Problem with varying sound levels
 

I'm working on a large project with 13 hours of voiceover material recorded on different dates, sometimes years apart. Some sections peak at -3dB, others at -9dB and there's a fair bit of variation.

I've got Audacity 2.0 and Sound Forge 9.0 and despite a fair bit of trial and error I haven't been able to find a way to produce a more consistent sound level. I thought using a Compressor or similar tool I would be able to remove a lot of the variation and come up with a consistent sound level. So far I haven't been able to do it and any help/suggestions would be appreciated.

Re: Problem with varying sound levels
 

Level is only one aspect and the tonality is probably what is not matching and you need EQ and a good set of monitors and ears to get that right, compression will help but it is more about finding an average that suits all of the elements.

However it may be that the use of different mics and environments may mean that you may never get a total match!

Re: Problem with varying sound levels
 

Ditto to Gary's comments but something caught my eye in your post...you mention how the peak levels are all over the place. Watching peak levels is certainly important to avoid clipping, etc, but perceived loudness is more a factor of the average level than it is of the peak. The old-fashioned averaging VU meter from the analog world is a better indicator of loudness than are peak-reading meters such as found is most NLEs. You said you have Sound Forge and it does give you the option of using VU/PPM meters as well peak reading meters.

Re: Problem with varying sound levels
 

Ditto to the comments above, and,

Most often I'd recommend against using Normalization until preparing for distribution. However, I believe you have an RMS method for your Normalize process in Soundforge. This can be useful for this kind of level matching. You might choose a fairly conservative level, say, -18 db, RMS not Peak, and run a few loud and soft clips through it. Maybe -18db is too conservative, you'll find out.

Re: Problem with varying sound levels
 

Just to add that if you are using clips from interviews as voice over it is quite natural for them to be different in tonal quality from the main voice over or narration on a doco which should always be recorded last once the final story has been established.

Most of this should be sorted in the sound dub rather than at the edit stage and I would always do a dialogue pre mix so I could sort our such level and tonality problems before adding any effects or music to the mix.

Re: Problem with varying sound levels
 

Many thanks for the comments which have been very helpful. Gary's comment about it being as much to do with tonal quality as peak levels is something I hadn't properly appreciated. I've had a look at the opening sections again and think I can see more of a way forward with this.

Thanks again

Re: Problem with varying sound levels
 

Hi Paul you are about 100 miles from me so if you want me to have a listen to some examples and give further advice just give me a shout.

I have a full pro tools audio dubbing set-up here with some decent reference monitors and about three decades of experience in pro TV and film dubbing! ;0)

Re: Problem with varying sound levels
 

Many thanks for the offer Gary that is very kind of you. I've had a reasonably productive morning with various alterations made and things are looking a bit brighter just now. I've taken on board your advice that items recorded years apart on different microphones aren't going to be a perfect match and I'm getting towards an acceptable end point.

Once again many thanks for your help which is much appreciated.