Is it better spending on a Hi-End Master Clock rather than on Hi-End Converters?

[Panos K]

Dear Sweetwaters hello!

I'm currently building a new room capable of 16 input rec and 64out playback with Nuendo 4.

So I'm trying to figure out the most cost effective hi-end converter/master clock combo available with the best performance/price ratio.

I'm wondering what would give me the best performance for my bucks: getting a more expencive clock and somehow "cheaper" converters, or buy the best possible converters I can with a relatively good clock?

For my clocking needs, I'm anyways thinking to buy an Antelope Audio Isochrone OCX or OCX-V, and a mix of Apogee, Lynx, Mytek, Lavry and Cranesong Converters.

I'm strongly considering to also buy Antelope's Atomic clock 10M, for which I've read good reviews and which is the perfect companion for OCX or OCX-V to upgrade their clocking. This unit is fairly expensive though.

So my question is: Should I buy both Antelope OCX and Atomic clock and then get Apogee and Lynx Converters (which are more cost-effective), or should I just buy OCX (without the Atomic Clock) and get the more pricey Lavry's and Mytek's?

Another issue is, is mixing converter brands recommended? Do you think they would all work together with no sync or quality issues, all slaved to the master clock? Or should I buy all of the same brand?

Let me tell you the initial setup I had in mind, and I would appreciate your input on that:

-(4x) Lynx AES16e PCIe cards running on a PC with Nuendo 4
-(3x) Apogee DA-16x
-(1x) Mytek 8x192
- 8-ch Lavry Blue A/D
- 8-ch Lavry Blue D/A
- Cranesong HEDD 192 (2ch A/D-D/A)
-TC Electronic System 6000

for a total of 16ins/64outs plus 2-ch Cranesong I/O. Total of 12 digital devices.

I was thinking of Antelope OCX or OCX-V (for a possible video upgrade of my studio), and possibly an Atomic clock if my budget permits.
But I don't know if the OCX could send clock to 12 devices simultaneously as it only has 8 clock outs. Do you think I should daisy-chain some of the devices, or would that not be recommended? (would daisy-chaining reduce sound quality or introduce any jitter?) So would I also need an additional clock distributor, or that would be a waste of money?

Last thing is that, If I try to "squeeze" Atomic Clock 10M on the budget, then I would probably have to eliminate the 2 Lavry Units and possibly also the Mytek, and buy an Aurora16 instead, so I can get the same number of I/O's. Would that be a better bet on quality, or not?

I'll appreciate any of your inputs

[DAS]

As a general rule better converters will benefit less from better clocks. I'd start there (at the converter). All of the converters you mention are good, and one is not necessarily "better" than another because it costs more. The differences are largely subjective and usually rather small. I don't quite follow how you plan to hook all that stuff up and what you are trying to do, which makes giving helpful answers harder. But with all of that stuff on line at the same time it is likely you will need a master clock anyway. As for daisy chaining, it is never ideal, but you can get away with it. A better solution is a second "master" clock.

[Panos K]

Hello DAS!

The whole idea behind mixing up the converters is that I could have a little flavour of everything and use it on the right tracks, on the right instruments, sounds, or even the right styles of music. All I know I need is 16inputs to record simultaneously, 64 outputs to mix simultaneously (using a 64 track SSL X-rack as a summing box), and a 2-track higher quality mastering converter.

So I thought, instead of getting only one brand of converters, to get several with the same I/O's and sync them together. I don't know if it's as good idea as it sounds, as I don't really know if there are gonna be any side effects of such a mixed setup.

What I don't know neither, is that if there's gonna be any difference at all in the performance of the converters if I put an OCX clock to the game, and even more, if I put an Atomic Clock 10M to the game as well. Maybe then I could buy Auroras or Apogees and have them sound like Lavry's (completely transparent) or such a thing would never happen?

Another issue I'd like to address, is the connection of the aforementined setup. What would you connect with what? How many word clocks out would I need totally from my word clock? Would I need to buy a re-distributor too? (such as Isochrone DA?)

Thanks.

[DAS]

Ahhh, I see. It's an interesting idea. I mean, those boxes do sound different from each other, but of all the elements in the recording chain the character that a given converter imparts on the sound is one of the more subtle things. Your SSL rack (as compared to any of a number of other summing solutions) will have much more impact on the sound than the difference any of the mentioned converters would, by far. And that's before we start talking about microphones, preamps, etc. But I do know engineers who have different converters they use for different things. The idea you have is valid if you have the budget for it. On to your other questions...

Most converters I've heard react to external clocking enough that it can be heard. The amount of difference varies, but the important point is that the external clock isn't always an improvement. It's subjective though, so I want to be clear that I'm not saying it won't be an improvement. It just...depends.

As I mentioned before I'd lean towards getting the best converter (for me) and then worry about clocking as needed. It's a fallacy to think that clocking a lesser converter (no matter how good the clock) will make it sound comparable to a better one. For starters there is much more to making a converter sound good than the clock. I think you need to get rid of that idea and concentrate first on the best signal path (converter) for you, which doesn't necessarily mean the "best" (most $$) converter.

Having said that...if you really want to be able to use several of those units simultaneously you're going to need to clock them to a common source anyway, so you're right back to the master clock idea. Getting to your hook up question...you would run a word clock cable from the master clock to each converter. If you end up with more converters than your master clock has outputs we'll have to look at the best solution for your specific set up and combination of devices. There isn't really an all in one answer to this. Again, it depends. You may not even have to use all of them at one time.

There are enough other questions hidden in here that I think it would be a good idea to contact your Sweetwater sales engineer and talk this stuff through before you go too much further.

