OT: RME Wordclock

[mr_es335]

Hello,
I would appreciate some assistance with a question regarding the following:

RME HDSP 9652
Behringer ADA-8200

In configuring the RME with the ADA, one of my acquaintances has told me unequivocally that I should be using the wordclock feature which is included on expansion board for the RME HDSP 95652.

As my present computer is built using a 1U chassis, using the expansion card would be either difficult or impossible. My present configuration is as follows:

2750

So, my question is, "Do I actually required wordclock in this situation?"
PS: If not, then maybe a brief description as just why this is so would be appreciated.

Thank you.

[cgrafx]

Hello,
I would appreciate some assistance with a question regarding the following:

RME HDSP 9652
Behringer ADA-8200

In configuring the RME with the ADA, one of my acquaintances has told me unequivocally that I should be using the wordclock feature which is included on expansion board for the RME HDSP 95652.

As my present computer is built using a 1U chassis, using the expansion card would be either difficult or impossible. My present configuration is as follows:

2750

So, my question is, "Do I actually required wordclock in this situation?"
PS: If not, then maybe a brief description as just why this is so would be appreciated.

Thank you.

Since your using both the Inputs and Outputs the ADAT clock will work just fine, you do not need to use a separate word clock.

I do recommend putting a dab of silicone glue between the ADAT connectors and the edge of the card or back of the ADA8200 to make sure your cables don't come loose. (make sure you do this AFTER you have inserted the cables)

[mr_es335]

Philip.


Since your using both the Inputs and Outputs the ADAT clock will work just fine, you do not need to use a separate word clock....Thank you Philip for your reply.

Always appreciated.

[Cary B. Cornett]

As a general rule, a coaxial word clock connection is more stable than optical. However,, especially for live performance work, this is not that big a deal.

Since you are using only one converter box, you might want to consider making the ADA8200 your clock master. This will likely give you the lowest possible clock jitter. Remember that where timing jitter really matters is at the converter itself, because that's where jitter can actually affect the sound. There is almost ALWAYS an increase in jitter when slaving to an external clock. This is because of how a PLL works. If you use the internal clock of the converter box, it isn't being made less stable by having to follow an external clock.

So, if you are using only one converter box, and you want the most stable clocking for the converters, set the converter box to its internal clock and have the interface on your computer being clocked from that. Optical will be fine for this.

[Dave Labrecque]

As a general rule, a coaxial word clock connection is more stable than optical. However,, especially for live performance work, this is not that big a deal.

Since you are using only one converter box, you might want to consider making the ADA8200 your clock master. This will likely give you the lowest possible clock jitter. Remember that where timing jitter really matters is at the converter itself, because that's where jitter can actually affect the sound. There is almost ALWAYS an increase in jitter when slaving to an external clock. This is because of how a PLL works. If you use the internal clock of the converter box, it isn't being made less stable by having to follow an external clock.

So, if you are using only one converter box, and you want the most stable clocking for the converters, set the converter box to its internal clock and have the interface on your computer being clocked from that. Optical will be fine for this.

Cary -- interesting. I did not know this.

Why wouldn't any jitter affect the sound equally wherever it occurs (notwithstanding the degree of jitter actually being worse in one place than the other)? The "tainting" of the signal is happening in the digital domain in either case; the conversion is a separate process from the "jitter process," isn't it?

[cgrafx]

1. ADAT clocks in almost all cases are perfectly stable.
2. The clock on the RME is a higher quality clock than the clock on the ADA8200.
3. Using the RME as master allows SAW/SAC to control the clock rate dynamically. Granted the ADA's only support 44.1 and 48k, but there are legitimate reasons to want to be able to switch from one clock rate to the other. If you set the ADA as the master than you will manually have to change your clock in three places anytime you want to change those settings (once on the ADA, once in the RME settings and once in SAC/SAW).

Lets not make this anymore complicated than it has to be.

1. Connect your ADA with ADAT input/output cables. (the output cable from the RME to ADA will supply the clock for the ADA)
2. Set the ADA SYNC to ADAT IN.
3. Set the clock mode on the RME driver to internal (that will allow SAW/SAC to control the clock rate).

At this point if your sync light is working your good to go.

You are now free to go make some music.

[Cary B. Cornett]

Cary -- interesting. I did not know this.

Why wouldn't any jitter affect the sound equally wherever it occurs (notwithstanding the degree of jitter actually being worse in one place than the other)? The "tainting" of the signal is happening in the digital domain in either case; the conversion is a separate process from the "jitter process," isn't it? First, let's look at what jitter IS: a very slight "wiggle" in the exact timing of each clock pulse. One pulse may be exactly "on time", another may be a bit early, and another may be a bit late. This may also happen to the pulses of the data stream itself.

