This is a continuation of my jitter test thread. I just wanted to pass on some things I learned from my testing.

All results are based on recording a wave from Sonar into DA7 digital in, and analog recording back to Sonar. This essentially is testing the DA7 DAC and Lynx One ADC combination.

A little background info: The DA7 DAC has a noise spec of –91dB and the Lynx One has a noise spec of –93dB, so technically, you could not cancel to less than –91dB. I tried recording dead silence, and it was exactly at –91dB, so that is the limit of testing.

The theory in this is, if you record an identical signal twice, you should get 100% cancellation if there is zero jitter in the clock. In this case, 100% cancellation would be –91dB.

I ran this test using DA7 internal clock, Lucid Genx6, and Ramsa AD96M clock, along with a high quality 50ohm wordclock cable, and then a good quality 75ohm cable.

The first thing I found to get good cancellation is, I needed to let the system warm up for at least 2 hours. This was one thing that fooled me in the beginning because as it slowly warmed up, you would hear differences. Once that was out of the way, I tried all the various combinations of cables and clocks. I found worst performance with GENX6 and 50ohm cable combination. With 75ohm cable, the differences between clocks were much smaller. As far as cancellation, the best cancellation overall was Ramsa AD96M clock, but even at this it is somewhat subjective. The difference between clocks was a high of –83dB and low of –85dB. This is so small and in the sub-hearing range, I would find it hard to justify a dedicated clock unless you have multiple digital systems to sync up. To hear the cancellation signals, I boosted it enough to hear the level, and what I found is in all clocks, the music signal tends to fade in and out at the –85dB level. With each clock it would fade in and out at different rates. If there were no jitter, you should hear only white noise at –91dB. So even the best clock showed some jitter.

Another thing I was wondering about was what the effect of power cables had on the sound. This has always been a point of debate. I have some DIY cables I made with high quality connectors and shielded belden cable. I recorded once with the DIY cables and again with the cheap cables that came with the equipment. I got cancellation down to –85dB. This tells me the power cables had NO effect. I tried this several times, with a similar result. The key to doing this test right is to run through the cycle of switching the cables at least a couple times without disturbing any settings on the affected equipment. The reason for this is, just turning on and off something could initially cause a difference.


Another test I ran was testing the ADAT interface between my EMU sampler and DA7. I played a sample from the EMU through the ADAT DA7 to sonar and did the same cancellation tests. This is an all digital path. I got cancellation down to –96dB on a single stereo track. Cranking up the cancellation signal, you could hear all kinds of weird clocking noise. With the Genx6 clock there was a noise burst on each hit of the high hat.
With the DA7 you could hear an overall faint level of music but less noise bursting. That tells me that the Genx6 does a better job of repeating the lower and midrange (more cancellation), but less on the high end, whereas the DA7 clock is overall more smooth across the spectrum , but overall less accurate(less cancellation). Still, at –96dB, this is still so below the threshold of hearing, it would be hard to hear the differences with the music playing at normal levels.

Finally, I ran a digital out from Sonar to DA7 and recorded digital in to Sonar. I achieved –132dB cancellation. The difference signal was dead silent except for some white noise sitting at the –132dB level.

Another interesting fact I found. If you run a digital signal from Sonar into channels 15/16 straight to L/R mains and record back to Sonar, and then run the same signal again into Bus 1 / 2 which then goes to L/R mains, there is a 10 sample difference. So beware when submixing, to compensate for the 10 sample difference. I didn't realize this existed until I ran these tests. I tried this a couple times, switching back and forth, and it was repeatable. Could this be a source of others not so 3D imaging problems?

The thing I like about these cancellation tests are, it takes the psychoacoustic brain games out of the equation. It's easy to fool your mind into hearing things that don't exist. I'm convinced after this, that at least in my system, I have probably hit the limits of improvement short of buying some different equipment.

If anyone is still reading this, one question to the audience. How can I get a dynamic range of 120dB on Ramsa when the noise level is at –90dB? It seems the best you could do is –90dB.

Cheers

cyberblue,

This is an interesting test with even more interesting results.

And even though I’m a staunch supporter of hi-def cables, I too have always questioned the esoteric A.C. power cord claims.

I’m going to run this thread by some of my audio guru friends and see what they have to say…

in any event…in a blind a-b test can your hear any differences between these converters and clocks?

Let us know.

Maric,

One of the things that prompted me to do these tests was, I sometimes thought I heard differences, but if I repeated the test multiple times, I would sometimes pick the opposite one I liked the most. You hear the least difference using headphones. When I listened through my monitors, I could sometimes feel there was a change, but if I turn my head a fraction of a degree, it will change somewhat. I think that's what makes these a/b tests so difficult to judge.

Just imagine you have several people in the same room, struggling to hear differences and someone moves or breathes, won't that affect the way the sounds are bouncing around in the room?

One thing that I am happy about, the 75 ohm cable, I used Canare RG-11, did show an improvement for both the DAC and ADAT tests, so in that respect, I was successful in correcting a problem (using wrong impedance cable).

One question remains. Would a Big Ben clock or Aardsync get me to the -91dB cancellation level? Anyone out there with one of those that could run a test similar to mine?