[Panos K]

Thank you for your help, things start getting more clear in my head now.
Let me just keep it simple and tell you the minimal setup I have in mind:

-4 Lynx AES16e
-4 Lynx Aurora16
-1 CraneSong Hedd-192
-1 Cranesong Avocet (the monitoring solution, it contains a D/A which I'm not sure If I have to connect it to the master clock or not)
-1 TC Electronic System 6000 (the same with this)
-1 Antelope OCX (with or without Atomic Clock 10M, not sure yet)

So how many word clocks would you think I should have in my master clock? Do you think I would need an additional clock redistributor such as the Isochrone DA?

[DAS]

I would clock all of that from one master clock. The OCX has enough outputs to do it.

[Panos K]

So my total devices are 11. Is there enough outputs on the OCX to connect all my gear? How would you connect them?

[DAS]

You probably should talk to a sales engineer here to get into those specifics. The pci cards don't require a word clock connection. The fact that you didn't seem to know that, while not a problem at all, is somewhat troubling given the amount of money you're talking about spending. It sort of feels like you're trying to design a very expensive system using nothing but posts on a forum. I don't think this is a good idea at all. One of our guys can help you sort through all these details and others to be sure to get exactly the right stuff that will work. It's a free phone call. Give us a call.

[michaelhoddy]

You probably should talk to a sales engineer here to get into those specifics. The pci cards don't require a word clock connection. The fact that you didn't seem to know that, while not a problem at all, is somewhat troubling given the amount of money you're talking about spending. It sort of feels like you're trying to design a very expensive system using nothing but posts on a forum. I don't think this is a good idea at all. One of our guys can help you sort through all these details and others to be sure to get exactly the right stuff that will work. It's a free phone call. Give us a call.

Yes, yes, yes. What he said. I mean, you really need 64 simultaneous inputs and you're running a native system? What are you doing, recording orchestras for film score?

[jmit01]

It was said earlier by DAS that "As a general rule better converters will benefit less from better clocks."

So, based on that general rule, what constitutes a "better converter ?"

The converters I see mentioned frequently are from companies such as Lavry, Apogee, Weiss, etc. I specifically have an Alesis HD24-XR, but really have no basis for comparison yet for the quality of the Alesis converters, other than a few opinions thrown around here and there in the various forums.

I guess my question is, are the XR converters considered "better converters," and how much will they benefit from external clocking ? Will a $1500 Antelope OCX yield a $1500 improvement ?
Will a $1400 Big Ben yield a $1400 improvement ?

[EC_Beast]

It was said earlier by DAS that "As a general rule better converters will benefit less from better clocks."

So, based on that general rule, what constitutes a "better converter ?"

The converters I see mentioned frequently are from companies such as Lavry, Apogee, Weiss, etc. I specifically have an Alesis HD24-XR, but really have no basis for comparison yet for the quality of the Alesis converters, other than a few opinions thrown around here and there in the various forums.

I guess my question is, are the XR converters considered "better converters," and how much will they benefit from external clocking ? Will a $1500 Antelope OCX yield a $1500 improvement ?
Will a $1400 Big Ben yield a $1400 improvement ?

A clock simply locks all digital equipment to the same time code. This makes sure that all of their clocks are lined up together so you will not get clicks, pops and other weird effects. With that being said...high end converters often have very good clocks built into them, and often have such clean conversions that they either do not need an external clock, or do not greatly benefit from external clocking.

The answer also depends on how complicated your chain is...if there's a lot of stuff going on, and things being routed all over hell and back.

[jmit01]

EC_Beast,
My application/question is not about using external clocks to synchronize multiple digital devices, rather, it's about whether to use a "better" external clock as a substitute for a potentially lesser-quality internal clock, in the hope of improving the AD conversion process.

My signal chain is nothing out of the ordinary: microphone/preamp/recorder

In my specific situation, I'm wondering how good the AD converters are in the Alesis HD24-XR, and ultimately if my recording chain would be improved by the addition of a good external clock.

[EC_Beast]

EC_Beast,
My application/question is not about using external clocks to synchronize multiple digital devices, rather, it's about whether to use a "better" external clock as a substitute for a potentially lesser-quality internal clock, in the hope of improving the AD conversion process.

My signal chain is nothing out of the ordinary: microphone/preamp/recorder

In my specific situation, I'm wondering how good the AD converters are in the Alesis HD24-XR, and ultimately if my recording chain would be improved by the addition of a good external clock.

The differences between say a Big Ben and an Antelope are really nominal, if noticeable at all. You'll notice a difference from the new clocking, I'd imagine, but it's impossible to say if it's "$1,400" better.

[DAS]

Will a $1500 Antelope OCX yield a $1500 improvement ?
Will a $1400 Big Ben yield a $1400 improvement ?

These questions really are impossible to answer. It's subjective. You will hear a difference, but you'll have to determine if that difference is worth the money. Maybe there are some other users up here that can characterize it for you somewhat. Still...with things this subjective about the only way to know for sure is to try it yourself. What I can tell you is that we've sold literally hundreds of Big Ben's to users for the specific purpose of making their existing converters sound better. If it didn't do something worthwhile those sales wouldn't stick.

[ELP71]

A clock simply locks all digital equipment to the same time code. This makes sure that all of their clocks are lined up together so you will not get clicks, pops and other weird effects. With that being said...high end converters often have very good clocks built into them, and often have such clean conversions that they either do not need an external clock, or do not greatly benefit from external clocking.

