I'm working on some MIDI files to make some simple backing tracks and I seem to be having a lot of difficulty with inserting multiple soft synths and switching between them, and also I am puzzled about exporting to audio. 
 
Here's an example of what happens to me:

1. Open MIDI file
2. Insert TTS-1 and assign each MIDI track's output to TTS-1
3. Hit play - all seems normal.
4. Insert another soft synth, say Dimension, and assign, say, the bass track to it.
5. Open the Dimension interface and pick a bass sound file.
6. Hit play - no sound from Dimension. 

 
If I start a simple project from scratch, add a MIDI track, insert Dimension, pick a patch, and play some notes on my keyboard, it seems to work as I'd expect - sort of. Except sometimes it seems to not make a sound, but if I record some MIDI data and hit play, it comes back.
 
I seem to have similar problems with some of the other soft synths like Session Drummer. I might insert Session Drummer and Addictive Drums, and alternatively assign the drum track to one or the other. Typically one of the choices gives me no sound.
 
If I just stick with TTS-1 I seem to have no troubles at all, but -- sometimes that's acceptable, other times it's not.
 
Often, I want to mute all the parts except drums, or drums and bass, and then export it to an audio file. The docs suggest that I can select the track and choose Track -> Bounce to Track, but I just get told that the track contains no audio data. I can freeze the synth and then it mixes the audio down, but if I do this with the TTS-1 it has to create audio tracks for all the parts, which takes an inordinate amount of town when I want to export just one or two tracks. If I was able to use, say, a bass from Dimension and drums from Session Drummer, I could freeze those and SONAR just has to render that data - but I cannot get that to work as described above.
 
Are these things that everyone struggles with? Am I missing something really obvious that I just can't see?
 
The machine is a pretty beefy one -- Intel i5-3450 CPU @ 3.10 GHz, with 16 GB RAM, running Windows 7 Home Premium 64-bit.
 
Thanks in advance.

Hello which version of Sonar X please? X3C? (Maybe worth updating your footer like mine so we know what you have).

Thanks

Thanks, Alex, good idea.
 
X3C Producer (64 bit), Win 7 Home Premium (64 bit), Echo Layla 3G (PCI), custom-built PC with Intel Core i5 CPU 3.10 Ghz, 16 GB RAM, Seagate ST1000DM003 drive (1 TB), Western Digital WD5000 drive (500 GB), Intel HD graphics

Hi AlChuck, when you assign the bass from TTS-1 to Dimension do you also reassign the channel and patch numbers? The new channel and patch numbers must match Dimension’s channel and patch numbers.
 
The same thing with your drum choices.
 
When you bounce to track you must select the MIDI track and the softsynth (TSS-1, Dimension, etc) track. Together they will produce a new audio track.
 
If you have already done this then sorry for the redundancy.
　

TTS is a multi-timbral instrument - Dimension and Rapture ain't.  Open DimPro from the browser - that will insert midi/audio channel.  Copy the bass notes and put them on the DimPro midi lane.  Pick a bass in DimPro and hit spacebar.  It should play.  I gave up trying to assign patches to DimPro what with a 1000 patches.

MarioD,
 
Thanks, I missed somehow that you had to select the MIDI and softsynth tracks to do a bounce.
However, I have a track with drums playing through Addictive drums and everything else through the TTS-1. When I select the four tracks corresponding to the drums and bass parts and Bounce to Track, all I get is the drums. If I leave its UI open, I notice that the TTS-1's output drops to zero when the bounce is done.

As AT already pointed out, you are missing the procedure for inserting the software synths. It is NOT an effect. You should use top menu "Insert". It will create MIDI and AUDIO tracks. Copy your MIDI data in to that midi track. You should have audio track for each MIDI track. I did not try to freeze or bounce any songs in polyphonic mode. I open individual instances of a plugins every time. But it can be done. Just make sure you assign your instrument audio track to Master, and have corresponding audio tracks for every MIDI.  

vladsyn,
That's how I do it, except that there's no need to copy the MIDI data from an existing track into a new MIDI track - you just need to assign a MIDI track's output to the soft synth's input. I do have a separate audio track that corresponds to each MIDI track, and each synth's output is routed to Master.

Alchuck,
 
again, you have to have the synth (DimPro) set up correctly.  DimPro will only recieve on the base channel it is set up to.  You can get 4 midi channels if you set DP to mult-timbral, but that will only play one element for each midi channel, 1-4 (or whatever you set the midi channel to in DP.  Also, some DimPro sounds are restricted in range and won't play a note because it is outside of its range.
 
The TTS is a general midi module (GM) and can use all 16 midi channels at a time.  You can play a whole performance/song through it, assigning different channels to different instruments.  It goes back to early sound card days when the internal sound cards were set up to play music (badly) using short 8-bit samples stored on the card.  GM was a means to organize the 3rd party music files so they would play on any computer card.  Channel 10 is always drums, where the different notes "play" the differnet drum sounds.  I think channel 1 is pianos, tho it has been a long time since I've used GM.  Also, GM sends midi instructions to the synth for patch information.  Back in the aughts (and before) there were only 128 instruments on the sound card.  Later, they figured out ways to send more by switching banks, like hardware synths - see LBS/MBS.  Few modern synths follow this configuration.  So your GM file will likely tell DP (which has a 1000 patches or so) to play its tuba instead of the piano the file wants to.
 
Join the 2nd decade of the 2000s and load in DimPro (it will open w/ both audio and midi tracks assigned, tho you can use an easier combo mode which only shows one track) and play or copy the midi to DP which is opened to the patch you want to play.  It should work like a charm then, for the one instrument you want to play.  You can open as many instances of DimPro as you need, on their own midi channels, playing their own instruments and midi lines.  You'll find that the sound quality is higher than any GM synth, including TTS.  If you need drums, use a drum synth set to Midi Channel 10, which is still commonly used, a left over from the GM days.  Or you could use SFZ, a free synth from Cakewalk, that uses soundfonts and SFZ files to play GM style files.
 
@ 

AT, I know about General MIDI, I've been doing this stuff (albeit sporadically) since the early 90's, with an early sound card (the first with a onboard wavetable synth). I wasn't expecting it to be a multitimbral synth and play all the parts - I was just trying to send one track's MIDI stream to it to play the bass line. I will check my routing again, it's probably something I'm overlooking. But my problems are wider than getting Dimension Pro to play - there seems to be some oddities with almost any of the soft synths I use if I'm using more than one in the same project. Maybe my sound card's drivers are a little wacky (it's an Echo Audio Layla3G PCI interface with ASIO drivers - the latest ones from Echo's site). Not much I can do there unless I replace it with something else and see if the problems clear up, which is cost-prohibitive at the moment.