VirtualMIDISynth 2.x alpha available for testing

It seems you're using an AdBlocker or JavaScript is disabled!

This sites offers only FREE software and it's supported by a few advertisement boxes (no intrusive popups).

This 10 seconds wait is to let you update your AdBlocker whitelist...

Got it, show me the content...
Please enable JavaScript!

Pages

Posts: 59
Joined: April 19, 2014 - 06:23
Re: CoolSoft VirtualMIDISynth 2.0.0-alpha14 released

Corrected link from previous post: https://drive.google.com/open?id=0ByIDTUMUvYUWTDhBVHJGbElDLU0&authuser=0

- Main laptop: Sager NP4658, Intel i7-4810MQ, 2*8GB DDR3 (PC3-10700), 512GB+512GB SSDs, Intel HD4600/NVidia GT840M

Posts: 1544
Joined: March 25, 2012 - 01:19
Re: CoolSoft VirtualMIDISynth 2.0.0-alpha15 released

v.2.0.0-alpha15 - 2015-06-07

  • NEW: Status box in About tab now updates itself more frequently.
  • NEW: New statistic added to status box: BASS update rate.
  • FIX: Dynamic voices limit algorithm fixed, now it only increases BASS update frequency without touching voices limit. Also removed configuration option, now it's always active.
  • FIX: Changed description of CPU limit setting (in BASS it's not a CPU percentage, like windows task manager, but render/play rate).
Posts: 59
Joined: April 19, 2014 - 06:23
Re: CoolSoft VirtualMIDISynth 2.0.0-alpha15 released

This version, especially with the assistance of the rendering time statistic has helped me discover more things.  The "max rendering time" limiter setting does do something, but it seems to be doing a very mild job, compared to how it's been working in my MIDITest-mod.  I've attached that below along with various versions for several render time limits.

I wonder if this is related to what I also just found, where there is this half song-slowdown and half-skipping note effect when the rendering time statistic doesn't even reach 100% and rather below the limit such as around 60-80%; I admit this seems to be part of how virtualmidisynth lag seems to sound better on version 2, but now I also have the suspician that this is really just part of a performance glitch or something.

Also, this alpha still has this one annoying bug that I've been getting since Alpha7, where if virtualmidisynth lags hard enough, the synth gives up until it is restarted.  Further testing with these statistics seem to me that it doesn't totally crash, but it just doesn't produce voices when this happens, the render time counter is still reacting to the amount of MIDI Data it recieves.  It looks like all it takes is freeing the VMS port and re-loading the port for the notes to start becoming audible again when this happens (aka. "stop" and "play" in vanbasco's).  Also, I found if I throttle just the VirtualMIDISynth.exe process itself, it happens a lot more easily.

Well, I've attached a little testing archive I've crafted, I have a feeling it'll be useful for experimenting with performance.

 

Suggestion wise, I think it'd help to have a feature (in the about tab?) to allow the user to force-free a VMS port, so a misbehaving application such as chrome 43+.  And, for user friendliness sake, name the process(s) that a VMS device(s) is using in the "One or more VirtualMIDISynth devices are in use" dialog.

Attachments (Only registered users)
BassMIDI Performance Toolkit.zip

- Main laptop: Sager NP4658, Intel i7-4810MQ, 2*8GB DDR3 (PC3-10700), 512GB+512GB SSDs, Intel HD4600/NVidia GT840M

Posts: 1544
Joined: March 25, 2012 - 01:19
Re: CoolSoft VirtualMIDISynth 2.0.0-alpha15 released

@Ray890: thanks for your detailed reporting and test files; will test it ASAP but... I'll be on vacation next week, so this (lask) week at work could be really intensive, leaving little to no spare time...

Ray890 wrote:
name the process(s) that a VMS device(s) is using in the "One or more VirtualMIDISynth devices are in use" dialog
I agree, will add it.

Ray890 wrote:
allow the user to force-free a VMS port, so a misbehaving application such as chrome 43+
Could you please elaborate on this? What you mean with misbehaving?

