|
|||||||
![]() |
|
|
Thread Tools |
|
|
#1 |
|
DriverHeaven Junior Member
Join Date: Sep 2005
Posts: 38
Rep Power: 0 ![]() |
3538J midi issue
Big thanks to all the Team for the wonderful piece of software.
Testing midis with my A2 ZS + Airfont 340 (78 Mb), I found strange behaviour, like some instruments pushes out others (as if there not enough channels). For example a title track from 'The Secret of Monkey Island' sometimes lacks flute. This can be fixed by Settings -> Synth Compatibility -> Check Legacy Release/Decay. Is this sound bank or driver issue? Other banks seems to be Ok, so I'm unsure. |
|
|
|
|
|
#2 |
|
DriverHeaven Lover
Join Date: Jan 2005
Location: Italy
Posts: 192
Rep Power: 0 ![]()
|
... the same here...
|
|
|
|
|
|
#3 |
|
Apple Fanboy?
|
IIRC onlu upto 32mb banks can be played at once
__________________
Chris - The Aussie Super Mod
Hardwareheaven Rules - Sig Request Thread How you can help HardwareHeaven by using Digg! Hardwareheaven Super-Moderator |
|
|
|
|
|
#4 |
|
Tail Razer
Join Date: Jun 2005
Location: Bernyurass, AZ - USA
Posts: 4,027
Rep Power: 0 ![]() ![]() |
Hmm - I just tried a 'busy' (read: uses LOTS of polyphony) midi file & an 8MB GM Soundfont. - and it seems 'note stealing' is happening pre-maturely.
Im wondering if 3538J - is splitting polyphony between the 2 synths now. - But, something did change. Im also wondering if theres a register that could display 'the number of synth notes' active at any one time? Id have to do more testing - But another possibility is; the 'priority' for channels have changed - could cause this to behave differently. But Im definitly hearing note stealing that DID NOT happen in 3538H. |
|
|
|
|
|
#5 |
|
DriverHeaven Addict
Join Date: Jun 2003
Posts: 257
Rep Power: 0 ![]() |
Maddog, radiocolonel, ONjA: If you open the kX mixer and select the "analyzer" (the bottom button on the left) it will give a graphical representation of how voices are being assigned. It should be fairly straightforward to see if kX is doing things in the most economical way.
I wonder if the change in note stealing behavior is put in place to combat the "beeping" issue? |
|
|
|
|
|
#6 |
|
kX Project Lead Programmer and Coordinator
Join Date: Dec 2002
Posts: 3,119
Rep Power: 75 ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
no, it is unrelated to 'beeping' issue (although I seem to suffer from it on my notebook now -- but only after resuming from hibernation - and it is easily solved by kxctrl -rv or disable/enable in the dev. man)
--- I will check if pre-3538j algorithm was better - perhaps, some 'forgotten/experimental tweaks' not still removed E. |
|
|
|
|
|
#7 | |
|
Tail Razer
Join Date: Jun 2005
Location: Bernyurass, AZ - USA
Posts: 4,027
Rep Power: 0 ![]() ![]() |
Quote:
This is good to know - but I never looked in 3538h - so I cant compare. But it DOES happen right after a boot-up. I found that putting a check in 'synth Compatibility' - Legacy release/Decay - resolves this. edit: - doesnt resolve - but helps. @Eugene: It doesnt seem any Channel Priority is implemented - is this correct? it *seems* it just uses a FIFO algo...?? But its hard to be certain. If theres IS priorities - AFAIK what is typicalwould be: 10, 1, 2, 3, 4, 5, 6, 7, 8, 9, 11, 12, 13, 14, 15, 16 Ch 10 being highest priority. - is this the case? If this is not so, we may need to 're-arrange' sequences to accomodate a change in note-stealing algo. Last edited by Maddogg6; Feb 3, 2006 at 04:33 AM. |
|
|
|
|
|
|
#8 |
|
Tail Razer
Join Date: Jun 2005
Location: Bernyurass, AZ - USA
Posts: 4,027
Rep Power: 0 ![]() ![]() |
Additional:
It seems (from looking at the 'KX anayser' that the release times could be preventing 'expected' results. Changing KX to use legacy release/decay - doesnt appear to change how it looks in the anayser - but seems to 'better' choose which 'NOTE TO STEAL'. I would need to revert to 3538h - and do further testing to see if I can see/hear whats different/changed. |
|
|
|
|
|
#9 |
|
HardwareHeaven Newbie
Join Date: Feb 2006
Posts: 7
Rep Power: 0 ![]() |
same here
I'm having the same problem here. It happened too with ver.3537 but only on 1 or 2 songs, now it happens on almost every song... win2k and SB0102 here... (if i switch to a 8mb/4mb soundfont, it works well)
The soundfont that produces the problem is Chorium (27mb), maybe the problem is there... as with other sf's it doesn't happen. I noticed that with this SF the bars on the Analyzer take longer to drop down, wich leads to kx not playing new notes cause it's out of voices. Maybe someone could point me to a good-sounding soundfont that doesn't has this problem? Hope it helps. Ferow2k. Last edited by Ferow2k; Feb 27, 2006 at 06:57 PM. |
|
|
|
![]() |
| Thread Tools | |
|
|