Hi Mr Katz. There's a statement in your book "mastering Audio" (great book and really appreciate your detailed and instructional views), where you say that a digital interface, even when clocked to a great great clock, will never outperform it's own, due to a separate process (pll), wich in a bad scenario, can only introduce artifacts to the sound. I believe that you also state one should clock externally only when recording thru external converters, and always clock internally when not. I assume that in a Hibrid Mixing stage, the benefits of great analog outboard gear will surpass the downfall of clocking externally. Would you elaborate a bit more on the topic, cause here in gearslutz, there's always been a popular consent, that a better external clock will improve your audio...if you have , let's say, a digi 002 and wish a Big Ben to make things better!?
Anyway. Thank you so much for stoping by and share your knowlwadge this month! Hope to have you around in the future.
Cheers....................Joaquin.

Hey, it's almost May... maybe there'll be another month :-). Others may of course chime in. My input is based on both technical measurements, theory, and listening. First of all, your statement above was a bit confusing. When you said "one should clock externally only when recording through external converters", it gets confusing. Let's rephrase that to, "an A/D converter PROBABLY will perform best when running on internal (crystal) clock" as opposed to being clocked externally." So, yes, PROBABLY, your A/D should be the master clock and your DAW should be clocked to the A/D. Notice how I avoided the term "externally" in that last sentence to avoid the confusion that you may have caused by referring to your DAW as "externally clocked" when I'm talking about the A/D as being internally clocked---which mean the same thing, of course.

OK, only measurements and very careful listening tests will confirm if your A/D performs better on internal or external sync. But I can say "PROBABLY" AND highly likely it will be better clocked interally. Apogee makes opposite claims about the Big Ben, and I have not seen a shred of objective measurement or evidence that would show it to be so. This is not voodoo, by the way, it is science. The ONLY WAY that an external clock can outperform an internal clock is:

---if the internal clock is inferior and poorly designed (which means the converter designer did a bad job). Because the clock signal goes through far more potentially degrading circuitry when it is generated external to the converter

Another good question is when you MUST clock externally, if an external clock can ever IMPROVE performance instead of degrading it. In my book I claimed that an external clock would always degrade performance compared to the intrinsic jitter of the PLL, and that is the common wisdom. However, another authority (in my mind), Eelco Grimm, recently pointed out to me that an external low-jitter clock with a very low jitter BELOW the corner frequency of the PLL can improve the low frequency jitter. BELOW the corner frequency, the external jitter dominates, above the corner frequency, the PLL's jitter dominates.

So, with an EXTREMELY low jitter external clock, and if the converter's PLL has a relatively high corner frequency, the low frequency jitter of the converter can be improved compared to other external clocks. A good converter, therefore, should have a PLL with a very low corner frequency. And the lower the corner frequency, the less likely that an external clock will improve and the more likely it will degrade the converter's performance. For example, Prism converter's have a corner frequency below 200 Hz while typical converters' PLLs are above 2 kHz! So it is highly likely that a very good converter like a Prism will not be affected at all, or possibly degrade no matter what external clock you feed it.

I'll be sure to include this addendum in my second edition.

-----


Now, what can you do if you do NOT have measurement equipment? This whole jitter thing is so subjective, isn't it? Well, you can be as objective as possible. Take a high quality source of stereo music (such as a 30 IPS analog tape, or an SACD). Feed it into your A/D and record that into your DAW. Transfer it twice, once on internal, once on external clock, and if you have several models of external clock, transfer it several times.

Now, line up all of the transfers on different tracks in your DAW, as closely as possible.

Then, with your DAW's DACs set to internal clock (which is likely to be most stable, and at least it will be consistent), solo back and forth between any pair of tracks. The track which has the widest image, most stable, with the most solid bass, purest, and warmest, is the one which represents the transfer with the least jitter. Make the comparisons blind if you can.

That's the best test you can do minus using measurement equipment, and if you perform that test with your DAW that carefully, and tell us about the results, then I'll believe you! And if you cite this testing method, then we will all benefit from some carefully-performed tests instead of the usual half-baked claims.