Actually VMS (both 1.x and 2.x) has a fixed list of blacklisted "clients" that load the driver unnecessarily, like "winlogon.exe", "userinit.exe", "taskmgr.exe", "svchost.exe", "spoolsv.exe".
These processes try to load the driver (I wonder why winlogon.exe needs MIDI synthesis, but I suppose it loads the whole multimedia stack) and they keep it in use, i.e. preventing updates.
So starting from 1.x I added a "blacklist" of processes not allowed to load the driver DLL; this list is internal and not shown.
Should it be configurable?

Posts: 59
Joined: April 19, 2014 - 06:23
Re: CoolSoft VirtualMIDISynth 2.0.0-alpha15 released

Here's a chrome bug report I've done on Chrome (as also linked in the previous post): https://code.google.com/p/chromium/issues/detail?id=497573 .  Chrome simply doesn't free the MIDI ports once it is finished with it as in when a person leaves a WebMIDI website tab, it seems to be a bug that's already getting worked on.  Having a configuratable blacklist sounds cool, but I was imagining the ability to force-free a VMS port that is already in use without having to interrupt other VMS port(s) and/or restart the synth itself.

- Main laptop: Sager NP4658, Intel i7-4810MQ, 2*8GB DDR3 (PC3-10700), 512GB+512GB SSDs, Intel HD4600/NVidia GT840M

Posts: 59
Joined: April 19, 2014 - 06:23
Re: CoolSoft VirtualMIDISynth 2.0.0-alpha15 released

Three new bug findings:

- Seemingly new to this version, I hear a bit of random stuttering during non-intensive playback; The "update rate" is 10ms during these points, as opposed to 1ms

- If a VirtualMIDI port is requested while no VirtualMIDISynth.exe process is open, it will open but the requested port doesn't get used until the port is re-requested

- If you re-openVirtualMIDISynth.exe immediately (within 0.25 seconds?) after you tell it to exit via taskbar menu, a message saying "Another instance of VirtualMIDISynth is already running, but can't find it" shows then keeps an invisible VirtualMIDISynth process open, rendering VMS unusable until the invisible process is killed manually

Along with the "giving up" issue mentioned above, bug #2 and #3 in that list are actually the two very issues I've been having issue with for a while, but I had no idea what caused those two latter situations until now.  I honestly think VMS2 will be pretty stable at least for my usage after those reported bugs are fixed.

- Main laptop: Sager NP4658, Intel i7-4810MQ, 2*8GB DDR3 (PC3-10700), 512GB+512GB SSDs, Intel HD4600/NVidia GT840M

Posts: 35
Joined: April 28, 2015 - 02:54
VMS alpha 15 memory leaking?

Hi, is VMS suposse to release memory when changing SF or is eating it even when is not active the other SF in the list?

I get same memory usage until I take out SF from list thats is not active. Sometimes crashes If I dont do that.

Thank you

Posts: 1544
Joined: March 25, 2012 - 01:19
Re: VMS alpha 15 memory leaking?

@all: I'll be on vacation next week, so my activity could be near to 0 (it depends to the mobile connection I'll found to my destination).
Please be patient, I really need to have a break ;)

@Abner: yes, VMS is supposed to release all the memory allocated by soundfonts when unloading/reloading them.
Will check it out ASAP.

Posts: 35
Joined: April 28, 2015 - 02:54
Re: VMS alpha 15 memory leaking?

I wish you a very pleasant vacation. I post Coolsoft VMS link in KVR and some other links. Alpha 15 is working better than official VMS btw, and memory eating I get away by ejecting unused SF.

Thank you

 

http://www.kvraudio.com/forum/viewtopic.php?f=257&t=440392&p=6157906#p61...

Posts: 59
Joined: April 19, 2014 - 06:23
Re: VMS alpha 15 memory leaking?

Just thought I'd say I've temporarily reverted back to Alpha14, just because of the occasional 10ms update rate stuttering issue introduced in Alpha15, which ended up being worse of a problem than I initially thought.

Cheers

- Main laptop: Sager NP4658, Intel i7-4810MQ, 2*8GB DDR3 (PC3-10700), 512GB+512GB SSDs, Intel HD4600/NVidia GT840M

Pages