Kriz, this time there was not a .5dB difference,. they were exact as far as level. but they did not sound the same. .... did you download and listen to them?

but a BB making a difference on BTD...hmm...

Apparently, you're wrong.

Look. Dave from Apogee made you an offer, didn't he? Why don't you run your own tests and get off my back.

Jasper

It's very simple. A clock is always clocking the whole system.
The BTD is a realtime process. As the word time indicates a clock is needed for this process.
Think of incorporating ext. devices such as synths during BTD. How should they be aligned correctly to the audio tracks if not by a clock?

Sorry Mike, I don't mean to be on your case. We're just having a discussion here. You put some data up and it's being analyzed, no big. I mean, that's the point of these forums.

What I don't understand is how different clocks could make a difference in the final product when you're doing a BTD. So, if you're hearing a difference, I think it must be from something else.

-R

If you are entirely in the digital domain, that is, everything has been turned into numbers, then those number will stay in the same order and spit back out in the same order regardless of the clock's jitter. When you are doing the BTD you are hearing the jitter of the D/A, but that's not affecting the bounce (unless you are going analog out to a summing box then back in). When you listen back to the bounce you hear the jitter of whatever playback system you are using. Again, it's not part of the file itself. However, when you are actually recording something to disc from the analog domain, then any jitter of the clock will be "imprinted" on the digital file, and you'd be stuck with it.

The way to test clock differences is to use them during a D/A or A/D conversion and then compare the results. In a BTD, that line is not crossed. That's why I asked Jasper for specifics. If he's using an analog outboard reverb, for instance, or running outboard synths live to the mix, then yes, jitter will be a factor.

-R

The ProTools Bounce to Disk is a realtime process. Someone has to say when it's time for the next sample. Again: How should e.g. my M3000 AES Signal (which is also just numbers) get onto my TDM-Bus correctly if not a clock tells the system where we are?

If an algorithm would calculate the mixdown (faster or even slower than realtime) - in that case you would be right. Just number - dig. Words with a timestamp, so the algorithm knows where to put them in.

while not an expert, i am pretty sure that RKrizman is right. i have seen other discussions where the people who know better than i do (nika, dan lavry) that the only affected by the clock, like a BB is the A/D and the D/A. and internal digital processing and D/D is "off the clock" -- which affects the chips ability to perform the conversion. in D/D and internal routing etc, there is no jitter, just numbers being pushed around. (unless i misunderstood those discussions)

but again, i hear a difference in the files. so, i am puzzled as to what is going on here.

As long as the numbers stay in the same order it doesn't matter if the space between them varies.

I wouldn't dig yourself in too deep on this one. Anybody who really knows anything about this will verify that I am right about this.

-R

I didn't mind your comments after the first test, because the first test was pretty sloppy. Basically, I wanted to know if I would hear a difference in a BTD mixdown. And I remember thinking, like you, that I would probably have to go through the analog outs of the PSX100SE to the MasterLink to capture the sound I'm hearing. Makes perfect sense, because then the D/A is in play.

But I thought what the hell, try BounceToDisk first. I was surprised by the results, reported that, and someone asked if they could hear the files, so I uploaded what I had at the time. You and Seymour questioned the methodology based on the .5db smoking gun, so I ran the test again.

Here's how I ran the second test:

I pulled up the Pro Tools session with the guitar and vocal I used in the first test, and highlighted the part I wanted to bounce. Then I set everything to the Big Ben clock. For this test, I kept the Kurzweil KSP8 out of it -- no hardware or software reverb of any kind was used, so it's dry, dry, dry. Also, no external digital components were used except for the Big Ben and PSX100SE.

1) Pro Tools LE "Clock Source" was set to S/PDIF. It was kept there for both tests.
2) The Apogee PSX100SE, connected to the Digi 001 via S/PDIF, was set to Word Clock Lock.
3) The PSX100SE was connected via BNC connection to the Word Clock output of the Big Ben.
4) The Big Ben was turned on and set to 44.1.
5) I did bounce to disk, set to 16-bit wav, 44.1. On the Master Volume of Pro Tools I used the PowR Dither plugin.

You can find the result at http://www.deceptivesound.com/BigBen.mp3

After that bounce, I left the Pro Tools software exactly as it was for the first test. Let me say this one more time: There was no Pro Tools LE software change from the Big Ben bounce to the PSX100SE bounce -- that means clock source was again left at S/PDIF.

