Can I normalize the sound of the whole clip?

[jestem]

Can I normalize the sound of the whole clip?
The avs file is made of different avis and avs'. Some of them are louder. Can I normalize the sound of the whole clip?

[Mr_Odwin]

Have you tried it? Just add Normalize() to the end of your script.

[jestem]

When I add "Normalize()" at the end of the script and load the file into Media Player Classic- it doesn't play anything. Perhaps it is not AviSynth problem, but my computer which is very lagy for a month. Dunno why- no viruses found, 1,5 GB free for swap ;/

[Guest]

Post your script.

[Mr_Odwin]

When I add "Normalize()" at the end of the script and load the file into Media Player Classic- it doesn't play anything. Perhaps it is not AviSynth problem, but my computer which is very lagy for a month. Dunno why- no viruses found, 1,5 GB free for swap ;/

I've found it can take some time for normalize() to do its stuff. Try leaving MPC for a while and see if it plays in the end. Will the script open in vdub?

[IanB]

Yes, Normalize() can take a while, it parses the entire audio track before it returns the first audio sample. It is not unusual for 60 minutes of audio to take 5 minutes to parse (12 times realtime).

For your problem you probably want to normalize each individual clip to get them all to the same level.

It may be beneficial to first use the Normalize(Show=True) parameter and then replace the Normalize's with appropriate Amplify's as you edit up your script. This will also allow some tweaking of the levels to suit your personel preferences.

[Mug Funky]

would this be a bad time to suggest a new "replaygain()" function be added? :)

so one could guarantee an 83dB mix level and a -27dB dialogue normalisation to feed an ac3 encoder with (perhaps it could take mix level as an argument with 83 as default?).

considering some of the cool stuff being done with BeHappy and other avs based audio tools, it could help to incorporate replaygain into avisynth (the source is available on www.replaygain.org IIRC).

just a suggestion, of course. could be pretty good for the DVD backerupperers, or people making their own discs and wanting to ensure a uniform level throughout.

[edit]

a peak limiter would be needed too, but soxfilter could probably handle that, though i've not successfully got it not to clip yet. probably my fault.

[laserfan]

When I saw this thread I thought, "this should be easy, just de-mux the audio, run it thru BeSweet, and re-mux again" but no one has suggested that.

Isn't it that simple--what am I not understanding? :o

[Mug Funky]

@ laserfan: yes, you can do that, but if it were an internal filter in avisynth it'd save the extra steps :)

though, to me "normalise" does the opposite to what one would want. peak volume very seldom says anything at all about the actual loudness of a source. a loudness scan followed by increasing or reducing the volume to a reference level (-20dBFS is a good average) is a better idea. you don't want fullscale audio coming through the same system you use for DVDs... it'll blow the windows out.