I've moved this post to the page more likely to be seen... I spent to much time preparing it to have it instantly relegated to the "old" page: :-)
I have prepared these illustrations to answer Jonbouy's question about how fixing a peak effects the transient.
Maybe we just have a different idea of what the important part of the transient is... and that's fine... but here's an illustration of my idea:
Here's an exported 24/44 .wav file of 4 bar drum loop I made with Session Drummer 2. I purposely set the velocity on one snare hit so that it would hit an over. This track is a mix of Kick, Snare, and Hi Hat. The over is the snare... but there are harmonics from the simultaneous Hi-Hat hit laying in the wave as well. I would normally have a dedicated snare track (which makes the edit even easier), but this might simulate some bleed from the Hi Hat (which makes the edit a bit more tricky).
Here's the front of that hit. I've selected 4 *primary* cycles to work on:
I want to preserve the character of this portion so let's just bring everything in the selection down 1.4dBFS:
Here's the result of that operation, it's very clear where the clipped material is:
Here's a zoom in to show the 4 actual clipped peaks. The longest clips is apprx 25 samples long. That's 0.000566893424036 seconds (a.k.a. 0.57ms):
And here they are after I used a hand drawing tool to edit the clip portions into facsimiles of what I suspect the wave form should have looked like:
A zoom out and a test play reveals the edits are indiscernible:
2 things:
1) I wouldn't have much optimism about using this technique to fix something with rich harmonic content.
2) If the track was a discrete snare hit the edits would have been shorter and easier.
So, in my opinion the transient, both the rise and fall, has been preserved as faithfully as possible.
I haven't had to do this sort of work in the since March 2007 (I wonder why I remember that so well???? Great Jazz trio, live audience, drummer playing wickedly soulful brushes but with a stick in his lap for quick rifle cracks), but here's what I'd like to emphasize... when I'm confronted with this sort of scenario I get to work and don't think for a moment how long it's going to take. This is just part of the job and I wouldn't trust this sort of repair to any automated process that I'm aware of. I don't know why I'm up for hard work like this but it would just offend my sensibilities to seek out an easy out solution that treated the sound with less respect.
Now I want to emphasize... if there is any practical way to record a retake while maintaining the vibe et al. I would, without a doubt, use that option rather than perform the edits I have just described.
I'm just discussing this specific process so that the OP (or anyone else interested) gets a more complete overview of the options available.
I don't know if any of this applies to the OP's situation but that's what I've been speaking of.
very best regards,
mike