What level do you normalize to?  Do not normaize to 0 or anywhere close. There is a good dynamic range even with 16bit audio so give yourself plenty of headroom and avoid any risk of clipping. Then bring the level up in the final mix if you need to. Also look at the audio Track Fx where there is a Compressor in circuit by default, albeit with no effect settings.

This can happen if you're using 44k audio in a 48k project, the resampling algorithm can mess with the levels slightly. Normalize to -.5 and that should take care of it.

There are basically two types of normalization: peak and average (loudness). Normaizing to 'peak' wiil bring the highest peak to the user-set ceiling (-0.0dBFS being the maximum). Average (RMS) can induce compression and expansion and clip if improperly set. Inexperiened users should stay away from average. As John stated, peak normaizing to -0.5dBFS should be ok, however lossy encoding could still clip the waveform if headroom is not taken into consideration. In VP, event normalization does not have the average option.

I will run a test with a cample clip to try to reproduce the clipping after normalizing. This I found especially evident when editing in Sound Forge - applying a couple of effects including Wave Hammer- and normalizing whithin. The saved clip in vegas would clip significantly, though I suspect part of the reason is that the default type of normalizing is RMS,  as well as  the fact that the clips were 48k in a 44k project probably.

I did also see the peaks when normalizing in Vegas, only unique moments (ie not frequently throughout the audio as it would when importing from Sound Forge.) I did check that all effects were in fact disabled at track level when this was happening. I asume this was due to the difference in the audio properties vs the project properties as John_Cline has suggested.

As it is I'm mixing VO with a music track, so have had to tweak volume envelopes for ducking.

So thank you all for your help with this audio thing...! Learning loads of new stuff and you guys really help me along..

Best regards,

Paul.

Just a quick update on how the project ended....

having sorted out the different audio mixes, ensuring that none peaked above 0db, I finally started a new project to render out the final video, including two preciously rendered avi and three more nested vegs

Thought all was hunky dory and rendered away: 4 hours for a 15 minute video...

anyway, played the video on a TV as a final check before delivery and... the audio was clipping...this was with a spanking new default veg with no new effects or volume meddling...

I confess I was seriously confused. 

I guess I need to normalise AGAIN within this final timeline before the final render....

So more patience was called upon and another 4 hours of rendering...

If I'm doing anything wrong I'd really appreciate advice. I also wonder if normalizing is always needed, why can't Vegas issue a warning when you are going to render saying something like..."hey! Before rendering, have you made sure you have normalized all audio??"

thanks again for all your help!

best regards,

Paul

I also wonder if normalizing is always needed...

Not at all. You certainly don't need to normalize in Sound Forge before bringing into Vegas, as I suspect that does entail more risk of clipping than normalizing in Vegas.

In many cases you would be best to adjust your final mix using the loudness meters in VP13 and 14 to achieve the right level. Sometimes it might be appropriate to normalize events prior to that then reduce the volume/gain of the event or track or whole mix, and in other cases you might simply want to increase the audio and not do any normalizing at all. It all depends what audio you are working with.

It's been ages since I did any loudness work, so this might be wrong or out of date, but from what I remember you would want to aim for -23 db LUFS Integrated loudness for most purposes including broadcast. For YouTube, where I'm competing with a lot of videos that have just been maxed out, I have often gone for -20.

I mix EVERYTHING inside vegas, limit tracks (and sometimes individual clips) with LOUDMAX, normalize per clip if needed, eliminate all peaks and never have a problem.  I pull my final output to -1, someitmes -2 or -3 depending on RMS, test play on everything from crapphone to home theater, all is good.  Last production was 37 audio tracks, all inside Vegas and worked great.  Vegas rocks.

-23 LUFS / -24 LKFS  for broadcast, 6 dB louder for internet, but limit the peaks!

The problem with -18 for internet is that if people are trying to listen on their laptop built-in speakers they may not hear anything at all. Often some external amplification will be required, which is something many viewers won't have. There is no reason to mix to anything below -0.001 for internet listening.

-18 LKFS equivalent is plenty loud for internet/PC listening, and can be considered a defacto standard, following iTunes' lead for the past decade

That's unrelated to either dBFS or RMS norming, so need to keep those units of measure clear; norming to -18 on either of those scales would probably be unacceptably quiet.

Worst case, peaks over -1dBFS need to be avoided at all cost because of intersample clipping, which most hobbyists are neither equipped to detect or understand.

FWIW, -18 LKFS (integrated) works out (for me) to be around -4.0 dBFS, however this depends on the amount of comp/limiting, EQ and other variables.

As MV stated, the LU/LK loudness scale is not to be confused with the dBFS (peak) or RMS (average) .... among others. Aside from LKFS integrated, there's also Maximum True Peak, Maximum ShortTerm and Maximum Momentary. Sorry if this is complicated for non-audio folks. The ATSC paper 'A/85, Techniques for Establishing and Maintaining Audio Loudness' may be of interest. Most of which applies to EBU-128 loudness (Europe), which for b'cast is -23LKFS. So if a project is within ATSC A/85 specs, it most likely would not be rejected for EBU broadcasts.

So based on the suggestions in this thread, as I understood them, I adjusted my latest video to -17 LUFS (actually it's more like -17.4 LUFS) integrated loudness. Here are the settings and readouts during playback.

Here's the published video on YouTube. Apologies if you have to skip an ad:

I think it's a bit quiet in relation to "the general recent stuff" on YouTube, and will probably go louder next time.

Anyone agree or disagree?

And have I got my settings "correct" and is my understanding of the numbers correct? All these 4-letter abbreviations get confusing.