Les horloges des cartes son.

Citation : il a l'air communement admis par quelques grands noms du monde audio que les convertisseurs fonctionnent mieux cloqués en interne à l'enregistrement[...]

Citation : et pourtant, il y a toujours quelqu'un pour dire que depuis que son converto est clocké à un Big Ben, le son est tellement mieux que blabla...donc ?

Citation : In the latest ed. of Tape Op Mag # 64 there is an article about Dan Lavry of Lavry Engineering. He more or less says to use internal clocks unless absolutely necessary to do otherwise . Quote " As a rule Internal clock yields less jitter,so use the internal clock when ever possible. There are times when one must use external clocks (such as the case of many channels) When doing so, the jitter goes up and the sound quality suffers a bit. That is a compromise we have to live with. Just do not fall into the false hype claiming that external clock improves your sound."

Citation : Here's what Dan Lavry had to say on another forum,

"Say you have a movie camera designed to take 100 picture frames per second, thus a frame every .01 second. You are pointing the camera at an ball moving from left to right at a constant speed, and you take a whole second of that motion, that is 100 picture frames. When you “play it back” with a movie projector that runs a picture each .01 second, everything would “be fine”, as intended.

But let us say that the camera is very “unsteady”. It takes a picture, then it waits .05 seconds, then it takes 5 frames within say .001sec, then a couple of frames at .01 second…. When you play it back, the projector “does not know” what happened at the camera, and the ball may like it is slowing at mid air for a brief time, then it zooms real fast some distance…. That is distortion due to timing error – time jitter. Note that if the camera was OK but the projector had jitter, that also would be a problem. So jitter counts at 2 places – at the camera (which is analogous to the AD), and at the projector (which is analogous to the DA). Jitter is important at the converters.

In one sense, jitter at the AD is more important, because once it takes place; it is in the signal forever. One can replace a bad DA with a good one, and that will eliminate the jitter issue of a poor DA, but what the AD does can not be undone.

Analogies can be misleading. In the case of movies, with enough still frames per second, the eye makes it looks like continues motion. In the case of conversion, it is the analog circuitry that takes the samples and makes them into a continues wave. But I chose the analogy of video, because audio and video (as well as many other applications) are fundamentally based on equal and precise time intervals. The time between each adjacent sample should be exactly the same, and if it is not, there is jitter, which will distort the outcome.

I said that jitter is important at the conversion. What about jitter in transferring say data from AD to a computer hard drive? The answer – it is not important, because we are just moving data from one place to another. You can move one frame every second, very slow indeed, or move a million frames a second, very fast. You can move half the data now, wait a while then move the rest of it… It does not matter, because you are not viewing it. But one you play it back you need the timing to be clocked precisely.

But some manufactures and sellers of clocks wanted to sell clocks, so they decided to convince the world that you need to clock everything. And with enough advertizing money, they where pretty successful doing just that. There are times when you need to use external clock box – when you want to have a lot of gear (AD channels) work together. But as long as you do not need external clock and you can use internal, use internal. It is not only cheaper, it is better!

What you need is the “best clock circuit you can make” that is very steady to be very near your AD circuit – short connection, good grounding… That is internal clock.

Say you take the same “best clock circuit you can make” and put it in another chassis. Will that be better? Not, it will be worse. You now have to deal with 2 chassis thus grounding issues. You have a cable that can pick interference, you have a cable termination imperfection, a cable driver, a cable receiver, and I did not even start… By the time your clock arrives from the clock box, it has so much jitter that it requires some “jitter cleaning circuitry” – typically a PLL circuit…. I pride myself for making very fine external clock circuitry, but no way can I make the external clock circuitry be as good as internal. Almost as good, yes, but never as good.

However, the clock BS’ers are still arguing that their external clock will improve the sound. There is some claim of a “proprietary clock signal” that will make things better. That is a crock if there ever was one! The clock box to the AD connection is a ONE WAY street. The clock “DOES NOT KNOW” what the AD is doing. What kind of a clock box signal is going to improve ALL the following an Ad's:

1. AD with a lot of jitter induce by 60Hz power line
2. AD with little jitter induced by 60Hz power line
3. AD with jitter induce from digital circuit noise
4. AD with jitter due to nearby radio transmitter
5. AD with jitter due to nearby power tools
6. AD with jitter induced from the digital audio data
7. AD with almost perfect timing
8. AD that is powered off…

This is analogous to a doctor that can cure all illness, doing so without any information about the patient…

One of the main offending marketing BS guys said that you can take a tone and have it sound better with jitter. You can alter a fixed tone with jitter, and can argue that you like it, or that you do not like it. But the alteration has to be deliberate for a specific constant tone (including fixed amplitude). You change the tone and the distortion changes... Jitter distortions are very complicated, and they are an INTERACTION between the clock timing AND THE MUSIC. The last I heard, music is not a constant fixed tone The simplest of jitter (random jitter) will increase your noise floor. More complex jitter makes for all sorts of undesired at frequencies that depend on the music, but at frequencies that are not musical harmonics, thus sound bad….

