iZotope Declipper vs ReLife

[marmistrz]

Hi
I came across a plugin that also performs declipping: ReLife (http://www.terrywest.nl/utils.html)

What do you think about it? Is it better that iZotope's declipper for you? (it's better at at least this aspect: IT'S FREE!)

[metal4life]

I've played around with it and I really like it so far and it's WAY faster than the declipper I've used previously.

[marmistrz]

have you compared this with izotope?

[jNive]

sounds interesting - I'll need to remember to give this one a try soon!

[44100hz]

Let's see if ReLife actually does anything, shall we?

I took a sample from an album with gratuitous amounts of clipping, just to see what would happen. Most fake de-clippers just modify the songs in a way that makes them sound crisper, and use something else to cover up the peaks. The way I test this out is by looking at the clipped peaks, and seeing if they have been logically reconstructed.

So, I took out a short sample which has a lot of clipping in it, and one with plenty of audio information next to it to reconstruct it if that is what is happening.

compare.PNG (49.86 KiB) Viewed 881 times

Nice. As you can see by the hardness of the lower sample, the clipping has not been removed, the audio has just been scrambled around a little. I can get the same visual results from high-passing at piece of audio at 10hz 48dB rolloff. It would probably even look "better" if that is the sort of "improvement" taking place.

In fact, it is pretty obvious that there is an exciter being used to "reconstruct" the high frequencies, as using this adds a little bit of visible bias, which indicates even order harmonic enhancement is in play.

In other words, you can get the same results by high passing a piece of audio in a way that doesn't alter the sound, and then applying an exciter to it all. Whoops.