The answer also depends on how complicated your chain is...if there's a lot of stuff going on, and things being routed all over hell and back.

More can happen with superior clocking than just elimination of artifacts (or the potential thereof). Improved stereo image, 'tightening' of bass, 'opening up' of high frequencies can be achieved via improved clocking, regardless of converters. In fact, some folks specifically choose the Big Ben over on OCX because of the frequencies it 'improves' or 'hypes'.

[jmit01]

... In fact, some folks specifically choose the Big Ben over on OCX because of the frequencies it 'improves' or 'hypes'.


OK, that was an interesting statement...

All clocks do is generate and transmit a string of alternating zeros and ones...hopefully at very near a 50% duty cycle. So how exactly would one clock influence certain frequencies over others via AD converters ?

I can possibly see how random clock instabilities might influence one frequency at one time, then another frequency at another time, and I could also imagine how non-random instabilities would effect overall pitch, with respect to playback on other digital devices. But I can't quite figure out how one clock would improve or hype frequencies differently from another clock ???

Please help me understand this one.

[DAS]

Clocks do not transmit a series of one's and zeros. They transmit square waves at a given frequency (this is potentially important, the signal that comes out of a clock is analog, not digital), and those are used by the receiving device to drive a PLL (in almost every case), which in turn clocks the unit in question.

Jitter, which is what we're in principle trying to reduce with good clocking, comes in many subtly different forms, and some of them do have specific frequency characteristics that can manifest as sidebands in the audio program.

Those sidebands can stand on their own as content added to the audio and they can potentially reinforce or cancel other content to a certain degree. Having said that, this would be virtually impossible to hear in a tangible way unless really taken to an extreme in a given clock. But, I can see how a listener might feel that certain elements (I don't know about "frequencies" per se) are subtly accentuated (or not) based on introducing or reducing certain forms and frequencies of jitter.

It has been argued that some clocks may not reduce jitter at all (certainly only certain types can be reduced), or only reduce it within certain frequency ranges (jitter frequency, not audio frequency), and in fact utilize jitter added to the signal to produce a euphoric effect that is subjectively desirable. To do either: reduce or add jitter, a designer would have to have a very good working knowledge of the PLL in the intended receiving device, which is tricky because they don't all have exactly the same characteristics. It's possible that a clock that might produce a pleasing effect on a MOTU interface could do nothing on a Digi interface (brands chosen at random, do not read anything into that). Or that the reactions could be different at different sampling rates.

[ELP71]

Clocks do not transmit a series of one's and zeros. They transmit square waves at a given frequency (this is potentially important, the signal that comes out of a clock is analog, not digital), and those are used by the receiving device to drive a PLL (in almost every case), which in turn clocks the unit in question.

Jitter, which is what we're in principle trying to reduce with good clocking, comes in many subtly different forms, and some of them do have specific frequency characteristics that can manifest as sidebands in the audio program.

Those sidebands can stand on their own as content added to the audio and they can potentially reinforce or cancel other content to a certain degree. Having said that, this would be virtually impossible to hear in a tangible way unless really taken to an extreme in a given clock. But, I can see how a listener might feel that certain elements (I don't know about "frequencies" per se) are subtly accentuated (or not) based on introducing or reducing certain forms and frequencies of jitter.

It has been argued that some clocks may not reduce jitter at all (certainly only certain types can be reduced), or only reduce it within certain frequency ranges (jitter frequency, not audio frequency), and in fact utilize jitter added to the signal to produce a euphoric effect that is subjectively desirable. To do either: reduce or add jitter, a designer would have to have a very good working knowledge of the PLL in the intended receiving device, which is tricky because they don't all have exactly the same characteristics. It's possible that a clock that might produce a pleasing effect on a MOTU interface could do nothing on a Digi interface (brands chosen at random, do not read anything into that). Or that the reactions could be different at different sampling rates.

...what he said.

In the end, engineers who can buy whatever they want in whatever quantity they wish A/B these clocking sources and utilize the best piece clock for the purpose at hand. They recognize the difference in timbre and choose accordingly...how nice would that be???

[jmit01]

Clocks do not transmit a series of one's and zeros. They transmit square waves at a given frequency (this is potentially important, the signal that comes out of a clock is analog, not digital), ...

Actually, they do transmit ones and zeros. Remember, a digital signal is simply a square-wave sequence of electrical (or optical) ones and zeros, switching between two voltage levels (e.g., +5 V and 0 V) representing the two states of Boolean values 1 and 0. Technically speaking, all digital signals are analog, but certainly not all analog signals are digital.


Jitter, which is what we're in principle trying to reduce with good clocking, comes in many subtly different forms, ...

I know of two forms of jitter in a clock signal: temporal and amplitudinal. Do you know of others ?
Temporal jitter affects the pulse (or square-wave) timing, unless the amplitude jitter is so extreme that logical ones are actually detected as zeros, or vice versa (refers to variations in the extinction ratio between the logic voltage (or optical) levels).
Since we know that temporal jitter in the time domain transforms to frequency jitter in the frequency domain, I go back to my question before about what effects might occur from random vs. non-random clock jitter. Intuitively, it would seem that non-random clock jitter would manifest as an overall pitch shift in the resulting audio tracks – as if recording at one sample rate then playing back at a slightly different rate. Am I right ?
Then, what about random jitter ? What effect might it have ? Seems to me it would cause brief and randomly occurring pitch shifts (noise) in the audio tracks. Right ?

So, unless I've overlooked something, we have two fundamental effects from clock instability (jittter); overall pitch shift and short, incidental pitch shifts manifested as noise. If, as you say, “jitter is added to the signal to produce a euphoric effect that is subjectively desirable,” how would addition of either of these jitter effects produce a more desirable recording ?

