Har-Bal Users Forum BBS

Hi

I am currently working with files that are over 40 minutes long (Electronic music pieces from the 70's and 80's).
I've tried to Harbal the whole file with some success but on a whim I thought I'd try to break the music into 10 to 15 minute pieces (at appropriate points in the music) and re-Harbal each section. This turned out to be way better but I did have transition problems at section joins.
Anyway, to the request: ( I don't think anyone has requested this)


I know in the past, a request for splitting the wave was suggested. Would it be possible to implement this. Also, would it be a possibilty, to implement spectrum morphing over a specific time interval at each split point I'm not talking huge changes to be morphed, but currently, without a timed morph, it is obvious at some splits splits that something has changed.

Cheers...Thanks for any reply

That would be pretty cool, actually this idea could be extended to some sort of adaptive approach that even could be useful for for short "normal" length tracks were the style varies (say a clean piece in a metal song).

For your problem I guess the intermediate solution is to make your selection within each track a bit wider so you can crossfade btwn the pieces when you combine them again...

DuncF,

What you are requesting is on the plans for the re-write. I've had a mind to do this for ages but it really needs a full re-write so that is why it is happening now rather than as a change in the current code base. Be that as it may, this re-write is a big job so don't expect to here much about it for at least another 6 months. I've made some good progress on fundamentals but there is a hell of a long way to go...

By the way, your request to have morphing is not a big issue. The filter engine essentially already does this and that is why you don't hear a click when you change the filter on the fly. It's just everything else needed for track splitting isn't there...

Paavo.

Hey Paavo that sounds good to me.

To be honest I kinda thought it would be too difficult to implement without noise artifacts creeping in.

If you need anybody to Beta this thing in the next year or so, I'm your guy