There was a hardware change. This is what I did.

1) I turned off my monitors (the change generates a lot of noise, man. People should be aware of this.)
2) I turned off the Big Ben at the power button.
3) I set the PSX100SE to Crystal, 44.1. Now when Pro Tools looked to the S/PDIF line for the clock, it would use the PSX100SE.
4) I turned on the speaker montiors.
5) I bounced to disk exactly as above.

The result was http://www.deceptivesound.com/PSX100.mp3

RKrizman wrote, "So, if you're hearing a difference, I think it must be from something else."

If that's true, then what do you suppose the something else might be? It's not external reverb, right?

Don't get me wrong, I can see your point. I'm not sure why there's a difference either. More specifically, I'm surprised that I'm hearing in the bounces something pretty close to what I heard through the speakers during the bounce (although what I heard through the speakers was from a 24-bit source).

I can't explain the result either, but I can tell you it's not because I did something stupid or dishonest during this last test.

Jasper
PS -- I guess in the spirit of full disclosure, I should point out that the S/PDIF connection between the Digi 001 and the PSX100SE is separated by a Midiman DigiPatch digital patchbay. Of course, the patchbay was kept at the same settings for both bounces, so essentially that's a straight wire. Also, the analog outs of the PSX100SE are connected to a Coleman M3, and then to Hafler monitors. That's pretty much every little detail I can think of.

Nope. You're wrong. I can hear that for myself.

So are we listening to a 16/44 wave file or an mp3?

Is there a way I can download these files and drag them into PT for comparison? (I can listen but can't figure out how to save them as files) You might try that yourself. Flip the phase to see if they null out. Then record the null signal onto its own track and normalize it so you can hear into it and try to figure out what's accounting for the difference, if in fact there is any. Of course, they won't null to complete silence because the dither is different on each pass.

Maybe my ears are a little stuffy today, but I'm not so sure I could distinguish these two files in a blind test.

-R

Perhaps you can hear a difference, I can't speak to that. I'm only questioning the reason for it. You've thought of everything, until you think of the one thing you didn't think of. Nobody's accusing you of being stupid or dishonest. But ask around about the ITB jitter issue--I think my understanding of that is pretty solid.

-R

I just talked to David Seymour on the phone. He said that I was 100 percent correct, and RKrizman was all wrong.

Just messing with you.

I don't know, man. I hear something with the Big Ben on that I don't hear with it off, and I hear it in the mix, and I like what I hear. That's all I really know.

Will Big Ben improve mid-end converters? I have to say yes. Would I be better off with a Rosetta 200 or a Mytek? I aims to find out.

Jasper

Good point.

I took the 16-bit, 44.1 files and converted them both to mp3 using iTunes.

Jasper

So try my suggestion with the orignal files. See how much they null.

-R

Doesn't matter if it's a realtime process or not. If there's no conversion (either A/D-D/A or sample-rate conversion) taking place then the clock won't affect the sound (unless something's not working right). All that's happening is that we're starting off with some data, processing it, and outputting a file. As long as that all happens there's nothing for the clock to affect whether you're talking about an internal bounce, connecting two digital devices, etc. The only time it would affect things would be if the jitter was high enough that it would cause errors. Either it would work or it wouldn't. Everything has to be locked to the same clock, but in your example (the TDM system and the M3000) as long as the clocks are locked together it doesn't matter how jittery the signal is. Every second 44,1000 samples go in, are processed, and go out. We can hear differences in jitter but as far as the file that results, it will be exactly the same.

This is the same thing. Like people walking through a revolving door...if the door is revolving at a certain speed, it doesn't matter if one person squeezes out as soon as they can and the next person just walks through and the next person waits until the last second...as long as nobody actually slows the door itself down or squeezes in at the same time as another person (which would be analagous to errors) the number of people that go through in a given period of time won't change.

I'd be willing to bet that that's where the difference is happening. If I recall correctly that box does do some sort of buffering/sample rate conversion...it probably messes with the signal enough that clock differences could make a difference. I remember that being a somewhat finicky box. I'd imagine that if you went directly from the PSX to the 001 you would not hear a difference between the files. You shouldn't.

In any case, that does show that the clock can make an audible difference. A better test would be a mix run through a mixer or summing box of some sort, where an A/D-D/A conversion is involved.

-Duardo

kriz, listen to the files.

Never mind my last thought on the Digipatch...I was thinking that audio was passing through it, which it wasn't. Just the clock, right?

