dexed
dexed copied to clipboard
Clicking on some patches
Taken from here: https://www.kvraudio.com/forum/viewtopic.php?p=6795306#p6795306
Looks like there are some patches which suffer from a clicking sound at the start of the note.
Contrary to what the poster said, I've managed to trigger the clicking issue also on Modern engine, although it is less pronounced and doesn't happen every single time.
I'm attaching the Sysex file here. The problematic patches are:
PHARAOH ENCOUNTERS
The patch "OB GenvivY" clicks only on successive notes and only if you play them at a certain time after the previous one.
I've found another patch that clicks really badly, the BASS 1 patch on the standard ROM.
I've attached two MP3 files (converted to MP4) so you can hear the difference between Modern and Mark I. You'll probably notice that the clicking is still present to some degree in Modern engine, but not as prominent as Mark I.
First one is Modern, second one is Mark I.
https://user-images.githubusercontent.com/51792528/222212472-d4cd47c3-f8f3-4d7d-81b7-0705e8a765a0.mp4
https://user-images.githubusercontent.com/51792528/222212478-31f365cf-f5d8-46c2-899c-8cd4ef56332a.mp4
Hi Luca, in the lower audio file (16sec) I can hear an additional click in the attack. The other audio (18sec) i can hear in release the same click but a little bit quieter
I hear this click as well on BASS 1 and other patches. The only way for me to remove the click is to turn the EG Level 1 down in the 1st operator.
I have that issue (default bank "Say Again" patch), too. 24-bit res engine somewhat less pronounced. It's on A LOT of patches. Primarily at key release.
Here is what I think (first guess). There are dx7 comparisons on youtube and the general consensus is that te original sounds warmer.
a) psychoacoustics - sudden changes in amplitude (loudness) have no clear frequency and are therefore perceived as noise. does anyone have a technical comparison whether dexed replicates correctly at extreme attack release settings?
b) I assume dexed uses mathematical models for the waveforms. I also assume the original isn't able to perfectly match those models due to electronics. I wonder to what extent electronics is impacting this.