microapp
Fix musicXML export bugs.

I was not aware MusicXML was supported in any form.  That is good to know.  My interest is on the import side, but export would be useful for somebody who composes in Sonar and wants to produce a polished score in a proper notation program.
 
Can anybody share real world experiences exporting a Sonar project into an external notation program with MusicXML?

From other posts, it seems even the export musicXML is not really functional.
IIRC, the poster said even the beats in each measure did not match the meter.
I have not confirmed this yet.
THere is not import whatsoever.
 
 

Hard to tell if this one is SOnar or MuseScore issue.
http://musescore.org/node/18892
These are definitely Sonar related
http://forum.cakewalk.com/Sonar-X1-Expanded-Problem-with-MusicXML-Export-and-MuseScore-m2418439.aspx
http://forum.cakewalk.com/Sonar-MusicXML-gt-Sibelius-m2435599.aspx
http://forum.cakewalk.com/MusicXMLAnyone-having-any-luck-m2437717.aspx
 

microapp
From other posts, it seems even the export musicXML is not really functional.
IIRC, the poster said even the beats in each measure did not match the meter.
I have not confirmed this yet.
THere is not import whatsoever.

 
Well, it is a starting point.  At least somebody at Cake has some code that does something with MusicXML.  A proper exporter would need to have a lot of options, such as quantizing, otherwise the resulting score would be a real mess. 
 
Initially I assume that the score would only contain notes that were in MIDI in the Sonar project, but eventually one could imagine hooking the Melodyne stuff in so that audio tracks could end up in the score.  That would require some thresholds to keep the noise factor down.

A true implementation of musicXML would need to convert articulations to MIDI and vice versa.
Also meter and accurate interpretation of things like dotted notes,triplets and ties.
Without this, you may as well import/export MIDI which is not very accurate as far as notation goes.
 

microapp
A true implementation of musicXML would need to convert articulations to MIDI and vice versa.
Also meter and accurate interpretation of things like dotted notes,triplets and ties.
Without this, you may as well import/export MIDI which is not very accurate as far as notation goes.
 

Are we talking the import side -- bringing in a score from a notation program?
 
That's why I suggest that the import process should have the option to obtain the schema from the MusicXML, but get the notes from a MIDI file.  In Finale, you can cause the Human Playback to save its controller commands in the MIDI file.  That takes care of all the articulations, crescendos, trills, rolls, fermatas, tempo changes -- everything.  Much of that is not in the MusicXML per se, and could only be inferred if the import process were as sophisticated as the Human Playback.  It seems to me that it would be a lot easier just to grab the MIDI that the notation program has already created.  Even without Human Playback, that MIDI should adjust for accents, staccatos, slurs, etc.

It applies to both although I would prefer musicXML import be more robust since a composition done in Finale and imported to Sonar Staff View/MIDI is likely to be more notationally complex than the other way round.
I doubt SV will ever be much more than it is now unless Cakes pairs up with a 3rd party notation company.
This brings up all kinds of issues. Would XML import even involve the staff view (a simplistic SV) ? Would XML import be to MIDI and then the MIDI is displayed in SV (the limited one we have now)? I am not sure if I had a robust XML import, I would even care about SV.

microapp
I am not sure if I had a robust XML import, I would even care about SV.

I think import and export are two entirely different audiences -- two entirely different jobs, even if there are similarities in the technical process.  For import, my interest is only in improving the final playback.  Finale with Human Playback is pretty good when used with the Garritan libraries.  But there is always something that isn't really quite right.  There is an ability to modify the MIDI directly under Finale, but it is tedious -- basically the same kind of no-support feature as notation-view in Sonar.  Moreover, Finale doesn't support VST effects at all.
 
I would love to have a process that would, with a few minutes' effort on my part, transport the essentials of my Finale score (and Human Playback) into an equivalent Sonar project, with some reasonable selections for synths on each track.  In my dream world, the XML import would pop up a dialog that shows the synths it is planning to set up by default and give me a chance to override individual tracks, or apply a template that I had created earlier.
 
Or maybe there would be a mapping file that says "Whenever an XML instrument looks like a harp, then set up this synth.  Whenever it is a trumpet, do this, etc."
 
I can do all of this now by hand of course.  But the MusicXML file would have all that additional information about tempo changes, key changes, meter changes, etc. 
 
I don't mean to trivialize this.  A proper importer or exporter would be a lot of work.  But it would open up a whole new market of customers, and in the end, it would probably be a lot more efficient than trying to build up the score view.  We know how that would go.  If Cakewalk were to address the first 15 issues, there would be another 30 issues brought up because the notation view is just not a general purpose notation program.

Here is a summary of a test I did last night.  As I have mentioned up-stream, my greatest interest is on the import side in order to get the best sounding renderings of compositions I do within a notation program -- Finale 2014 in my case.  Finale can do a pretty good rendering when using the Garritan libraries.  Finale proper allows precise control over interpretation of articulations like staccatos and marcatos.  The Human Playback feature does a great job of introducing volume control and tempo variations to interpret most of the markings in the score.  Finale includes a basic mixer and the Garritan Aria player has a basic reverb.  The results can be acceptable, but nothing like what you can do with fine-grained control in a DAW.
 
I was always afraid of of importing Finale's MIDI into Sonar because:
 
a) That takes a lot of project set-up
b) I was afraid it would be a real mess and would require a lot of touch-up after importing.
 
So as a result, I have done my rendering in Finale, and sometimes send the resulting WAVs into Sonar or Ozone for some refinement.
 
I had an arrangement that just wasn't working out using the Finale rendering.  I could never get anything like a good mix.  So last night I bit the bullet and did the process of sending all that MIDI into a Sonar project.  Here are my observations:

1. It took about 2 hours to get the Sonar project set up with all the right routing.  But the next time I do a project for a similar instrumentation, I can reuse that template.
2. The MIDI actually was pretty clean.  The keyswitches for mutes did not come through.  I had to add those manually.  There were a few cases where the volume controller actions didn't work out, but they would be fairly easy to edit within Sonar.
3. The overall result was very successful, with the end product being a far better mix than I could get through Finale.

So it is definitely viable.  The issue is the amount of work involved.  I don't believe that tempo changes or meter changes come through.  I had none of that on this project, but if I had to fix that manually, it would be a lot more work.  IMHO, this is too much work to do at intermediate steps of the project -- and that is when I really need good renderings to review with the client.  It is probably worth the effort for a one-time process once the Finale project is frozen as a final deliverable.
 
So I believe more than ever that this is a perfect candidate for automation to make it quick and easy to render into Sonar at any stage of the Finale (or any other notation program with good support of MusicXML and MIDI) project.
 
If anybody is interested in comparing, here are the two MP3s.  Please note that I am not going for the perfect mix, just a mix that reasonably illustrates the arrangement, making all the parts audible.  I can't do much better on the Finale version, but I could do a lot better on the Sonar version with another hour or two working on it.
 
Render in Finale: https://app.box.com/s/t7h24tjnub0fn59p9b7mq0qpbhhx5j5y
 
Render in Sonar: https://app.box.com/s/b76wfqky74no3drnxo2rsgojv95n94n0
 

cparmerlee
That is quite a bit of work to go through! The Sonar rendering is quite a bit better IMO. We shouldn't have to go through the grief to import/export from Sonar/external programs. I hope CW is listening....