Smoothing Sound files

[Coriolanus]

If I create and audio book out of cassette tape I have a problem withe uneven sound levels on a Cc.

Although when I record from the cassette the sound will be at the same level for that tape, when I join the tape together to get the lengthe for a CD - I have portions of the CD with varying sound levels.

I have been "eyeballing the file" and where I see the levels are not the same I have been using AUTO GAIN on portions of the file so that they all have an equal level.

Is there any ONE step process I can use to do this? I tried GROUP NORMALIZE on Adobe that that didn't seem to work for me. I have tried MAXIMIZE etc. but they don't see to addjust or smooth out the audio levels. Is there anything that will?

[DougDbug]

Try The Levelator (FREE!!!). This is a non-linear effect similar to Auto Gain, but (I assume) a different algorithm.

For matching perceived volume, nothing is as good as the human ear... Two or more humans might not even agree which file is louder if the files are different, such as one music file and one voice file, different styles of music, or one music file with constant loudness and another music file that starts-out quiet and ends loud... Complicated stuff!!!

Normalizing (including GoldWave's Mazimize) works on the peaks, but loudness does not correlate well with the peak level. These might help, or they could make things worse... One extra-loud peak can throw-off the whole file.

I don't know what Group Normalize does, but it might use the loudest peak in the whole group as a reference, boosting/cutting all of the files by an equal amount. You'd want to do that on an album to keep quiet songs quiet and loud songs loud as originally intended.

Also, any linear effects are applied equally to the whole file, so they can be used to match/adjust different files with different volumes, but if one file has loud parts and quiet parts, these tools won't even it out.

Linear adjustments are generally used for music to retain the original dynamics (loud parts and quiet parts of a song). Non-linear techniques are more common with dialog, although (non-linear) dynamic compression is used (or overused) everyday in modern music production.

GoldWave has a Match Volume effect (also linear) that helps match the average volumes of different files. The difficulty here is that you can end-up boosting a file into clipping (distortion). So, the "right way" to is to Maximize all of the files first. Then scan all of the files with Match Volume, noting the average dB level and canceling the effect. Then choose the file with the lowest average level as your reference. Go back and use the Match Volume effect to bring-down the other files to match.

Other linear volume matching tools are MP3gain and WaveGain. These tools use a very sophisticated algorithm that not only takes into account the average volume, but the frequency content. Our ears have different sensitivity at different frequencies, and relative frequency sensitivity also varies "2-dimensionally" with loudness (Equal Loudness Curves).

Dynamic compression (non-linear) can also be used to make the quiet parts louder, or the loud parts quieter. So for example, it can bring-up the average level without boosting-clipping the peaks.

With GoldWave - Effect -> Compressor/Expander:
1. Volume Maximize (to get a starting point).
2. Compress (use the Reduce Peaks or Reduce Loud Parts preset).
3. Volume Maximize ("make-up gain" to bring-up the overall volume).
(Repeat steps 2 & 3 as desired).

[Coriolanus]

Thanks for the info. I will give it a try.