Not being confrontational...just trying to learn something.

Also, I’m not sure what you mean by the word “sidebands” in the context of jitter frequencies manifested in the audio program.

[DAS]

Regarding the first point (analog or digital), the discussion is now down to semantics I think. Obviously all digital signals are manifested as analog voltages, but it's important to make the distinction that with clock signals there really isn't any binary information being conveyed. There is no "information" at all being conveyed, other than raw timing info, and the square waves themselves convey that without any manifestation of ones and zeros. Yes a clock signal COULD be thought of as a series of alternating ones and zeros I suppose, but in actual implementation of external word clock signals it is not. It simply drives a PLL, which is an analog circuit.

By forms of jitter I am speaking to jitter that can be correlated in some ways (as opposed to completely random jitter, which is another type that may or may not manifest as noise). There are a number of things that can cause different forms of correlated jitter. But regardless of the causes it is true that when signals are modulated by other signals sidebands result. Yes, you are right to a degree that we're talking about short and long pitch shifts of various amplitudes here, but when those pitch shifts are in the form of consistent modulations, whether short or long, (see FM synthesis) sum and difference frequencies are produced that become part of the signal. Sometimes those show up in the audible range and sometimes not.

So with regard to jitter as pitch shift... (semantics) yes AND no: while technically true that jitter manifests as pitch shifts it is not necessarily true that it shows up to the listener as pitch shifts. Particularly when those pitch shifts become systematic (correlated) and are producing sidebands in the audible range. In those cases it shows up to the listener as additional content that interacts with the already existing (desired) content.

I hope this makes sense. It's not an easy subject to explain quickly, and unfortunately I just don't have time to provide all the background info.

[Rad]

Well, from some of the opinions, it would seem that there are people here who think that jitter mostly will manifest itself eiteher as noise or as pitch shift. Before I jump in, I just want to make sure we can all agree on some basics, because I think that statement just isn't correct, technically speaking.

What happens when there's timing irregularities in the clock pulse is several different things, of which one is just plain old distortion, but the rest is much nastier as it can involve aliasing of certain frequencies back into the audible range (the sidebands DAS is taling about). Distortion is relatively straightforward - just think of jitter as shifting the correct point where you want to sample the analog wave somewhat aside, so instead you detect (for example) a zero crossing where there really isn't one.
The aliasing thing is more complicated but the basic idea (if I remember correctly) is that during playback, higher-than Nyquist frequencies end up in the playback and are aliased back into the audible range in a non-harmonic fashion (but for details you'd have to look up a tech manual - I used to know this, but right now I'm busy with remembering other stuff).
I can see neither of these effects necessarily corresponding either to pitch shift or noise (because noise is random, and aliased frequencies aren't).

So, I'm a bit at a loss in this whole discussion about pitch shift and people going for Big Ben as opposed to OCX because of "frequencies it "hypes"".... man, this whole thing looks terribly misguided to me, and if there are folks out there really people buying Big Ben thinking that it will act as an EQ box, I just don't know how they'd survive the thought that this is, well, so to speak, wrong :) :smokin:

[DAS]

If you sample at one frequency and playback at another the pitch will be changed. Jitter does change the sampling frequency so technically it does introduce a pitch shift, but it is never (or rarely) ever characterized as a change in pitch shift because: A) The sample frequency would have to move a lot in order for us to perceive a change in pitch, and B) it would have to move to a new frequency and then stay around that frequency for a while...long enough for your pitch sensitivity to lock in on it. With typical forms of jitter neither A or B is true, so the listener (almost) never perceives anything going on with pitch. Instead we "hear" other characteristics.

[ELP71]

Well, from some of the opinions, it would seem that there are people here who think that jitter mostly will manifest itself eiteher as noise or as pitch shift. Before I jump in, I just want to make sure we can all agree on some basics, because I think that statement just isn't correct, technically speaking.

What happens when there's timing irregularities in the clock pulse is several different things, of which one is just plain old distortion, but the rest is much nastier as it can involve aliasing of certain frequencies back into the audible range (the sidebands DAS is taling about). Distortion is relatively straightforward - just think of jitter as shifting the correct point where you want to sample the analog wave somewhat aside, so instead you detect (for example) a zero crossing where there really isn't one.
The aliasing thing is more complicated but the basic idea (if I remember correctly) is that during playback, higher-than Nyquist frequencies end up in the playback and are aliased back into the audible range in a non-harmonic fashion (but for details you'd have to look up a tech manual - I used to know this, but right now I'm busy with remembering other stuff).
I can see neither of these effects necessarily corresponding either to pitch shift or noise (because noise is random, and aliased frequencies aren't).

So, I'm a bit at a loss in this whole discussion about pitch shift and people going for Big Ben as opposed to OCX because of "frequencies it "hypes"".... man, this whole thing looks terribly misguided to me, and if there are folks out there really people buying Big Ben thinking that it will act as an EQ box, I just don't know how they'd survive the thought that this is, well, so to speak, wrong :) :smokin:

I don't think people are buying a Big Ben over an OCX as an EQ box. What I was referring to was people who have both and use one or the other for specific program material and the effect the specific clock has on the mix.

Semantics & theories aside, if you clock with a Big Ben and then clock with an OCX and you percieve the mix as sounding different, then it sounds different.

[Rad]

I don't think people are buying a Big Ben over an OCX as an EQ box. What I was referring to was people who have both and use one or the other for specific program material and the effect the specific clock has on the mix.

