So I am finished with a track made up of all softsynths.   Is it best to bounce each track to an audio track and then maybe hide the MIDI?   
 
I understand for safety, it would probably be best bet, but is there any difference on the audio side I need to worry about, or can I start my mix phase right from soft synths, then export to audio for mastering?

Thank you!

The only benefit I know of would be to free up some system resources by freezing the synth tracks. If there's plenty of horsepower on your system and the project can handle playing all your synths in real time without any glitches etc then I'd leave it as is. FWIW I've always left all my midi tracks as is without issue.
 
(Having said that, I have had one recent weird midi export experience - there's another thread on that if you're interested - but not sure of the cause. It decided to fix itself before I worked out what happened! )
 
 

Down the road a few years, when your system and plugins have changed, the midi based soft synths may not load up the same as they used to. If you have frozen tracks, or bounced to tracks, at least you still have the audio rendition in place to carry on.
 
"can I start my mix phase right from soft synths, then export to audio for mastering?"
 
I have done both, and have found out that mixing and mastering don't always get along inside one project.
My opinion is that it's best to export to audio, then master that exported audio in a new "mastering" project.
That said, I do have simpler projects where mixing and mastering inside one project makes total sense.
 
It's also been stated many times on this forum that you can export the audio at a higher bit depth, then use dithering when mastering to a lower bit depth (depending on the medium / end result needed)
 Advantage ? . . . some say it sounds better doing this.

I take it that bouncing midi to audio means freezing it.
 
In my case, it's only necessary for Session Drummer 3. Many times in my project during looping or start/stop editing on other tracks, SD3 will lose audio for a while, like it's muted. Then after a few bars, audio shall be back again.
 
And if I inserted a few instances of SD3, I must freeze the SD3 tracks that I'm done with, other wise Sonar may hang - Edit : upon next startup of the same project with multiple SD3s.

What I do is freeze all midi tracks, copy and paste them to a new (audio) track and name them, unfreeze the original and archive it and drag it to the bottom of the screen.  Then I work from an all audio version.  That way it is simple to get the midi back and I'm working all with audio.  That is how I learned to work - on tape.  And early digital, where the computers couldn't handle more than one soft synth or reverb.  But it just seems neater to work in all audio.
 
But that is the great thing about DAWs and home studios, you can develop your own best practices.  If you prefer midi and it doesn't choke your computer, there is no reason not to work that way.  Whatever makes you feel comfortable.  And yes, you just export the entire song and SONAR converts the midi to audio w/ no action necessary on your part.
 
@

I think so, I do it all the time, usually as soon as possible. I like to separate building a track into distinct phases: composition and sound design, mixing and effects processing, finalization and mastering. The boundaries tend to be a bit fuzzy for me because I like to start mixing as soon as I've got enough of the track programmed to make it worth mixing. A lot of times I'll listen to the sofar mix and decide I want to add a sound to it, I wouldn't be able to make that decision without hearing the mix.
I do a lot of pop music and I generally like a hotter signal so I set everything to around -12 dbfs for each distinct sound I want in a mix, I've found that for the plugins I use the most -18 dbfs is not enough signal. Sometimes I'll use the same soft synth for more than one part but I want to separate the parts for the final mix. A lot of soft synths like Z3TA will add subsonic rumble even when when the part is living above 400 Hz, without adding an EQ on the audio or instrument track you can't clean that stuff up pre-bounce and if you put the EQ on the synth, you may decide you want the rumble after all, then you have to bounce it again without the EQ.
For example: I usually do between 4 and 20 percussion parts each with their own MIDI track sometimes using multiple percussion synths, if I mixed them in the soft synth I'd have to deal with tiny faders with limited resolution and I still want to add processing. On top of that, I'd have to deal with overs due to summing (kick, snare, clap, hat, blip thingy). Even if I end up gluing them on the sub, I still want to be able tweak the parts individually.
Once I get the drums/percussion and bass tracks sketched out, I'll start bouncing and mixing. Then it's easy to start filling in the blanks.
Once I'm around half way through the mixing part my MIDI tracks are muted or archived and I'm dealing almost exclusively with audio.
YMMV, pick whatever workflow works for you and don't worry about it.

AT
What I do is freeze all midi tracks, copy and paste them to a new (audio) track and name them, unfreeze the original and archive it and drag it to the bottom of the screen.  Then I work from an all audio version.  That way it is simple to get the midi back and I'm working all with audio.  That is how I learned to work - on tape.  And early digital, where the computers couldn't handle more than one soft synth or reverb.  But it just seems neater to work in all audio.
 
But that is the great thing about DAWs and home studios, you can develop your own best practices.  If you prefer midi and it doesn't choke your computer, there is no reason not to work that way.  Whatever makes you feel comfortable.  And yes, you just export the entire song and SONAR converts the midi to audio w/ no action necessary on your part.
 
@

That's my method as well. I keep the softsynths frozen unless I need them and work only with the copied and renamed audio tracks in the mix. Additionally I use shortcut key H frequently as I proceed so I can hide any tracks I'm not using. So the frozen ones are out of the track view unless I need to back to the source midi or softsynth to adjust anything, which is easy to do. I prefer to only look at the active tracks for the mix rather than distractions like frozen tracks.

I convert all MIDI to audio so I can EQ the audio tracks where needed.

Skyline_UK
I convert all MIDI to audio so I can EQ the audio tracks where needed.

You can just as easily EQ any soft synth without bouncing/freezing. You don't need the audio to work with.

I usually need to bounce/freeze softsynths to audio due to lack of horsepower.
I try to keep the drums as MIDI, as I tend to edit them to the last minute.
 
Skyline - EQing the softsynths doesn't require bouncing/freezing. Do I miss your point?