Can you post a few seconds somewhere? The frequency and nature of the whine can often provide clues as to the cause.

Might help if, as well as a few seconds of the resulting audio showing the whine, you could post a few seconds of each of the original audio files too.

I'm guessing that you may somehow have an inaudibly high-frequency pilot tone one one or more of the input files that is getting aliased down into the audible frequency range for some reason in the conversion to MP2. But that's just a guess at this point - a look at/listen to the files should help give us a clue, as johnmeyer says.

Mark

I'm not at the computer at the moment, but I will try to post a clip asap.

The whine is definitely not the HD. The PC is purpose built for on-location audio and is dead silent. Additionally, the whine is there on other systems.

Ironically, I may have fixed it but have been unable to determine the source of the problem in the first place. I had left myself a considerable amount of headroom (-20dBFS). It wasn't really intentional...the audio feed that I would normally have captured myself was provided by a different audio guy this time and he kept the levels very safe.

In any case, apparently Vegas was auto adjusting this to peak at 0dBFS. Somehow, or for some reason, this noise was a by-product of this process. (Auto gain control??) When I brought the audio track up to peak at -1dBFS, the whinning went away (for the most part).

The original audio is dead quiet. It's a professional concert hall with Schoeps and Sennheiser mics into True Systems preamps - dead quiet. The noise floor (after being brought up to -1dBFS) is at -70dBFS. So, I know it's definitely not some inherent inaudible content being made audible by the gain adjustment.

Strangely, during very quiet passages, the whine comes back in again. The level of the whine seems dependent upon the level of the audio signal. If the audio signal is weak, the whine becomes loud and vice versa. During quiet passages, you can literally hear the whine changing amplitude by a range of 6dB or more with slight variances in dynamics of the program material.

I'm afraid that there is some processing going on during the compression that I'm not terribly happy with. Usually this won't be a problem. The typical output for our studio is straight to DVD where I'd use Sequoia to edit the audio for the DVD. Otherwise, it's straight to YouTube and frankly, I'm not overly concerned with the audio quality since they compress the heck out of it anyway. (I still want it good, but a little bit of a whine doesn't translate over most peoples crappy computer speakers...)

Thanks again for all of the advice and help. If I can easily post a clip, I will definitely do so.

Cheers-
Jeremy

"In any case, apparently Vegas was auto adjusting this to peak at 0dBFS"

Vegas does not auto adjust anything!

Having said that I'm quite uncertain exactly what you've done. You had a pristine analog audio signal, excellent and I wish I can ever be so lucky. But what did you use to convert that into the digital domain?

If you recorded that with 20dB headroom you raised the noise floor of the converters 20dB. If they are the best money can buy then not much of an issue. If you fed that pristine analog into a camera to record it, what a waste.

Other thing I've noticed is high compression audio codecs do evil things to low level audio. That might be where your problem is coming from.

Bob.

I may have missed something - was the audio mix recorded to the camera or something else?

If recorded on the camera tracks, are you using a dual DI box with the grounds lifted between the mixer balanced out and camera balanced inputs? This is essential, because you may have an unintended ground loop that is picking up induced whine from the camera motors.

As mentioned, when you raise the audio level in post, you are raising the noise floor by an equal amount. Starting with -20dB headroom and then boosting it later also has the unwanted effect of reducing your effective bit depth significantly. I shoot for -6dB headroom, knowing that there may be an rare clipped peak. It of course depends on your program.

> apparently Vegas was auto adjusting this to peak at 0dBFS.

Vegas would not usually do this. Have you perhaps set all your audio events to be Normalised for some reason?

> When I brought the audio track up to peak at -1dBFS, the whinning went away (for the most part).

Sounds like the whine may some kind of distortion by-product. If you are mixing more than one signal which are already at 0dBFS the resultant will be greater than 0dBFS and you don't want that happening - clipping will result. Reducing them to -1dBFS has given you enough headroom to (mostly) avoid this, though you will still probably be clipping at the places where peaks in the two (or more) input signals coincide. Normally (pun not intended) you want to be mixing with considerably more headroom than that.