...cue another long, ill informed, debate about the big ben improving your convertors. I think we can file that in the same 'fact' file as UFOs, the Bermuda Triangle and The Loch Ness Monster.

PS -Bob, did Apogee ever send you that big ben to 'evaluate' like the promised a year or so ago.

Bob,

> Take a high quality source of stereo music (such as a 30 IPS analog tape, or an SACD). Feed it into your A/D and record that into your DAW. <

I always assumed that even the finest 30 IPS recorder will have jitter 100 times higher than any stock computer sound card - caused by the "stick-slip" behavior of analog tape as it passes over the playback head. Do you know if anyone has ever actually measured this, to equate "scrape flutter" to numbers that can be related to jitter?

--Ethan

Thank you very much for all the new info to ponder. I'm a long way from truly understanding the mechanics of the issue, but all of this expands the basics for further comprehension and informed decicson making. Thanks!

I'll do the test with the instruments and AD/DA that I have available, and a new converter (Rosetta800) that I hope to acquire in the near future.
Just to be in the clear, I wasn't stating any conviction, but only trying to have a better understanding of the topic in your book.
For the Test, I assume, that the AD/DA has to be a constant and the only thing changing has to be the clock...I'll post once I'm able to perform the tests.
Thank you again.........................Joaquin.

We have a variety of analog and digital gear that are commonly found in pro mastering studios. Most masters done here go out of our computer (DAC) through our analog gear and back in (ADC) to the same computer. Digital gear is inserted where needed.

This is a copy of a post that I put on the Brad Blackwell mastering site:

"We used an Apogee PSX-100 here for many years (yes, I know, not the best...). Two years ago, we heard that the Big Ben could improve the sound of our combo analog/digital studio.

We got the Big Ben in for trial and it took me about two minutes to decide that the sound was better when clocked by the BB.

Fast forward to yesterday. We replaced the PSX-100 with a round trip of Lavry Gold and of course decided to check out the claims that clocking to the ADC was the best way to go.

Our test was a quicky, not too extensive, but enough to realize that there is little difference, if any, between clocking from the Lavry ADC or the Big Ben. My assistant and I thought we could hear the difference, well, sometimes, so inconclusive.

So, we will probably sell the Big Ben and clock from the Lavry.

BTW, we were correct in our conclusion that the Big Ben improved the sound of the studio while using the PSX-100 but I'm sure it was b/c the PSX-100's clock was not as good as the Big Ben's. The PSX-100 has been around for almost a decade(?) and clocking technology has obviously improved since then.

I remember a few months ago posting a reply on this webboard stating that an external clock would help to achieve a better sound in a studio with various pieces of digital gear. Obviously I was not in possession of all the facts b/c my observations were based only on my specific situation! This is a good example of generalising a situation and believing it to be the gospel for all situations."

So, it appears that Bob, and many others, including Dan Lavry, are correct, but only if you have a well-designed clock in your ADC.

Andy,

Silverbirch Productions

No. One of their dealers was going to if they did not, but I forgot who it was.

Makes sense. But be sure to blame the right equipment, don't shoot the messenger. Those who are thrilled with their external clocks should be depressed instead by their converters!

A properly designed $25 crystal oscillator, proper power supply and grounding INSIDE a converter can beat the pants off of a $1000 external clock. One is a cure, the other is a bandaid. Guess which one?

While analog flutter and digital jitter are related, the degree and main frequencies of analog flutter and its affect on perception are very different from the frequencies of jitter. So one won't mask the other. When in doubt, use an SACD as a source :-).

Bob,

Good point about the different jitter frequencies. Thanks.

--Ethan

Far be it from me to go toe to toe with Mr. Katz so I submit humbly that.... if the Band-Aid stops the bleeding and that is all you have in your med kit.... well.....

heh

What I mean is if the external clock does make the poorly designed converter sound better then, isn't that the point? I mean I completely agree, make sure you lay credit and blame at the feet of the proper gear but in the end does it really matter if the finished recording sounds better?

