MMultiBandHarmonizer change in sound between versions?

[ericzang]

I am suspecting that there has been a change in the Harmonizer's "vocoder" sound compared to version 1, or perhaps compared to version 2 or 3. Unfortunately I don't have the older versions anymore, I am on the latest 4.01c.

May I be sent version 1 for my own testing purposes?

I recall that there was only a very subtle difference between the 3 different vocoder quality modes. Now I hear a big difference between all of them, and I'm not sure it is a good change.

I think I preferred the old sound, but I am not sure, perhaps I am remembering wrong. I would like to test them next to each other.

Thanks very much!
Eric

[VariKusBrainZ]

I thought there has always been noticeable difference in sound between algos.

Are you using the same oversampling setting as this can have a huge effect on the sound?

[ericzang]

Do you mean the the "High quality upsampling" option inside the Settings? Changing that I hear no significant change.

One of the problems I am noticing now is that the sound quality changes depending on pitch. Testing with my voice, some pitches are not even coming through.

[MeldaProduction]

Hi Eirc,
I checked and found no difference, but maybe it depends on audio material. I'll upload some previous version, for which platform you want it?

Cheers!

[MeldaProduction]

Wait, I'm taking back, there seems to be some change after all, I'll check it out and let you know.

[ericzang]

Thanks for checking it out. I'm on Windows 7, using 32bit in ableton. (64 bit for sonar, but not using it there for now).

[MeldaProduction]

Ok, found and fixed, hopefully. Please check it here:

http://www.meldaproduction.com/download ... onizer.zip

Please let me know how it looks.

[ericzang]

Great! Thank You!!

[ericzang]

I run at 32 samples in Ableton for live performance (using RME digiface). I'm getting a few pops now with the fixed version. I don't think there was that problem before.

[MeldaProduction]

Ok, I reuploaded the file just in case. But honestly I don't think you can make it work with 32 samples latency, that's just too low. On MacBook I wasn't even able to play a WAV file in that latency .

Usually 256 is fine for realtime. Sometimes I go for 128 when drumming, because there is some MIDI latency as well and attack of drums is just too fast. But 32 samples means less than 1ms, that's much less than you will ever be able to notice, and I believe the plugin was never able to handle that (actually I believe most plugins in general would have a very hard time to do that).

[Sampleconstruct]

Ok, found and fixed, hopefully. Please check it here:

http://www.meldaproduction.com/download ... onizer.zip

Please let me know how it looks.

Is that like an "official" fixed version for all users?

EDIT: Ahh, I see it's a Windows only file...

[ericzang]

The last upload has the sound quality problem again. Fortunately I still have the good one you first uploaded.

The clicks are inconsistent, I'll need more time to play with it. At the moment I've only gotten a few pops at 32 samples with 4 shifted voices activated. Maybe it is even better when using only 1 voice.

[MeldaProduction]

Are you sure about that? I just checked and it seems fine. Ok, try this one:

http://www.meldaproduction.com/download ... nizer2.zip

I just renamed it, just in case your browser cached it or something. Then please try to compare it to current release, which you can download the ordinary way. I believe the performance (and clicks for too low latency) are the same and the quality problem has been fixed.

[ericzang]

Hmm, this last one also has the poorer sound quality. I restarted ableton and have the .dll renamed, so I don't think things are getting mixed up.

[MeldaProduction]

And is it the same problem as before? Anyway try also to delete the file and start Live without it. Then put it back and start it again. Obviously in the first case Live shouldn't find it.

If this won't help, please send me some audio files with the good sound and bad sound to info@meldaproduction.com. Also potentially the original audio.