Not sure why you'd hear such a difference...but something's odd, as a digital bounce should be unaffected by the clock.

-Duardo

Doesn't PT use the converters when it you do a BTD? If I remember right, PT asks if you want to use "Interface 1&2" or something of the sort...

If so, wouldn't the clocking be used, and possibly be affected by the BB?

Just curious...

I did, and have commented. See above.

-R

PT asks you what bus you want to bouce to disc from. If you choose converter output 1 and 2 for your bounce that just means that all the information being bussed to that converter output will be the information that the BTD process will use for the BTD. No conversion takes place. That's always confused me too.

That analogy about the revolving door explained it best. As long as the bits are in the same order, the files are identical.

-R

Another test? And what about? Null? Naaaw. I'm getting a headache. To make matters even more complicated, I discovered the Prism converters last night. That's got to be overkill, right?

"No, not all, they're essential!" said the Prism salesman. If he ever sees this thread. Probably.

Tell you what. I'll upload the two 16-bit wavs I made for the second test. Then maybe someone else can null them. They're 17mb each, which means expensive bandwidth, so I'll leave them up until Sunday night some time - Oct.30. After that, they'll be deleted.

Now. Somebody else do a damn test.

http://www.deceptivesound.com/PSX100.wav
http://www.deceptivesound.com/BigBen.wav

Jasper
PS -- I've got to admit, as I'm uploading these two .wavs I'm thinking, man, things sure have improved since 1996, the first year I climbed onto the Internet. Here I am, uploading a couple of sound files I made in my computer-based audio studio so people all over the world can grab them and listen. A lot of **** has gone down in the last ten years, but definitely some good stuff too. Definitely some improvements, if you stop and think about them.

I'm getting all misty eyed here.

Somebody from Apogee mentioned that Rosetta 200 alone will match any converter in the market. I would add: IN SIMILAR PRICE SEGMENT !

If you have plenty of cash try DCS or Weiss, they actually sounded better on most of materials and it could be easily percepted at first audition.

Widely used Apogee, Mytek, Prism, Lynx, Benchmark, Lavry and other higher end converters might differ among themselves in how they translate from domain to domain, but not to such extent to call any of them definitely better for all aplications.

I use Apogee all the way incl. BB (for other digital devices) simply due to fact that for number of channels I need for typical project the best available conversion would cost me a fortune and final result does not justify such difference.
I think that Apogee succeded to reach unrivaled balance of high end level quality at comparatively affordable price.

Seems that one more long and boring thread again didn't clarify anything in relation to the subject.

GYang

oh, hey, SORRY! missed that post.

i don't have a Big Ben. i have a different clock. so my test wouldn't apply to this thread.

Oh, I don't know about that. I was convinced that I no longer need my digital patchbay.

Hey... I got something out of it.

Jasper

haha you got that from the back of the big ben manual. . . yeah, I read it

but that is a GREAT example.

There was a point I was completely against the Big Ben, but it was the fault of crappy BNC cabling and improper termination. Now that I have my Lynx Aurora properly clocked to the big ben, I think it is a wonderful device.

I don't doubt it. I think quality converters are just as important as your mic and pre, it's only as good as the weakest link in your chain. I have a Trak 2 apgoee and it KILLS the old 888s, even kills the 192s in my opinion. Even the Mini Me by apogee sounds really great, hell I would say sell one of your mediocre pieces and get a mini me if on a budget. And a Big Ben will definitely help your converters, but it is even better to start with something decent. Rosetta 200 could even clock your other gear and make them sound better than they do. That's my .02

Actually the Apogee AD16x and DA16X do have the Big Ben clock built in. Only the Rosettas do not, though their clocks are not bad, but the BB is better.

Raul,

I bet you will hear the BB make a noticable difference in your other converters for the better. That said, the point about buying better converters in the first place is extremely valid. Try this: the big ben is only about $1200 versus alot more $ for better converters. Buy a BB and see if it improves your sound enough to justify it, If it does, keep it, if it doesn't, they hold their value really well and you can resell it for probably close to what you pay for it if you shop around and get a good deal. If you don't keep it, then buy some better converters altogether. And of course you have to consider how many channels you really need, how are you interfacing with whatever DAW you use, all that you have to figure into how much you will have to spend on new converters, I think the cost of the BB is small for what it can do and when you upgrade overall converters, you can sell it and get $95% of your money back make sense?