Let's say a guy or gal already has an older Apogee AD-16 and a DA-16 (not an AD-16X or a DA-16X). They can take a huge leap in $ to upgrade for 16 channels of I/O in the X range (something like $5000 when I last looked) to get... oh off the cuff, a 30% improvement in the finished recording. If the Big Ben gives them a 25% improvement for $1200 or so who cares were to place the blame right? They upgraded and got better sound out of the poorly designed older units.

Well, having recently turned my Digi 002 into a dongle by using both UA 2192 and Aurora 8 converters, I wondered about the clocking situation.

So I asked Bob Katz whether clocking the Aurora 8 from the UA 2192 (that is therefore the masterclock) via wordclock cable is the right thing to do when both converters will be in use.
For drum tracking, I will use the 2192 for say OHs and the Aurora 8 for the rest of the kit.


My question:
I recently bought UA 2192 and Aurora 8 converters. So far I clocked the Digi 002 I'm recording to from the 2192. The Aurora also got the clock from the 2192 via Wordclock cable. I'm often using both the 2192 and the Aurora for say drums. Is it necessary to clock the Aurora from the 2192 or is this unrelated to the 'external clock' debate?

Bob's answer:
It's related. If you're using both converters at once, one of the two converters must be externally clocked. You have to test and/or measure to see what's the least compromise.

Sure, if it's an economicallly good decision, even if it's not a scientifically-wise one :-).

Well-designed converters don't need external clock bandaids, but they cost a lot more. And they sound a whole lot better. But do they sound 10 thousand dollars better? I doubt it. There is a price/performance curve.

But for only 2-4 k$ more you can get a superior converter, while not the best one, it will be considerably higher on the performance/sound curve than a cheap converter outboarded with an external clock.

In 2-channel, take a look at Benchmark, for example. Jitter is one of the buzz words over there, and the new ADC-1 performs IDENTICALLY on internal and external sync. Do they make an 8-channel model?

I just did a shootout (on internal clock) between my HEDD-192 and the Benchmark ADC-1. It's a tie! The two sound identical to me (get that... why is it so difficult for someone to dare to write they could not hear a difference between two components?).

So that means that the Benchmark is at least as good a 2-channel A/D converter as one of the most highly-regarded A/D converters around, the Cranesong HEDD-192.

Other solutions, like Mytek, for example, may not perform equally as well on external clock compared to XTAL, but they will probably beat soundwise and performance-wise on internal clock what you can get with a "cheap model", hotrodded with external clock or otherwise. If I had the time, I'd be doing these kinds of test and sound reports for a good magazine, but I don't have the time. However, from time to time I do shoot out 2-channel converters over here.

BK

Thanks Bob.

I have an AD-16X and a DA-16X, I stepped up to the plate and got the best I could afford and so far I am very happy (so much so that I sold off my HEDD 192, not that is was not better or worse it was just redundant).

I guess the point is that I have heard too many people say that the Big Ben has helped their system. Could be "I paid x $ for this thing so it has to sound better right?" syndrome but I think some people with lower end converters do get some benefit from the thing regardless of the reason. In the end if it helps them make better recordings, good for them I guess.

I do see your point, get a better converter and you can skip the external clocking but some folks just don't have that kind of cash and at that point the Big Ben might make more economical sense for them.

Hi Bob,

Good to see your comment on this issue in a straight forward manner.
When I first came out with the claim that as a rule, using internal clock is a better way to go, I got a lot of resistance from people that sell clocks.

I agree with your observation that it is very unlikely (though possible) to have an internal clock operate worse then external. What are the chances to have external better then internal? Not very high. Why?

The task of designing a fixed internal clock is VERY MUCH SIMPLER AND EASIER then getting the same performance out of the circuitry used by external clock -
VCXO (not a fixed clock but a variable one) plus a phase lock loop circuit. Add to it the fact that you send a clock over a cable, between chassis of different grounds....

The whole concept of having an external clock improve conversion is nuts. The external clock box gets to send clock signals to the AD, but the AD does not communicate with the external clock.

