed thru since Josh was having troubles with his DMic.  I'm pretty sure others have done it.  And the INA217 (or THAT1510) are drop-in replacements for the SSM2017 chip, which is dip-socket mounted into the DMic, so swapping them is a breeze.

[Chuck]

My e-mail to Graham Patten:

I do field recording and need to be able to sync two stereo sources to the same clock for four channel, digital, recording.
I have a Grace Lunatec V3, which has SPDIF (RCA) and AES (XLR) outputs. I also have a DMIC-20 available, but I don't have it in my hands yet.

What I need to know is... Can I run the Lunatec V3 SPDIF or AES output into the AES input (on the back of the DMIC-20) to sync the DMIC-20 to the V3?

My V3 does not have a word clock output, so that is not an option for me.

The goal is to have the digital recordings made from both devices, in perfect sync, so that they can be edited together without the clock drift they would have if they were recorded on two separate machines with two separate clocks. Is this possible and if so, am I thinking about it the correct way?

Any help or advice would be greatly appreciated.

Thanks,

Chuck Miller



This in from GP:

Hello Chuck,
 
You will have no problem using the DMIC-20 AES input with S/PDIF for Reference.
 
Mike 
 
Mike Sauter
Test/Service Technician
(888) 622-4747
www.gpsys.com


I am in business!  I will report back when I get the DMIC-20 to confirm...

[gratefulphish]

I am really not sure what that reply meant, even though your question was written clearly.  First of all, as I indicated yesterday, this just does not seem to be what their manual says.  Second, in his answer he seems to suggest that you can just plug a SPDIF source into the AES input, and that would just "work."   As I have understood it, and I have spent some time reading about it, (as all of my DAT decks were SPDIF, but the R-4 Pro is AES) the signals are different, at different Ohms, and to truly convert data from one to the other can be accomplished with a cable that will harshly perform the function, or boxes costing hundreds or even thousands of dollars to properly do the signal conversion. Finally, he just said for "reference."  Reference in their manual seemed to only be sampling frequency. I have seen nothing in there about word clock.  Is this guy the senior tech?   Hasn't anyone else ever experimented with this function.  It still doesn't sound right to me, but give it a try (I would try to borrow the DMIC first, and find someone with an unmodded V3) and you can find out for sure.