So after all that, they still did not fix THESE bugs...

Well, that was the last time I'm ever going to bother spending any time documenting bugs in SONAR. I just tested with 8.5.1 and these bugs are still present.
 
Oh well, so much for hoping Cakewalk can get SONAR to do basic MIDI functionality properly...
 
In short, the bugs are:

• Bug #1: When a VSTi plugin is configured with Enable MIDI Output turned on, and it passes incoming MIDI events straight through, the incoming buffer of MIDI events sometimes become corrupted. This results in stuck notes.
  
• Bug #2: If the input of a completely unrelated MIDI track has its input set to that of an external MIDI controller, and that track is put into record mode, it not only records the input from the MIDI controller (as expected), but also what the unrelated VSTi is outputting at the time (not as expected).
  
• Bug #3: If a MIDI track's input is set to an external MIDI controller, and its output is fed into a VSTi that has Enable MIDI Output turned on, it will record ghost notes of everything that is played on the MIDI controller. This might just be the same bug as #2 above but just manifesting in a different way.

 
The ref number of the bug report is CWBRN-1336, and it was logged on 2/24/09.
 
I mean, what else do I need to do...?

EDIT: The main problem here is that up to now I have not been able to find a VSTi plugin that ships with SONAR that supports sending MIDI out (does anyone know of such a plugin?). So to demonstrate this, you need to use a 3rd party plugin that does support this. Of course this conveniently gives Cakewalk the opportunity to attribute this to bugs in 3rd party plugins and then to ignore the bugs. It is interesting to me to note though that all plugins that I have tested so far (that support MIDI out) expose these bugs. These include Jamsix 1, Jamstix 2, RealGuitar2, RealStrat as well as my own plugins as described in the bug report.

Malevol3nt


I'd like to report the following bug report:

Who cares? What does that have to do with anything?
 
EDIT: There, I fixed the website . 

kp


Doesn't Session Drummer 2?

I tried with it but could not get it do send any MIDI out (I could not find any settings for this). I'm going to try again right now...

I have not been able to find a VSTi plugin that ships with SONAR that supports sending MIDI out (does anyone know of such a plugin?)

How about the Arpeggiator plugin?

SilkTone

Bug #1: When a VSTi plugin is configured with Enable MIDI Output turned on, and it passes incoming MIDI events straight through, the incoming buffer of MIDI events sometimes become corrupted. This results in stuck notes.

Interesting. I've noticed this as well, I believe. Whenever I'm using a sequencer with Reaktor to sequence my Roland D-50, I almost always have to record it a couple times, because I get stuck notes very often. Though, I always thought that this was the D-50's fault, and not SONAR's because if I reload the patch or reset the D-50, the stuck note stops. Hmmm.... I'll have to look more into this as well.

candlesayshi

SilkTone

Bug #1: When a VSTi plugin is configured with Enable MIDI Output turned on, and it passes incoming MIDI events straight through, the incoming buffer of MIDI events sometimes become corrupted. This results in stuck notes.

Interesting. I've noticed this as well, I believe. Whenever I'm using a sequencer with Reaktor to sequence my Roland D-50, I almost always have to record it a couple times, because I get stuck notes very often. Though, I always thought that this was the D-50's fault, and not SONAR's because if I reload the patch or reset the D-50, the stuck note stops. Hmmm.... I'll have to look more into this as well.

 
The thing with switching patches is that often the synth will send an all-notes-off message. This will then also clear any stuck notes regardless of what caused them to become stuck in the first place.
 
I wonder how many times people run into these sorts of problems, but because a 3rd-party plugin is always involved, no-one ever suspects the bug to be really in SONAR. Especially with "bug #2" as described above, you'd be hard pressed to argue that the bug is not in SONAR. Even a buggy plugin should not cause MIDI events from an unrelated track to be recorded on a different track whose input is clearly set to be from an external MIDI controller. How on earth can you write a buggy plugin that can cause that behavior?
 
 
 

How about the Arpeggiator plugin?

It Rocks. Im playing with it. I hope i dont go blind
Cj

Wow.  You've made my day by reproducing that bug.  I've been going back and forth with Support for weeks.  I've sent about 10 minidumps and given them exact steps to duplicate the problem.  I sent them demo versions of the plugins and I'm still not sure they believe there's a problem here.
 
Finally last night I got Camtasia and recorded numerous crashes.  That's when I finally twigged to the multiproc thing.  I can crash it at will by turning on multiproc.

My work around for the VSTi crashes was to purchase energyXT and subhost all my MIDI plugins through that.  It's tedious to set up, but it avoids many of the problems.  It's also far more flexible to be able to route MIDI signals with nodes.  Even when they get the bug fixed I'll probably stay with this for complex routings.

Thanks again for the confirmation.

bitflipper

I thought about that too, but the Arpeggiator is not a VST plugin, and instead seems to be integrated into SONAR.

I'm not talking about the new arpeggiator feature in 8.5, but the old arp plugin (arpeggiator.dll), which looks like a VST (based on a peek at the file header), but obviously is not a VSTi.

