Two peculiarities with File-Export-Audio

You need to set a time range to export in addition to selecting the tracks. You might to to add a slight fade on the clips if the audio was not cut when the wave was at a zero crossing.

So, I know about the sounds you are talking about that end up at the beginning of bounces from where you last hit stop. Like the audio that was in the buffer just before you hit stop comes out of the buffer just after you hit play, even when you hit play in a new location in the tune.
 
In preferences you can set it to play tails after stop. This can help in a lot of cases. For some midi stuff, I usually click on a little icon near the clock in the control bar that resets the audio engine. This will always 100 percent clear your audio buffers.
 
When you are working with a very long project and exporting tracks, if you do File > Export > Audio > Tracks > Mono and uncheck most of the boxes about applying fx and automation etc, you can export the tracks of one tune to be imported into another project. It sounded like this was what you were trying to do.
 
Did you find that the audio buffer was an issue with this?
 
To avoid issues with the audio buffer, I often let an extra measure of nothing exist at the beginning of my project. Then if I bounce the project with audio in the buffer, I can just slip edit it out of the beginning of the track.
 
There are probably better ways. I think the main purpose of maintaining the audio buffer is to be able to restart from where you stopped and pick up right where you left off. Obviously it is problematic when you move within the timeline or execute a bounce to tracks where the audio in the buffer is written to the start of the new wav.
 
BTW, I never use Audacity for anything. I'm not sure why that is part of your workflow. You should be able to do everything in Sonar, unless you are changing to another OS (where Sonar doesn't work). On Linux I use Ardour.

gswitz
BTW, I never use Audacity for anything. I'm not sure why that is part of your workflow. You should be able to do everything in Sonar, unless you are changing to another OS (where Sonar doesn't work).

Agreed.  I was only going to use Audacity in order to get rid of the garbage in these WAV files, but I got to a work-around, so Audacity is not in the work flow at all now, I'm happy to say.  I'll experiment with your ideas for clearing that audio buffer.  I don't really understand why the buffer would not be cleared in every case automatically -- it could never be helpful, could it? -- but if there is a setting that will have effectively purge the buffer, that will be great.
 
And just for clarification, I wasn't trying to re-import the tracks.  I was just trying to put out WAV files so I could convert them to MP3s.  I would export them directly as MP3s, but SONAR is EXTREMELY slow at that.  Exporting to my SSD as WAVs is pretty quick, and then when I have all the songs exported, I can just run them through WinLAME to turn them into MP3s.  That's a lot faster than using SONAR.  But other than that, my entire workflow is under SONAR now.  I'm elated.  My improving workflow doesn't take too much longer than what I had previously done with Audacity (to break a concert into individual MP3s for each song) and the sound quality with SONAR sounds so much better.  Thanks Cakewalk !!
 
Now I am ready to move on to 8-track recording of concerts, which is only going to add about 20% more time to this workflow with SONAR, compared to processing stereo recordings  :)
 
One final question, if I may.  Part of my workflow is to split any stereo WAVs into separate tracks so that I can apply effects to the channels independently if needed.  For example, in my latest stereo recording, the bass trombone was obnoxiously loud in the left channel.  The bass guitar wasn't high enough in the mix.  I was able to compress the left channel and then boost the bass guitar with EQ on the right channel.  Even though the bass guitar was playing from the left side, most of the material in the right side was higher frequencies, so it worked out well.
 
With Audacity, there is a simple command to split a stereo track into two separate tracks.  I can't seem to find a single command in SONAR to do this.  So what I did was:
 
1) Clone the stereo track
2) Use Process-Effects-Gain to make the first track 100% left material, and likewise to make the second track 100% right material.
3) Convert stereo to mono for each of those tracks to save disk space, and adjust the pan settings hard left and hard right respectively.
 
There must be a faster way.  Am I missing something obvious?

Scook's got great answers.
 
The reason to have the audio buffer remain loaded is if you are going to restart your listening from the exact point where you stopped listening. Otherwise, I don't see value in it. If you are restarting from the exact point where you stopped, you have reverb all loaded into the buffers etc from audio previous to your start point.
 
I also think that VSTs cache audio outside of Sonar's control. The VSTs don't know where in the program you are. IMHO, Sonar should reset the audio engine (forcing the VSTs to clear cache) automatically when you move in the timeline after stopping listening. This would resolve this problem forever. Does anyone know why they don't? (I'm asking for a valid responses... nothing snarky).

Where a project restarts depends on how SONAR is configured. Better get used to the buffer issue, I doubt it will ever change.
 
Also, if you know in advance that you are going to split tracks to mono, why not record them that way in the first place? Alternately, it is possible to avoid bouncing tracks, send the left and right sides of a track to separate buses for processing.

scook
Also, if you know in advance that you are going to split tracks to mono, why not record them that way in the first place?

These are live recordings.  I tried doing some live recordings with SONAR on a notebook.  It was just too unreliable and too complicated for working under pressure.  So the present recordings are 2- or 4-track on a Zoom H4n.  I don't think that device has the option to save as mono tracks.
 
I am looking at the next step being Allen Heath ICE16, which can record 16 tracks.  It appears that box saves each channel as a separate mono WAV.

cparmerlee

These are live recordings.  I tried doing some live recordings with SONAR on a notebook.  It was just too unreliable and too complicated for working under pressure.  So the present recordings are 2- or 4-track on a Zoom H4n.  I don't think that device has the option to save as mono tracks.

There is an option on the audio import dialog to import as mono tracks.

There are a couple of per VST settings you can make in the PlugIn Manager that may help the buffer clearance (Suspend on play/suspend on stop IIRC).

Kalle Rantaaho
AFAIK using Simple Instrument Tracks you should export by selecting the track clicking the track number ( and also selecting the desired range in the timeline above, if necessary), not the clip.
I only use separate audio and MIDI tracks with soft synths so I'm not 100% sure.

Right.  When I click on the clip, it automatically selects the track and it also sets the time range to be exactly the same beginning and end times as the clip.  In that situation, File-Export-Audio produces the error dialog.
 
But if I were to do a  Select All (Ctl-A), in my example that would set the beginning of the range to the empty measure 1 and the end of the range to the end of the clip.  And that works.  It only fails when the time selection matches the clip exactly -- and obviously that is the case the really OUGHT to work.
 
I never noticed this happening until after I activated the MP3 license.  It seems unlikely to me that activating the license would cause this behavior, but it is possible.
 
And just for another curiosity, you might notice my clip is just a little short of 2 bars.  I copied that same clip to measures 4-5, which leaves a little gap at the end of measure 3.  If I select both of those clips, causing the range to go from the start of the first clip to the end of the second clip, it also fails.
 
Before I report this as a bug, it would be nice to know if anybody else can reproduce this problem.  It only takes about 30 seconds to set up this test.  I'll upload my project if that makes it easier.