Yes, you're right. The snare is what's causing the major +dB, because it's mixed in on the 2 overheads also. The rest of the tracks ARE only a few + dB when the snare is muted. The type of music is just general rock n roll, so it has "driving" drums.
I'll check out the Ren Compressor. Thanks for the tip.

It has also confused me, but what I guess is happening, is that the 'overs are accomodated by the overhead available with the high bitdepth in the 'processing'.

This is also evident when you go into the red in plugin processing.

It seems when rendering, the output file is scaled so that the loudest peak is not clipped.

I would appreciate confirmation of my theory too !

I'm sure someone else could explain this better.....but I believe that the processing Vegas uses to mix the tracks knows the combination is going to be over zero by "x" amount and indicates this on the meters. When you render to a certain bit-depth, a zero dB level is essentially all bits *ON* (I think 2's complement is used however, but for the sake of argument). Now think about it, if all the bits are set when you are at zero dB - whether it be 16, 24 or whatever depth render, there is no way to encode anything over zero dB. Thus your rendering will not show the overs as they were mixed. I assume overs shown after recording/mixing are detected by consecutive samples at 0 dB (??).

Bottom line: Stay out of the red! If you need more loudness etc. - compression and peak limiting are required.

I'd love some clarification on this also. If Vegas uses its internal 32-bit resolution to prevent distortion on the output, how does it scale the 32-bit data back to 16 or 24 bits without distortion, and what's the meaning of the clip indicator? Does Vegas have something like an internal limiter on its master channel?

I think you missed my point on my last post - probably because I did a poor job explaining it. Zero dB is all bits "on", "1s" whatever. If you go over zero dB you have all the bits turned on and no room left to encode any more information - this is the same for 16 bit, 24 bit or more. More bits just gets you finer resolution\more dynamic range - not louder signals. If you take a signal that peaks at exactly zero dB in a 16 bit file and convert it to 24 bit it will still be zero dB. After the conversion you will have 24 1's instead of 16 1's. This is also true if you go from 24 bit to 16 bit (or any other bit-depth).

I am fairly certain Vegas does not have an internal limiter on the master bus. You are just seeing the effects I described above. After you reach zero dB and all bits are "on", there is no room to go. I think what confuses people is that the meters indicate you are over by "x" amount yet the rendered file shows only +.3dB. I have to guess that the Vegas DSP that is going on when mixing is aware that you are clipping and displays the amount. When you render the file it is clipped and the playback after that does not show exactly how much it is clipped as there are no bits left to encode higher levels.

It is not correct to make the assumption that you have no distortion when running in the red with Vegas as you are. You may not be "hearing" the distortion - but trust me it is there. Clipping very short, sharp transients may not be very audible. Using a peak limiter such as the Waves L1 actually intentionally clips the signal but does it more gracefully. The peaks are "shaped", you have control over the out ceiling, and dithering/noise shaping are used when re-quantizing. By strict definition the L1 is introducing distortion on the signal, albeit controlled and more "pleasant".

Also, processing at a higher bit-depth than your source files (as is done in Vegas) is done to reduce distortion - but not clipping. DSP processing is simply math done on the audio data whether it is mixing of tracks, EQ, compression etc. When the math is performed more resolution is needed so that for example, when you divide 22 by 7 you get 3.142857143 instead of just 3. The end result is re-quantized to bring it back to the original bit-depth though adding dither can reduce the distortion caused by this. The fact that the internal processing is at 32 bits in Vegas does not eliminate clipping that is shown on the meters - it just improves the accuracy of the processing.

I hope I did a little better job explaining this time. I am no expert by any means but it does help to understand some of the basics of digital audio when using these products. I suggest finding a decent book on the subject. "Principles of Digital Audio" by Ken Pohlman is a great book though a bit on the advanced side.

Two comments:
1. It seems that in Vegas, there's no audible distortion even if the signal on the master level meters goes significantly above 0dB for a significant amount of time (ie. more than 3 samples). Of course, it could be that my ears are shot, but others have noticed this as well.
2. It's perfectly possible to use 32 bit resolution to increase your dynamic range above the 0dB of a 24 bit encoded signal. It just depends on how you define the relationship between the number that encodes the signal and the dB scale.

I thnk what happens is that the 'over' relates to the file's original bit-depth (recorded files *will* clip on an over at the recording stage), but subsequent processing that pushes the signal 'over', is absorbed by the processing headroom (presumably a 32 bit float).

More significantly, it seems almost impossible to clip even a large mix. I'm sure what happens is that the rendered file is normalised to the highest 'mixed' peak, and teh resultant file is dithered down from there.

i would love to have this confimed, as red 'over' alarms in digital systems aren't conducive to low stress.

geoff

I don't think Vegas normalises to the highest peak, since it cannot know in advance what the highest peak will be (you may decide to change the volume of a track while Vegas is playing back, for instance). It's more likely that some of the 32 bit dynamic range are used to absorb overs. The 32 bit signal is then mapped to the 16 or 24 bits of the output device (hopefully without dithering). This would mean that if you render your project and play it back through something like WinAmp, it would be louder than playing directly from Vegas.