Ah, I see. I just tried it with that arpeggiator and it does not reproduce. Most likely it is because that arpeggiator is an MFX plugin, not a VSTi as you already mentioned. So my theory is still that this is specific to MIDI buffers being passed around to VSTi plugins.
 
Somehow I have a feeling that Sonar's test matrix doesn't include testing MIDI out from a VSTi, or if it does, it is limited to simply verifying that they can select the VSTi's output as an input to another track, but it doesn't include a test case where they confirm that other MIDI tracks can still record properly when you configure it that way.

8.5.1 is a huge step for Cakewalk, setting them up for an amazing next feature drop. Yea, some bugs aren't fixed yet, but there's not a lot of people working on this software (check the credits easter egg) and they have other projects going on that in the next year will come to fruition and make the Sonar experience a million times better, IMHO. They haven't let us down yet, so don't lose faith. I haven't, and I'lve been on for a long time with Sonar.

to OP:
Hmm dont know but I just checked with Catanya full bought version and I followed your procedure but still there's no MIDI where it shouldnt be.Pehaps Im not following your procedure correctly?
Let me clarfiy:
-Catanya has on MIDI track - MIDI output enabled and input set to receive MIDI from my MIDI keyboard
-Z3ta+ has an audio track and a MIDI track set to receive incoming MIDI from Catanya's output
(by having set MIDI in -  Catanya OMNI
Of course z3ta+'s MIDI track is set to "echo on" )
-I then proceed to add another MIDI track which is set to receive MIDI from my MIDI keyboard.
Then what do I have to do to witness the bug? I tried recording the lastly added MIDI track by hiting keys in my keyboard and indeed it DOES record ONLY what I play on my keyboard.

thegeek


to OP:
Hmm dont know but I just checked with Catanya full bought version and I followed your procedure but still there's no MIDI where it shouldnt be.Pehaps Im not following your procedure correctly?
Let me clarfiy:
-Catanya has on MIDI track - MIDI output enabled and input set to receive MIDI from my MIDI keyboard
-Z3ta+ has an audio track and a MIDI track set to receive incoming MIDI from Catanya's output
(by having set MIDI in -  Catanya OMNI
Of course z3ta+'s MIDI track is set to "echo on" )
-I then proceed to add another MIDI track which is set to receive MIDI from my MIDI keyboard.
Then what do I have to do to witness the bug? I tried recording the lastly added MIDI track by hiting keys in my keyboard and indeed it DOES record ONLY what I play on my keyboard.

You might try or confirm the following:
 

• Make sure you have some events on the Catanya MIDI track, and make sure Catanya sends out events. Don't rely only on your external MIDI keyboard to trigger Catanya for this test.
• Make sure the last MIDI track where you are recording from the external MIDI controller has its input set to "External controller - OMNI". I found that when you select a specific MIDI channel as the input, as opposed to OMNI for the keyboard, it does only record the external data as expected (but see note below). 

 
Note: Even though you can select a specific channel on an external keyboard as input, and that this is somewhat of a work-around, this is still a bug because selecting OMNI for a keyboard is valid as some keyboards send multiple channels. Also, this only masks some of the manifestations of this bug.

to OP:

bug confirmed - I had not recorded any content in Catanya's MIDI track in my previous post.Tried today and what happens is pretty much what the OP describes.A MIDI track receives MIDI messages from the MIDI out enabled plug in although its input is specified as ONLY from MIDI keyboard

SilkTone

tvCasualty


    it's just not realistic to say the Bakers "haven't let us down yet".

You're acting as if they get your bug report and say "Meh..."
I'm sure they get so many show stopping bugs, and with only a handful of developers you can't expect them to get to every bug for the main release. That's what patches are for.

 
The only thing Cakewalk uses patches for is to fix the most serious of bugs right after a major release. The non-showstopper bugs are always put on the backburner, in some cases for years. And that is why people are saying that a 1-year release cycle is too fast. They barely fix bugs that have been around in multiple versions. Every time they release a new version, there is a flood of people complaining about bugs that have been present for multiple versions and have once again been skipped (the envelope bug comes to mind). Instead of adding new features, why not release patches and fix those bugs instead?
 
And a point release, which is what 8.5 is, is supposed to be focused on bug fixes, not new features.

I think with the advent of 8.5 we can forgo any previous idea of what a "point release" is.

Cakewalk have moved the goal posts, but it's not the end of the world.

Hopefully the positive side will be they now have more time to concentrate on some bug fixes.

Fret Wizz

 

Maybe Malevol3nt  was just pointing out the fact that bugs can appear anywhere,
even in the web page of someone who is complaining about bugs in another program.

I wonder what your reaction would be if Cakewalk replied as you have in this post?

 
 
Yes the irony wasn't lost on me, except it is completely irrelevant in this case, because:
 

• No-one pays to visit my website that doesn't even contain anything useful, in contrast to me paying Cakewalk $179 each year for software (ok, $119 this year, but also add in the initial price of the full version - must be approaching $1500 by now).
• I fixed the problem almost immediately, in contrast to Cakewalk that doesn't seem bothered to look into the problem.