> The level of the whine seems dependent upon the level of the audio signal.

Somewhere along the line you have got something acting as an AGC, either in the analog or digital domains (possibly in the camera). If it is obvious enough for you to hear its effect then you certainly don't want it, but until you discover where in the chain it is being applied we can't really give you any specific hints about how to turn it off.

Mark

Sorry - let me clarify -

The recording went into Lavry Converters into an RME Fireface via digital into Sequoia DAW. From there, I sync the digital audio with the video in Vegas once I have a 2 track bounce.

In this case, the bounce was -20dBFS peaks (normally I don't mix this low, but I had completely missed this since I'm working in a new product (Vegas) with new monitors.)

There is no chance that I went OVER -0dBFS. Again, my peaks were at -20dBFS.

With the noise spec on the Lavrys being nearly -120dBFS at 24 Bit, raising the whole signal 20dB would not be a problem. Room noise would be more of an issue. However, the room was dead quiet as already mentioned.

However, the whine was much "hotter" than the peak of -20dBFS. Unfortunately, I can only partially recreate this now. The whine is there and quite audible. However, in the original problem (which I apparently deleted), the whine would peak at or darn near -0dBFS.

This led me to believe that there was some auto-gain or auto normalization going on.

I've got two samples.
The first is an MPEG (MPEG2) and it has the issue that I mentioned but not as horrible as it was. It is after the applause.
http://www.sublymerecords.com/Sony/Mozart.mpg

The second of these is using .MOV format with uncompressed PCM audio. I hear no traces of the issue which leads me to believe it may be a compression artifact. In my many years of audio, I've heard all sorts of nasty artifacts from compression, but never one like this. That certainly would be disheartening. Perhaps it was a horrible combination of poor level adjustment (my fault) and compression artifact (digital's fault).

Second clip:
http://www.sublymerecords.com/Sony/2.mov

Cheers and thanks again for the replies!

J.

"Somewhere along the line you have got something acting as an AGC, either in the analog or digital domains (possibly in the camera). If it is obvious enough for you to hear its effect then you certainly don't want it, but until you discover where in the chain it is being applied we can't really give you any specific hints about how to turn it off."

Absolutely agreed. This is my thought to a t.

The only problem - I didn't record to the camera. I recorded directly into a DAW. If I were a noob at the audio portion, I'd say "gee, it's got to be how I recorded it" but the audio is something I'm definitely not new at. The only "unknown" in the equation is the Vegas Pro 9. I can't find anything about AGC or auto leveling/auto-normalization anywhere.

I'll keep plugging away and see what I find. In the mean time - thanks again!
J.

"Sounds like the whine may some kind of distortion by-product. If you are mixing more than one signal which are already at 0dBFS the resultant will be greater than 0dBFS and you don't want that happening - clipping will result. Reducing them to -1dBFS has given you enough headroom to (mostly) avoid this, though you will still probably be clipping at the places where peaks in the two (or more) input signals coincide. Normally (pun not intended) you want to be mixing with considerably more headroom than that."

That's definitely not the case - I didn't clip. Besides, digital clipping sounds way different - not a whine at all. I had to raise the level significantly to get it to peak at -1dB, not bring it down to -1dBFS.

Thanks!
J.

I don't hear a LOUD wine in the MPEG sample you posted. During the quiet period after the applause, I do hear some sound which sounds like the motor in your camcorder. I put the audio in Sound Forge, and there is a very distinct single frequency noise at exactly 600 Hz in both channels.

This is way down at -80dB and while I could hear it, I really wouldn't call it loud, as I said above.

If you are hearing something really loud in the MPEG sample, then the problem is in your sound card or computer and NOT in the file itself. That's actually good news because once you figure out what is causing the problem, you should be able to proceed and get a really good result.