VirtualMIDISynth 2.x - Release Candidate 2 released

Jaxxoon Yeah wrote:

However, I've recently noticed that setting a soundfont's volume level to any value under 100% mutes it completely for some reason.

It was a bad casting, easy to fix and just in time for Beta4.
Thanks for reporting it.

Thalyn wrote:

Unfortunately it's been that way since the feature was first added during alpha.  I was certain I'd actually reported it way back then but I can't for the life of me find the post in my history, though I can see Supergod referencing it on April 30 last year (alpha 13).

Can't also find it in my TODO list... who knows ;)

Here are some interesting thoughts from Chrome:
https://code.google.com/p/chromium/issues/detail?id=580522
 

coolsoft wrote:

 
As a quick workaround I'll skip that message only for chrome.exe, waiting for a better idea to come up.
 

toyoshim wrote:

I'm not sure Chrome really open the same device twice. I believe Chrome closes it before opening it again even in the case I mentioned at #7. Otherwise, other physical devices should not work because Windows does not allow to open the same MIDI device while the device is opened. But, there could be virtual synth specific race issue at the driver stack level. Anyway, I agreed that this bug should be open for a while.
 

toyoshim wrote:

 
I find this thread; http://coolsoft.altervista.org/en/forum/thread/222, and now I'm suspecting that this problem is a race issue inside the VMS2. Here is an excerpt that I'm interested in.
 

VirtualMIDISynth 2.x indroduces a deep and important architectural change: Driver and Synth components have been splitted in two distinct parts, while the Configurator is now included into the new Synth. The driver is still a DLL, loaded by the MIDI application as 1.x, but now it only contains the "windows multimedia driver" parts needed to identify the virtual MIDI device ("CoolSoft VirtualMIDISynth") to the system and receive MIDI streams from MIDI applications outputs. The received MIDI stream is then forwarded to the new Synth component: a new, isolated and independent process that listens to VMS2 driver instances and plays the received MIDI streams. Syncronization between multiple driver instances could be done easily, from a single point.
 

 
So, my guess is when Chrome close and open the "Driver" part in a very short period, the open request happens before the previous close request finishes closing the "Synth" part that my be processed asynchronously.
 

Steve Wellens wrote:

Sometimes, I would loop a track and then change the patch on the fly to see how different instruments sound...

Steve Wellens wrote:

Try this link:  http://stevewellens.com/MidiProblem.avi
Sorry it's so huge.

Still can't reproduce the bug on my side, sorry.
I've trued both with .MID and .KAR files but when I right click that column the instruments dialog appears successfully without any issue on the playing track.
I'd like to try with the MIDI you're playing in the video; could you send it to me? Use the contact form if you like to keep it private.

BTW: this is the configuration of VMS in my Record Producer:

coolsoft wrote:

While working on it I found a better "workaround" for the "Device already in use" message that works for any other process:

• If the driver is used by the same PID (processID) that is trying to reopen it, instead of immediately showing the message it'll wait up to 2 seconds for the previous CLOSE to complete.
  If still locked, then the message is shown.
• If the driver is used by another PID, the message is shown immediately.

Attached here you'll find a special Beta4 version with these fixes.
Since they're strictly related to this "Chrome bug" I haven't created a new Beta for it.

Please let me know if it fixes the issue.

Bravo! I installed the beta4-chromebug version and the "Device already in use" message still does not appear when I switch my keyboard on/off in Chrome.

It sounds to me as if toyoshim was actually right. Your fix fixes what I think he said was the problem. Maybe you misunderstood him -- but we now have the right solution anyway. Great stuff! :-)

I'll tell toyoshim.

notator wrote:

Bravo! I installed the beta4-chromebug version and the "Device already in use" message still does not appear when I switch my keyboard on/off in Chrome.

Thanks for testing and reporting it, will integrate these fixes in next Beta5.

tomekleon wrote:

Does in VMS 2 will be able to change in real time on individual instrument sounds MIDI instrument tracks (something like in attach. pic)? I really want this. Please.

This feature was already requested in the past but I delayed its implementation because I still think this feature is much more "MIDI Player" than "MIDI (virtual) Device" related.
I mean, changing the instrument should be done at the player side, not at device side.
Anyway I'll keep it on my TODO list and we'll rethink about it after first official 2.0 version will be released ;).