Semantics & theories aside, if you clock with a Big Ben and then clock with an OCX and you percieve the mix as sounding different, then it sounds different.

I agree that good clocking can sometimes make an audible difference relative to bad clocking. But when it comes down to comparing two really good clocks like Big Ben and OCX... I just don't see enough variables to allow for dramatic sound differences. And when it comes to subtle changes, people start to hear all kinds of things that don't exist. I have experienced this myself so I'm a bit skeptical about such differences. In any case, even if you are able to hear a difference between the Big Ben and the Isochrone in the studio, I am more than willing to bet that it will get immediately wiped off by inaccuracies of the clock of any regular CD player when your CD hits the shelves.

[ELP71]

In any case, even if you are able to hear a difference between the Big Ben and the Isochrone in the studio, I am more than willing to bet that it will get immediately wiped off by inaccuracies of the clock of any regular CD player when your CD hits the shelves.

Probably accurate, and I tend to subscribe to this point of view. However the fact remains that in spite of our collective skepticism 'big time professionals' are A/Bing mixes with different clocks as I have described and coming to their own conclusions on what works for their mix.

[jmit01]

If you sample at one frequency and playback at another the pitch will be changed. Jitter does change the sampling frequency so technically it does introduce a pitch shift, but it is never (or rarely) ever characterized as a change in pitch shift because: A) The sample frequency would have to move a lot in order for us to perceive a change in pitch, and B) it would have to move to a new frequency and then stay around that frequency for a while...long enough for your pitch sensitivity to lock in on it. With typical forms of jitter neither A or B is true, so the listener (almost) never perceives anything going on with pitch. Instead we "hear" other characteristics.

Yes, your point B is what I referred to as the result of non-random, or systematic jitter.


analog or digital, the discussion is now down to semantics I think.
Semantics ? …yeah, probably. I think the word we’re both looking for is ‘binary.’ Clock signals are binary (two discrete levels). But clock signals do carry information…granted, it’s not in the format of bits/bites/words, but they certainly carry real-time info about the current clock cycle start time and duration.

I think I understand what you’re saying about signals being modulated by other signals – as in the way a carrier wave at one frequency carries (is modulated by) a much higher frequency signal (frequency modulation). Some kinds of noise are manifested in this way.

I certainly understand you don’t have time to go into this topic at depth, but it would be interesting to actually quantify some of the qualitative expressions being thrown around when describing the audible effects of clock instabilities; expressions such as “euphoric”, “muddy”, “highlighted” and “enhanced.”

One other question:
Do I understand correctly that if I use an external clock (into my HD24-XR), that it will completely override (substitute itself for) the internal clock ? In other words, the PLL circuitry wouldn’t just compare the existing internal clock to the external clock then make adjustments, does it ?
When some talk about improving their recording chain by addition of an external clock, it almost sounds like the external clock might serve to improve (not replace) the internal clock.

[jmit01]

Well, from some of the opinions, it would seem that there are people here who think that jitter mostly will manifest itself eiteher as noise or as pitch shift. Before I jump in, I just want to make sure we can all agree on some basics, because I think that statement just isn't correct, technically speaking.

Actually Rad, it wasn't a statement to be judged either correct or incorrect. Look back and you'll see my post was in the form of a question.
But yes, a pitch shift is equivalent to a frequency shift, technically speaking. And a clock timing instability (in the time domain) transforms to a frequency shift (or pitch) in the frequency domain - not necessarily an audible pitch shift, but a pitch variation nonetheless.

[DAS]

]One other question: [/FONT][/COLOR]
Do I understand correctly that if I use an external clock (into my HD24-XR), that it will completely override (substitute itself for) the internal clock ? In other words, the PLL circuitry wouldn’t just compare the existing internal clock to the external clock then make adjustments, does it ?
When some talk about improving their recording chain by addition of an external clock, it almost sounds like the external clock might serve to improve (not replace) the internal clock.

It sort of depends on what you define as the "clock." (danged semantics again). Arguably the PLL is part of the clock circuit in a device. An external clock doesn't replace this. It simply becomes the reference the PLL uses for timing rather that whatever the internal reference would have been.

[Rad]

Actually Rad, it wasn't a statement to be judged either correct or incorrect. Look back and you'll see my post was in the form of a question.
But yes, a pitch shift is equivalent to a frequency shift, technically speaking. And a clock timing instability (in the time domain) transforms to a frequency shift (or pitch) in the frequency domain - not necessarily an audible pitch shift, but a pitch variation nonetheless.

That's right. My point was that for this to become audible as a pitch shift, the clock should have periods where it runs consistently faster and periods where it runs consistently slower, and I think on pretty much all modern clocks this kind of variation happens much faster than audible limits.
Interestingly, however, I've heard of radio stations using deliberately overclocked processors that would run at say 44.3 instead of 44.1 so they can fit more songs in an hour. I guess this doesn't qualify as jitter though, even though I find it totally objectionable (much easier to just play one song fewer, they are paid per advertisement and not by the song...)

[jmit01]

That's right. My point was that for this to become audible as a pitch shift, the clock should have periods where it runs consistently faster and periods where it runs consistently slower, and I think on pretty much all modern clocks this kind of variation happens much faster than audible limits.
Interestingly, however, I've heard of radio stations using deliberately overclocked processors that would run at say 44.3 instead of 44.1 so they can fit more songs in an hour. I guess this doesn't qualify as jitter though, even though I find it totally objectionable (much easier to just play one song fewer, they are paid per advertisement and not by the song...)