That would be interesting.

C-blue
Hey my name is Sam and I have a studio in Whittier, CA. I'm very interested to see and hear these things and how you have your stuff set up (being another DA7 user) I would be willing to trade since I am in S.D. alot, my fiance' parents live there and we are there alot planning the wedding. If you were to come to the L.A. area you would be very welcom to come and check out my set up as well. I always find I learn a lot from experiencing someone elses work space. Let me know.
sam
earwitnessrecords@yahoo.com

Cyber:

Interesting thread; I like your systematic approach. About 2 years ago I had a chance to work with a cool studio in Nashville that had a big sonar / protools system. They would always turn on their master clock on for at least 2 hours to "warm up". I had asked why, and what they told me was it gave the crystal that controls the clock a chance to warm up and stabilize. Reading your post now makes a little more sense. Another point one of the engineers made was NOT to daisy chain your clock from the master generator - i.e. run parallel lines to each piece of equipment to help reduce jitter.

I know from experience in my home studio that when switching to Tara labs optical cables (thanks Long!) made a big difference over the stock HOSA cables I had been using (I was using the DA7 and 3 ADATS at the time). I can't say the improvement was reducing "jitter", but it seemed like I was removing a veil from my speakers when I used the tara labs (maybe this would be the same if I had used anything other then the bottom of the line hosa optical cables).

Alan

Alan,

I did try this test with a cheap optic cable and some monster cable on the ADAT. They both canceled to -96dB, but the noise at that level sounded different. I didn't listen critically at normal levels, so I don't know if it really makes an audible difference. With my EMU unit, it has both analog outs and the ADAT, but I can say for sure, to me the ADAT sounds much cleaner than the analog path. The noise floor on the ADAT is silent. i wish it had an AES option, but I use it for mostly bass and midrange sounds, so I don't think it's too critical. The ADAT seems to have the most difficulty in cleanly reproducing very high frequencies, at least that's what I'm hearing from others.

Sam,

I'll contact you by email.

50-ohm cable is not the correct cable for word clock. You must use 75-ohm cable. It's no wonder you got poor results with it.

Rick,

I know that, but at the time I had access to free 50ohm cable with BNC cables, and it sounded okay. It actually is not all THAT bad, only slightly worse than the 75 ohm, but I see your point. This is only 3 to 6ft. The 75 ohm becomes way more important as the length increases.

But it seems pointless to test incorrect cables.

Would you bother testing a 3-prong power cord as a balanced audio connection? No, because it makes little sense. (Though it would work in some cases...)

The testing could cause someone to assume that 50-ohm cable can be used, especially if they don't read the whole message. (You didn't say explicitly that it was the wrong cable, only that results were mixed.)

If it came with XLR connectors and was free.

I was curious to see how it would work. Anything wrong with being curious? I'm an engineer. That's how I think. If I thought something was dangerous, I wouldn't try it. At the time, I didn't know I needed 75ohm, and I just had some 50ohm available, so when I decided to buy some 75 ohm cable I wanted to see if there was any improvement at all. I wanted a baseline to see if there was a difference. Based on what I found, the difference was pretty small, so it kind of tells me buying an esoteric upgrade from a good 75 ohm would also be small. Does that make more sense now? I could go on for hours on other experiments but you probably wouldn't be interested, so I won't list them.

I keep reading about how the image is wider with this cable or that wordclock, or how jitter is damaging someone's sound. I wanted to know what jitter sounds like so when I make a change, I know if it is really an improvement. I also tried a bare wire with no shielding on the wordclock, just to see how it sounded. Even that didn't sound MUCH different. It was all subtle at the audible level, and hard for me to tell. I was only suggesting a more objective method of testing a setup instead of using subjective listening tests.

Isn't anyone curious if these esoteric cables are really worth the money? I am trying to point out that at least for me, using good quality cables of the proper impedance is good enough.

I read that some esoteric audio cables, not wordclock, sound warmer mainly because they have high capacitance, and actually roll off the highs. I don't think anyone mixing music would want to roll off the highs through their cables.

I just wanted to pass on my experience so anyone else out there with the same questions could learn from my mistakes, and maybe try their own experiment if they wanted to. I sense from reading several audio BB boards that there are alot of folks questioning their gear, and looking for some magic cables or word clock or mic-pre or microphone or ... to solve their problems. Maybe they COULD use some improvements in gear, or maybe they just need to improve their own mixing skills. For me this exercise was just a away of giving me some confidence in my equipment.

YMMV

"Isn't anyone curious if these esoteric cables are really worth the money?"

Hmmmmm, that's a slippery slope dude...... I'd have to say that the folks who have already spent the $$ have already made up their minds

OK, sounds sensible.

I think the biggest thing is the blind listening test. If people can't tell the difference (beyond simple chance), then there's no need to make the change.

A lot of people's monitors aren't good enough to expose the differences either.

Which tweak cables? Digital, mic, line, power?

Long's mic cables vs. standard Canare (maybe it was Mogami, I don't remember) made as much difference with an SM57 as a lot of mic cables would make.