Now let's look at the EFFECT of jitter. In a digital audio data stream, each pulse happens for a certain period of time, then there is a transition to the next pulse. Lather rinse, repeat, over and over. The device receiving the pulse actually "reads" that pulse somewhere in the middle of the pulse. A very slight variation in pulse timing won't have any effect because there is a decent amount of "wiggle room" in which the pulse can be correctly read. If the jitter exceeds about half of the pulse's designated timing "window", the WRONG pulse will be read. This would require a very large amount of jitter, so large that the system would have to be really messed up for it to be that far out of whack. Therefore, jitter does not cause trouble in an existing data stream. Inside your computer, this becomes irrelevant because buffering is used to allow what had been synchronous data to be processed in an asynchronous fashion. The computer system keeps up by being way faster in processing than the information coming in or going out.

Where jitter becomes important is at the converter, either A/D or D/A. Going from analog to digital, you have a continuous waveform coming in, and it gets sampled at regular intervals. The important word here is "regular". If a sample happens a little early or late, it is reading the wrong part of the wave, so it gets the wrong level for that instant, which means now we have distortion. The more early or late the sampling point gets, the more distortion you have. There is no nice broad "window" where timing error does not matter. The same problem exists for going from digital to analog, because feeding the sample out at the wrong time puts the sample in the wrong place. This will seem strange, but the right level put out at the wrong time is, in fact, the wrong level. Again you get distortion.

This is why the place where clock stability is most important is at the converter.

[Cary B. Cornett]

1. ADAT clocks in almost all cases are perfectly stable.

There is no such thing as a perfect clock. All clocks have some jitter. In the better clocks, the jitter is just very small. Any time a clock is "carried" from one place to another, an increase in jitter may be introduced. This increase is more likely, and may be worse, in an optical connection as compared to a copper connection. Were it not for this difference, no manufacturer of any converter with ADAT optical interface would bother with providing a coaxial word clock connection.

Now, that said, I have never heard any audible problem with clock carried over ADAT optical. Then again, jitter problems are usually very subtle in their effect. In a PA system, I just don't worry about it. For recording purposes, I may want to be more careful.


2. The clock on the RME is a higher quality clock than the clock on the ADA8200.
I am quite certain that this is true. Still, if using the RME as master clock, it is still best to use a coaxial clock connection if you can. If you are using two or more converter boxes, having the RME be master makes sense, because that is the only way to have the quality of clock for all channels be the same. And when you do this, if you can, coaxial is better, especially for recording purposes.

Now we get to the tricky part. In any given device, almost invariably the most stable clock is an INTERNAL clock. And yes, even for cheap converters this will be true. Why? Because the internal clock, and the way toe device carries it around inside itself, has some minimum amount of jittler. Using a Phase Locked Loop (pretty much everybody uses this) to "slave" the clock to an external source is, especially in cheaper designs, LESS stable than the internal clock, even if the internal clock is already less stable than the external clock source. In most converters, more especially the cheap ones, the PLL adds jitter, so using an external clock is not likely to be much advantage over the internal clock.

Yes, using the internal clock on the converter box is less operationally convenient, but if having the best possible clocking is important to you , you accept the inconvenience and just remember to take the extra steps when necessary.

There is something to be said for keeping it simple, especially if you are working PA instead of making a recording, but it is always best to know and understand the tradeoffs in the choices you make. You decide for yourself what is most important, but you should understand the consequences of your choice.

Oh, and if you want a good "lookup" reference about clocking, look in Bob Katz's book "Mastering Audio".

[cgrafx]

Yes in a high-end environment with very expensive converters, speakers, amplifiers, etc, I'm sure you would be able to measure the small amount of distortion caused by clock jitter (particularly if you have a crappy clock, bad cable connections or other poorly setup configuration).

In the real-world systems which 99% the rest of us have it simply won't matter.

The place where coax certainly has an advantage is its significantly more robust a connection, and thus my recommendation to put a dab of silicon glue across the ADAT connector to ensure it doesn't come loose.

Keep in mind the question relates to a $175 Behringer 8-channel AD/DA that is being used for live sound (thats just a little under $11/ch).

It is silly to discuss what amounts to functionally non-existent clock jitter distortion in this context.

I can guarantee connecting the ADA with Coax and clocking from the ADA instead of from the RME card will cause stupid problems down the road.

At some point the RME card or the Audio software will end up out of sync because the clock rates won't match and it will happen at the most inopportune time.

The ADAT cables have to be connected anyway, so just use the clock thats already there.

[Cary B. Cornett]

Yes in a high-end environment with very expensive converters, speakers, amplifiers, etc, I'm sure you would be able to measure the small amount of distortion caused by clock jitter (particularly if you have a crappy clock, bad cable connections or other poorly setup configuration).

In the real-world systems which 99% the rest of us have it simply won't matter.

For a live sound setup, if no recording is being made with the same system (as in a SAC/SAW link), I wouldn't worry.