Interesting. I'd never heard of radio stations doing that before, but it makes perfect sense. But then, I've also never heard of a radio station that thought it was even possible they could be running too many commercials ! ...but that's a whole other topic.

So Rad...let me get your opinion...

I'd like to record as accurate and realistic recordings as possible, given my budget and given the fact I'll be recording to an Alesis HD24-XR. I won't be using any other external digital devices that need clocking, like ADCs or other recorders, just a basic mic/preamp/recorder audio chain.
If you are me, do you add an external clock (such as a BigBen or OCX) ? ...or put that money into a big-screen TV and a couple cases of beer ?

[michaelhoddy]

"Big time professionals" are usually too busy mixing and doing work to swap clocks or whatever between projects or songs. Seriously, don't take this wrong, but the only people I personally know who have the time to get all worked up about this stuff don't have any work.

I'd like a show of hands: how many of you (besides DAS) have actually worked with your own two hands with the clocks in question here? I mean, you've done a record with them.

I'm all for a good master clock, I have one. But this thread is beginning to get silly- this stuff simply doesn't matter that much once you get past the entry point of getting a decent piece of gear. The heavens will not open up and a beam of light shine down on your mixes upon adding a certain clock to your rig. Anyone who says otherwise is selling snake oil.

My experience says that a good master clock has two main uses: getting a little extra mileage out of mediocre converters, and keeping everything synchronized in a studio with lots of digital gear. Past that, we're back to the age-old limit of good record-making being determined almost entirely by the respective talent levels of the artist and the engineer.

[michaelhoddy]

EC_Beast,
My application/question is not about using external clocks to synchronize multiple digital devices, rather, it's about whether to use a "better" external clock as a substitute for a potentially lesser-quality internal clock, in the hope of improving the AD conversion process.

My signal chain is nothing out of the ordinary: microphone/preamp/recorder

In my specific situation, I'm wondering how good the AD converters are in the Alesis HD24-XR, and ultimately if my recording chain would be improved by the addition of a good external clock.

Lemme answer this one, because I have some personal experience. Alesis converters and internal clocks in general fall squarely into the "mediocre" category, and every piece of Alesis gear I've hooked an external clock up to has noticeably benefitted from it, including the HD24.

But... If you're looking to spend $1500, I'd tell you to get a converter instead, it'll be a better investment, and you can clock the rest of the HD-24 from it anyway. One of the Lucid GenX clocks is perfectly adequate for your scenario for half the money, and they really sound quite good.

[jmit01]

"Big time professionals" are usually too busy mixing and doing work to swap clocks or whatever between projects or songs. Seriously, don't take this wrong, but the only people I personally know who have the time to get all worked up about this stuff don't have any work.

Yeah, well, we're not all "big time professionals" here that can afford to test drive several different clocks to see whether they make a difference (or not) on our "mediocre" recording equipment. So, in order for those of us unwashed masses who "don't have any work" to be able to learn - that is, to educate ourselves about this stuff - we have to "get all worked up" and start asking questions on a forum like this.

[michaelhoddy]

Have you used either of the master clocks in question? Have you used any decent master clock? (To keep the record straight, I hadn't even heard of the Antelope clock until this thread, not because it's not worth hearing, but because I don't have time to stay up on all this stuff). If not, your conversation is like the difference between talking about what Japan might be like while sitting in a Starbucks, versus getting on a plane and visiting Japan.

There's a popular military saying that goes something like this: "Amateurs talk about strategy. Professionals talk about logistics." In other words, professionals talk about getting the job done. Amateurs talk about all the different ways the job could be done.

What I'm NOT trying to do is dump on anyone for whom recording is a hobby. Your level of involvement does not give anyone a reason to look up at you or down on you. What I AM saying is that everyone could take a page from the lives of working pros, and make sure they're having the right conversation.

Anytime I see a clock/converter discussion devolve into discussion about jitter specs or square waves, or whether one clock makes the mix sound "warm" or "hi-fi" or whatever adjective, people are having the wrong conversation. It's that simple. I mean, did you ever buy a car because it was 4 inches longer than a competing car, or because it had 247 horsepower versus 243 horsepower?

Call me a jerk, but those kind of conversations are absolutely pointless.

[Joseph Hanna]

Yeah, well, we're not all "big time professionals" here that can afford to test drive several different clocks to see whether they make a difference (or not) on our "mediocre" recording equipment. So, in order for those of us unwashed masses who "don't have any work" to be able to learn - that is, to educate ourselves about this stuff - we have to "get all worked up" and start asking questions on a forum like this.

Yes but if one gets so worked up over mis-guided priorities it's easy to miss what really is important.

My two cents of course and YMMV but a discussion on how and why a particular EQ setting or compressor parameter rescued a crappy sounding guitar take is worth FAR more in the sonic landscape than endless splitting hairs on expensive clocks.

As some of you know I'm blessed with a great job that keeps me mixing 10 to 15 hours a day 6 and 7 days a week. I have a gaggle of people who run through and critique each and every 1/4 frame of audio I mix. Because of that scrutiny I'm on a constant hunt to up my game. To stay competitive keeps me employed. I'm always dropping by other engineers bays just to see what they're up to and to hopefully pick up a new trick or a cool new technique I hadn't thought of. Through it all however I can honestly say none of us have ever engaged in the level of clock/converter discussions as this thread has endeavored into.

Sometimes it's just talent and elbow grease that is the missing Holy Grail.

[ELP71]

"Big time professionals" are usually too busy mixing and doing work to swap clocks or whatever between projects or songs. Seriously, don't take this wrong, but the only people I personally know who have the time to get all worked up about this stuff don't have any work.