I first stated that internal clock is best a few years ago, and had to deal with a lot of attacks on a forum I was moderating. I insisted that the technical folks come in, instead of the marketing types, and sure enough, the technical types backed off after a short “fight” because they had no leg to stand on. A couple of years later, Digidesign wrote a paper about clocks, and they second me by saying that internal clocks are the best (when you can use internal clocks). I pointed that out and that brought about more attacks… The low jitter crock (I meant to say clock) goes on, and people are clocking with external clocks a lot of stuff they do not need to.

When your AD is using the internal clock for conversion, you are doing the best you can. The data sent forward to a computer, DAW or what not, is “after the conversion” so it does not need to be clocked with special care for jitter, and a “standard” link (say AES or SPDIF) is just fine for data transfer.

There are times when you need to use external clock, and when you need to, use external, when it is a needed trade off. For example, say you want to clock 2 or more AD chassis together... But other then that, internal is the better way to go!

Regards
Dan Lavry"

&

"First, about clock absolute accuracy:

If you view it from the point of view of pitch, clock accuracy of the cheapest crystal is more then good enough, because the ear can not hear 1 cent of pitch deviation, not even a sudden 1 cent pitch. A cent is around 1 part in a hundred between 2 notes on the tempered scale. So you end up with somewhere around 400 parts per million or so yielding a much better accuracy then needed to hear a pitch change.

Therefore, those that try to sell you an atomic clock (such as rubidium or cesium clocks), are going to charge you a lot of money, and not do a thing to improve your audio!

What one may need is for various gears to clock together. Foe example, if you have 2 chassis of AD's and you want to use them simultaneously, you may need to lock them together. If you do not, and they 2 chassis are just slightly off, after say an hour, the difference in times can accumulate to be way too big. So we may need to lock gear together, but that does not make the absolute time is so important.

You would not care if an hour performance was slowed down by say 10msec, but you sure do not want to mix 2 tracks that are time a aligned at the start, but are off by 10msec an hour later...

What IS IMPORTANT is jitter, and that is a "sample by sample" problem. Say one sample is early (relative to where it should be) by 1 nsec. By the time it happened, it is too late to do anything about it. In fact, at about 8 inches of wire (or trace a way) you are already 1nsec "behind the time". And if you try to control jitter for too far - send a clean clock a long distance, you will pick up a lot more jitter in that transmission path, then you started with...

Regards
Dan Lavry"

Citation : "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".



Hello Panos,

You are considering an atomic clock for audio. That is analogous to a “gold comb for a bold head”.

Atomic clocks have their place in the world. Their strong attribute is the absolute time accuracy. Say you want to have a clock that is very accurate, to say sub nano seconds over a day or a few nano seconds over a year, then that atomic clock is the way to go. The basic principle of operation of atomic clock makes it independent to temperature, and that is great. It is easy to figure out many cases where you need one or more clocks to be very time accurate.

But you do not need such time accuracy for audio. What you need most is low jitter, and for that, use internal clock for your AD whenever possible, and use external clock when you need to synchronize multiple chassis together. The absolute accuracy buys you nothing and costs you a lot of money.

Do you really care if an hour of recording is 1 second too long or too short? If you speed or slow down the recording, the pitch will change, but at what point is it even detectible by the ear? A great piano tuner can not tell pitch to better then 1 cent, not even with “a switch”, going between 2 notes 1 cent apart. That 1 cent is around 600ppm (parts per million), and that is not “high precision” when we are talking about clocks. Most crystals are way more accurate then that. When is the last time anyone complained about the pitch of the music played via crystal oscillator clock device?

Say your clock is 100ppm slow, and then an hour of music will be “stretched” in time by 0.36 second. A three minute song will be 18 msec (milliseconds) too long. From pitch standpoint, the error is less then 20% of what anyone can perceive.

So since timing accuracy is not about pitch accuracy. When you use internal fixed crystal, you get good enough accuracy. But what about the cases that call for external clock? When you use external clocks, you need to MATCH the timing of the various clocked devices. You need the “relative timing” to be precise.

If all the chassis are slow (or fast) by say 100ppm you are fine. The problems will show up when multiple chassis are not frequency locked together. Say you have 2 tracks, and one is slower then the other by 100ppm, which is .36 seconds delay after an hour, or .018sec in 3 minutes. That relative delay would cause problems, and that is why we need to lock multiple chassis to the same clock.

The key to clocking quality is low jitter, and the technology suitable for absolute accuracy does not necessarily yield the lowest jitter. In fact, it can not compete with a good internal clock. If you need external clock, use it, but the reason is to have all the channels "track".

The often stated “better sound” is just a hype. Digital audio is based on taking accurate samples at “equal time intervals” (low jitter). If each sample value is very accurate, and the time between samples is always the same, we can reconstruct the audio signal. If the sample values are less accurate, and the jitter is higher, we end up with errors, and any claims that those errors are better sounding is bogus. The errors are unpredictable, they are a result of very complex interaction between the jitter (there are many jitter types) and the music itself.

You do not need an atomic clock for audio. The jitter level is mostly determined by the internal circuits inside your AD, both for internal and external clocking.

Regards
Dan Lavry