However if one is also recording with the system, and the recording might be for serious use, we don't know that the recording WON'T be heard with a serious and accurate system somewhere else. I can usually get along with only the ADAT connection carrying the clock, but if one has the luxury of taking the extra trouble, it doesn't hurt to be sure. I've done it both ways, depending on system and circumstance.

In my current portable recording system, where my small rack has a Presonus Digimax FS feeding a Focusrite Scarlett 18i8 using ADAT optical, I use the internal clock of the Digimax FS as master. Yes, I have to set the clock rate in the Digimax, the Scarlett and in SawStudio, but I go to the extra trouble and I'm happy to do so. This has worked well for me.

[cgrafx]

For a live sound setup, if no recording is being made with the same system (as in a SAC/SAW link), I wouldn't worry.

However if one is also recording with the system, and the recording might be for serious use, we don't know that the recording WON'T be heard with a serious and accurate system somewhere else. I can usually get along with only the ADAT connection carrying the clock, but if one has the luxury of taking the extra trouble, it doesn't hurt to be sure. I've done it both ways, depending on system and circumstance.

In my current portable recording system, where my small rack has a Presonus Digimax FS feeding a Focusrite Scarlett 18i8 using ADAT optical, I use the internal clock of the Digimax FS as master. Yes, I have to set the clock rate in the Digimax, the Scarlett and in SawStudio, but I go to the extra trouble and I'm happy to do so. This has worked well for me.

The question is does it actually make any difference. My highly educated guess is no, but you've spent the effort to setup your system that way, so if it works for you that's great.

I prefer the added flexibility I get with a clock that can be controlled from within SAC/SAW.

I'm 100% certain my recordings would not sound any better using an externally connected COAX clock and I have 4 ADA8200 interfaces, so I can't let them clock themselves.

Acoustic treatment, proper mic placement, and quality musicianship play an infinitely more important role.

Again in the context of the question asked, there is no need for mr_es335 to worry about shoe-horning in a space for the daughter RME card just so he can connect an external word clock to his ADAxxxx.

Also in the context of the question asked, my opinion and that of a lot of others (including Bob L) is to set the RME card as the master clock source and let the software control the clock speed.

[Cary B. Cornett]

Acoustic treatment, proper mic placement, and quality musicianship play an infinitely more important role. I completely agree about that.


Again in the context of the question asked, there is no need for mr_es335 to worry about shoe-horning in a space for the daughter RME card just so he can connect an external word clock to his ADAxxxx. Yes, for what I understand he is doing, he should not need separate word clock for his setup. If he is recording his gigs, though, he might want to make the converter box the clock master. Not a big difference, but small differences accumulate in the recording process.


Also in the context of the question asked, my opinion and that of a lot of others (including Bob L) is to set the RME card as the master clock source and let the software control the clock speed.
Which is, in fact, what I have done most of the time, and I did that at least some of the time when I used an RME interface.

[mr_es335]

Hello,

Thanks to both Philip and Carey for your helping me in getting this matter all sorted out.

Since I believe that 1) I will never have a need to change the clock settings from 44.1, and 2) since am I only planning on using a single ADA, I will 3) go with the ADA as "slave" and the RME as "Master".

So, thank you both again for your invaluable assistance. It is always very much appreciated.

[operationwhat]

lol like a BOSS :cool:
You're awesome!!!!

Sir, You've just did the opposite of what they've advised you to do lol

shortened down....
- use ADA clock if you use only 1 unit
- if multiple units used, then go for RME clock or
- use COAX from ADA1 Set as Master - into ADA2 - ADA3 - RME Card
why? Because, if you some problems with droputs accouring on a live gig ,
it's probably that clocks can't sync...

I can confirm that my rig is more stable with ADA's clock than with RME's,
why??? probably some few things but I don't care, it's Behringer ADA and a
discontinued RME card....pfffffffff

While using my Lynx Arora16 converter in the studio, the company advises you
to use the the internal clock of the converter box, than the one that is in the
PCI Lynx Aes16 Card, and all because of jitter.

Guess they know something.....

Cheers!

[mr_es335]

operationwhat,

Sir, You've just did the opposite of what they've advised you to do lol...Am I seriously missing something here? operationwhat, what in the world are you talking about?

Who are the "they" that you are referring to here? There has been two apparently opposite views presented here, one for Philip, and one from Cary.

From cgrafx:

Also in the context of the question asked, my opinion and that of a lot of others (including Bob L) is to set the RME card as the master clock source and let the software control the clock speed.

I gather that you - operationwhat, are saying that the ADA should be master and the RME the slave??

[Dave Labrecque]

Keep in mind the question relates to a $175 Behringer 8-channel AD/DA that is being used for live sound (thats just a little under $11/ch).

Wha... ? ;)