I'd like a show of hands: how many of you (besides DAS) have actually worked with your own two hands with the clocks in question here? I mean, you've done a record with them.

I'm all for a good master clock, I have one. But this thread is beginning to get silly- this stuff simply doesn't matter that much once you get past the entry point of getting a decent piece of gear. The heavens will not open up and a beam of light shine down on your mixes upon adding a certain clock to your rig. Anyone who says otherwise is selling snake oil.

My experience says that a good master clock has two main uses: getting a little extra mileage out of mediocre converters, and keeping everything synchronized in a studio with lots of digital gear. Past that, we're back to the age-old limit of good record-making being determined almost entirely by the respective talent levels of the artist and the engineer.

Who's worked up? I've used both the Big Ben and the OCX on sessions. I guess you could call them 'professional' since I was paid. 'Big time', not so sure :D I threw out that term because it added to the silliness of the discussion.

My point was, Nyquist theory/dither/jitter/pitter/skitter/PLL/ones and zeros aside, the deal is that 1) stable clocking makes a difference and 2) some guys actually do use different clocks for different sessions as strange as that may seem (from the horses mouth - these are not hobbyists - major labels, major artists). So to debate the issue based on square waves, etc seems silly to me and if you can hear a difference there is one and thats about all there is.

[Rad]

I'm totally with Michael and Joseph in that once you have a good quality clock, it really doesn't matter that much what brand it is as long as it does the job. I think that's what Michael was trying to say.

Now, to get to jmit01's original question,


I'd like to record as accurate and realistic recordings as possible, given my budget and given the fact I'll be recording to an Alesis HD24-XR. I won't be using any other external digital devices that need clocking, like ADCs or other recorders, just a basic mic/preamp/recorder audio chain.
If you are me, do you add an external clock (such as a BigBen or OCX) ? ...or put that money into a big-screen TV and a couple cases of beer ?
so here's what I think

If the Alesis HD24 is something you can't change, then it might benefit from a good clock, but it's not worth going as high as Big Ben for that (it's like putting an engine for a Porsche in a Honda Civic). I'd go with the genx as suggested by Michael.
But at this point, it also seems to me that if you're going to be spending 600-800$ on a clock it's much better for that money to get a decent PC-based interface and record on your PC instead. For that money you'll get all together a preamp, clock and convertor, and if you choose well, there are interfaces out there with very decent converters. For example, in such setups as you say I have successfully used the Macke Onyx series (which by the way uses the same AKM5385 convertors as found on the RME Fireface) with very good results. Most people who've used this convertor agree it has sufficiently good clocking (and I have the convertor spec sheets so I can also tell you it has 114 dB dynamic range and an impeccable linearity graph).

[jmit01]

Have you used either of the master clocks in question? Have you used any decent master clock? If not, your conversation is like the difference between talking about what Japan might be like while sitting in a Starbucks, versus getting on a plane and visiting Japan..

Actually, no, I haven't used either clock, nor do I have access to either clock until I shell out one-grand plus, hense, the very reason I'm here on this forum asking questions...to learn ! An audio salesman highly recommended the addition of a good external clock to my recorder. So, before I run out and spend money based on one person's recommendation, I'm doing my homework by asking those knowing more about audio than myself, what they think about external clocking to digital recorders.

I just can't figure why this disturbs you so !?!?!


There's a popular military saying that goes something like this: "Amateurs talk about strategy. Professionals talk about logistics." In other words, professionals talk about getting the job done. Amateurs talk about all the different ways the job could be done..

OK, so, your military saying notwithstanding, I'll just let you know that I AM a professional ! A professional electrical engineer (BSEE, MSEE), not just someone who CALLS themselves an engineer, but someone with 20+ years for a well-known research facility that rightly requires it's scientists and engineers to "talk about all the different ways the job could be done" prior to jumping in, and spending the company's millions on a project before adequate planning and before bringing on board those having substantial expertise in the supporting areas.
Now, scale that particular business model down to my personal location-recording business (not hobby), and I do believe it makes a lot of sense to go out and learn everything I can from those with more expertise than I have, BEFORE dipping into my personal budget for a new piece of equipment.

I gotta say, MichaelHoddy; you truly raise the art of condescension to a new level ! ...I'm impressed !!!




Call me a jerk, but those kind of conversations are absolutely pointless.

OK, you're a jerk ! ...but I'm not going away.
I'm going to continue with what you call my "absolutely pointless conversations" about square waves and jitter and whatever other topics I may need to educate myself before going out to spend my limited budget on something I may or may not need.

Suggestion: try spending a little more time helping a guy who's just trying to learn, and less time criticizing what you consider to be his "pointless" questions.

[Rad]

Allright, allright people, let's try not to be so bitter about this.

Of course comparisons like "if you use Big Ben, you'll get one kind of sound, but if you use an Antelope OCX, that will give you a different sound vs both the Big Ben and the Antelope rubidium clock" to me are meaningless, and for the reasons I explained, but none of them were made by jmit01.

I personally see nothing wrong with asking questions, this is what this forum is about, and I myself have learned quite a lot from it, including on difficult to fix studio situations.