So the clock box DOES NOT KNOW what the AD is doing. It does not know if the AD is near perfect, or if it has 60Hz jitter, or some radio frequency related jitter, or data induced jitter...

Since the External clock box DOES NOT KNOW what the AD is doing (it does not even know if the AD is on), it can not send any "special" signals to improve the AD. It is analogous to a DR. trying to treat you without ever seeing you or talking to you.

That alone means that the best you can ask from a clock box, is to supply a clean clock. So why not take the same clean clock circuit and put it INSIDE THE AD CHASIS where it belongs. It sure beats having to have the same clean clock in another box, and then have the signal exposed to a cable, and one more set of circuits as mentioned above.

Of course there are times when one needs to use external clocks, such as when using many channels - more then one chassis full. But that is not an improvement. It is a tradeoff! A little more jitter for the ability to operate many channels.

Regards
Dan Lavry
http://www.lavryengineering.com

Hi Dan and Bob. Thanks for sharing your knowledge here.

What happens to the clock quality when a digital interface to a DAW is slaved to the clock of external AD converters?

For ex. - a digital interface like the 192 receiving external clock from the external AD?

From what I've read here, even the clock of a digital interface should be compromised when "sync'd" to external clock.

Correct?

About my question above - the reason I ask is that the digital interface would be receiving audio digitally from the external AD... so I'm confused about whether or not it's better to leave to external AD set to its own internal clock - or have it receive clock from the interface.

It seems that the external AD's clock would suffer if slaved to the clock of the interface and vice versa.

Which clock should be set to internal, and why?

Thanks again.

Sort of correct. But entirely unimportant! It's important to realize that once a signal has been converted from analog to digital, jitter is unimportant. Only at the conversion stages, the interfaces between analog and digital, does jitter become important. All digital processing, for example, is not at all affected by jitter. Because digital processing is performed according to the VALUE of the samples, not according to their intersample timing. There is one other exception, and that is asynchronous digital sample rate conversion, which can be affected by jitter, leading to more distortion.

There can only be one MASTER in any digital system. All the rest have to be slaves. So, in general, choose the converter to be the master clock and slave all the digital devices (including the DAW) to that clock. In the case above, the ADC should get set to internal clock, and the DAW (through its digital interface) on external sync.

Max Gutnik, Apgogee's Sales Manager, said publicly that he would send you one for testing. I'm surprised that he did not.

I'm not heh

Does sonic studio solution enable an asycrous conversion on the same Daw ( playback and rec )?? if using an external finalizer as a lavry Le3000S that convert from 96 to 44 ?? what happen to the clock ?? Does it real need another daw to record it ?? if not ..the daw should be locked on the last conversion clock ??

Thx

1) No. You'd need two DAWs to do it with Sonic. Sequoia can do it with two instances and two interfaces. And you'd need the external hardware box to have one playback and the other record.

Personally, it's easier for me to get a handle on this one with two DAWs and a KVM switch.

2) "Finalizer?" Oh, I guess you mean the SRC facilities in the Finalizer. You definitely need two DAWs or at least two interfaces in the same computer. The box takes in 96 and puts out 44. It doesn't matter if the technology of the box is synchronous or asynchronous---either way you need to capture in a separate DAW or interface.

my english isnt good enough sorry for my translation

Thanks !! -)

I think there is some confusion re "Finalizer". I believe the poster was referring to the Lavry L3000S Digital Optimizer. On the lavry website, it says Digital Finalizer so it must have been called that at one time.

In fact, I just bought one but can't use it b/c I play and record in software (Pro Tools) that only works at one sample rate per session.

So while I've got you, do you, or anyone, know of a sound card for a Mac G5 that can accept 2 digital inputs so that I can run Spectra Foo and the Lavry L3000S SRC off it at the same time?

So that means that I could be running PTs at 88.2 out of the HD unit into another digital sound card that would be controlled by another piece of software running at 44.1k. This would avoid a second computer, if this is possible.

And lastly, I gather I would still be able to clock the whole setup off my Lavry ADC?