[jmit01]

Now, to get to jmit01's original question,

so here's what I think

If the Alesis HD24 is something you can't change, then it might benefit from a good clock, but it's not worth going as high as Big Ben for that (it's like putting an engine for a Porsche in a Honda Civic). I'd go with the genx as suggested by Michael.
But at this point, it also seems to me that if you're going to be spending 600-800$ on a clock it's much better for that money to get a decent PC-based interface and record on your PC instead. For that money you'll get all together a preamp, clock and convertor, and if you choose well, there are interfaces out there with very decent converters. For example, in such setups as you say I have successfully used the Macke Onyx series (which by the way uses the same AKM5385 convertors as found on the RME Fireface) with very good results. Most people who've used this convertor agree it has sufficiently good clocking (and I have the convertor spec sheets so I can also tell you it has 114 dB dynamic range and an impeccable linearity graph).

Thanks Rad !
This is just the sort of advise I was looking for. From reading many of your posts in the past, you do seem to have quite a bit of experience/expertise in the area, so I appreciate your input.

The main reason I'm trying to put this particular configuration together with my existing (Alesis) recorder instead of directly to a PC, is because about 90% of what I do is on-location live recording. Rack mount equipment, such as the Alesis, and quick setup is a must for several reasons, mostly related to logistics. Because of that, my approach has to be get in, set up quickly, record the tracks, pack up then get out. All my editing, mixdown, mastering, etc. HAS to occur back at the studio - never on location.

I realize the Alesis is not the best setup that money can buy, but it's very practical for me for so many reasons.

[DAS]

Well, I'm not sure how much more there is to be gained in this discussion at this point.

Couple of thoughts: I generally agree with ELP71 that if you hear a difference then there is one for you (I don't necessarily agree that clocking always makes an audible difference though). Having noted this, I do sometimes feel that certain people on forums hide behind the "I can hear it" subjectivity in order to get around having a slightly more rigorous understanding of what is going on. Time and time again I see people reach conclusions based on what they hear when in fact their "tests" are flawed once all of the details become exposed. Further, I've learned over the years that quite a few people make claims about things they hear when they've in fact not actually done the hearing, but instead are repeating things they've heard (or read) others say or, in some cases, simply making it up based on some other personal bias or agenda they may have. I'm not suggesting any of that has taken place here, only that I do take people's claims of what they hear with a HUGE grain of salt generally.

jmit01 - I certainly understand your need to do your due diligence on this. Hopefully the discussion has been helpful. But at this point... I may be wrong, but I don't think there's much more concrete information we can convey here that will be more helpful in making a well grounded determination. The good news is if you buy something from us (and at this point I sure hope you would buy it from us) and it doesn't make a big enough improvement you're welcome to return it.

[michaelhoddy]

Suggestion: try spending a little more time helping a guy who's just trying to learn, and less time criticizing what you consider to be his "pointless" questions.

Jmit1: Read my other reply. I have used exactly the gear in question, and I answered your question here:


Lemme answer this one, because I have some personal experience. Alesis converters and internal clocks in general fall squarely into the "mediocre" category, and every piece of Alesis gear I've hooked an external clock up to has noticeably benefitted from it, including the HD24.

But... If you're looking to spend $1500, I'd tell you to get a converter instead, it'll be a better investment, and you can clock the rest of the HD-24 from it anyway. One of the Lucid GenX clocks is perfectly adequate for your scenario for half the money, and they really sound quite good.

Maybe you missed that?

I'm not being condescending. The internet is a horrible place for communicating the subtleties of speech so you have no way of seeing my intent as I write. It's certainly not to be condescending. I'm not trying to shoot you down or chase you out- that would be rather self-defeating for a forum, wouldn't it? However, your engineering background is a double-edged sword here- while it may give you an understanding of the science behind the technology that others may not have, it can also lead you away from the reality that recording is mostly a somewhat subjective art, not a hard science, and spec sheets and measurements don't even begin to tell the whole story. You will not be able to make this decision well based on spec sheets and internet conjecture. That's what I desperately want to help you avoid (simply, because I've owned a lot of gear I didn't need because of the way this thread was going).

Bottom line, as DAS says, get the gear in question in your hands (Sweetwater is a great place to do this), and see how it affects your process. If you don't see improvements, return it.

[jmit01]

Jmit1: Read my other reply. I have used exactly the gear in question, and I answered your question here:

Maybe you missed that?

No, I didn't miss that, and I do appreciate that particular piece of the discussion. And you certainly have my apologies if I've mis-characterized you from your other piece of the discussion.

Let me just say this...
As an engineer, it's never been good enough for me or any other scientist/engineer to stand up in a technical interchange meeting before department directors or customers, and simply tell them that "my design is better...meeting over !" That would simply never happen.
On the contrary, I would be compelled to defend such a subjective statement with an explanation of why exactly my design was better ?...what does my research show that would support such a statement ? ...what are the results of trade studies and testing ? ...what prototyping and test methodologies did I use to come up with the performance specs that I'm asking them to buy into ? ...etc.
My point is that at some point in the real world, the subjective must give way to the objective. The claims that "this thing is better than that thing" must be supported at some point with real evidence, and more times than not, that boils down to getting at that evidence through discussions down to the level of nuts and bolts, electrons and holes, and yes, this kind of jitter vs. that kind of jitter.

Once you arrive at some fundamental level of understanding, then you're better able with good certainty, to either accept or reject such subjective statements as

"yes, I can sense frequencies above 20kHz", or
"that recording is warmer because it has more harmonic distortion", or
"that recording is more airy because it was sampled at 192kHz".

I'm pretty sure I didn't make any subjective claims like that during any of this discussion...there's already way too much junk science out there that can't be supported with evidence !

[michaelhoddy]

I agree with you. I think the trouble is, that clocks in particular (and to a lesser extent, converters and mic preamps) are hard to nail down on objective terms like that. One person's "better" is different than another person's "better." Which is why boutique manufacturers in all these fields do good business.