Thanks,

Andy,

Silverbirch Productions.

Hi,

The Model 3000S can be used many ways. It is more then a sample rate conversion. As a SRC, it may be worth the trouble to do the final SRC with the 3000S, especially when you need a 2:1 SYNCHRONOUS down sampling.

Also, one of the LE3000S best features is the Acoustic Bit Correction. That feature can "run at the same rate", or it can be combined with an SRC operation, and it is very useful for the final step - reducing from say 24 bits to 16 for a cd. There are 4 noise shaping curves, and for classical music and acoustic instruments I recommend NS2. For heavy metal NS4.

The Model 3000S is also a great meter, just like the Gold AD. But while the Gold AD can show you what is going into the DAW (before processing), the 3000S can show you what is coming out of the DAW (after processing).

Regards
Dan Lavry

We would leave the 3000 in front of the Sonic so it always saw clock no matter what we were running the rest of the room at. Different source rates or processing rates or even turning things on and off didn't matter since Sonic always saw the 3000. At the time I was running a Pro Tools TDM system and Sonic HD on the same computer, each with their own cards and I/O boxes, each running independently of each other; quite convenient and surprisingly stable.

Now I'm running a 002 rack more often as source (still run a PTHD in the mastering room occasionally), and it's less convenient with Sonic HD which never got out of OS9 (so I need a 2nd Mac), but fine with newer Sonic products that are OSX. Each DAW's I/O works through firewire. I haven't put it through its paces too much yet, but it seems to be workable. The other thing with new Sonic is that you can (or reportedly will be able to once beta is through) run multiple instances on the same machine concurrently, so you can play out of one into the other and be editing on a third. This I haven't tried or seen yet, but it seems like a good idea. On a G5, you could use the built in toslink for one of them. There are a variety of ways to set up a workflow you like.

I think the point is that in most professional situations where multiple digital devices are run simultaneously, there can be only one master. With Big Ben performing this function, correct cables and termination, it provides an eloquent solution. This solution is far superior to daisy chaining multiple devices. Even if you were to slave Big Ben to one of the "superior" converters, it would still be a better solution than daisy chaining.

While the Apogee X series are easily in the catagory of the converters mentioned here
(the sound as good or better than my HEDD, TC6000 or Troisi's) a great clock is essential in situations like scoring or post. I recently completed a 2 hour special for network broadcast.
Because I had Big Ben, I was able to slave Beta SP and Ben to black burst video ref and still send word clock (a better reference) to my converters, via Big Ben.
If someone is running a small rig where they don't have to run multiple devices and clock propogation is not a risk, then any of the above scenarios will be fine.
When running multiple digital devices, Big Ben provides a variety of simple solutions to complex problems and still maintains sonic integrity.
If someone is using Lavry, Prism, HEDD's or Apogee X, they can still pick their favorite clock to be the master.
I won't beat a dead horse any more but it is important to address the needs of larger setups and growing studios and Big Ben does just that. I know this isn't theoretical, it's practical and some of us need practical (users) while designers are entirely free to think out of the box. I would imagine that's how Big Ben came into being.

Respectfully,

chap

I don't know much theory, but I have found Big Ben to give audible improvements each time I have tried it in a system, primarily 192's, Rosetta 800's, AD8000, etc. I haven't tried Big Ben with a Lavry or a Prism but I haven't seen too many of them around.

I find Apogee not explaining why / how Big Ben works as it does very interesting. Maybe they know something they aren't telling the rest of us? I would also be surprised if Apogee ever sent out units to be evaluated by a competitor's advocates. That's a losing game.

You can't "hear" theory, but you can try various products yourself to hear what sounds best to you. If you don't think it improves your sound, don't use it.

Just my opinion, ymmv.

Best....H

IIRC, Fletcher didn't like what the BB did to RADAR.

Similarly, I don't enjoy what it does to SADiE. But that's no problem as the BB will happily source clock from SADiE and feed everything else. The ergonomics of the BB are pretty good.

Haigbabe