Earwax

mudgel
It's not hard to do, that is to record your VSTi in real time, you just have to be prepared to deal with the issues that it brings up. Send your monitor source out of your audio device and loop it back onto another track making sure that that track is out of your monitor path. Voila, you have a live recoding of your performance. If you are going to add further material you'll have to nudge the recorded VSTi audio tracks.

No one is saying what you describe hasn't been tried. Those who like to record "live" have been doing it (or some variant) for years. A substantially more direct method is what is being suggested. Some get the concept, and some don't. Some prefer the methodology, and some don't. I'm okay with that. People work differently. But to suggest (and I'm not saying you are), that one method is superior to the other, or that people who prefer one method over the other are somehow brain-damaged, is ludicrous. 

I merely suggested a way to do what you want. I certainly didn't equate the desire for your way of wanting to do it with anything of a derogatory nature neither do I say one way was superior. At this time there just isn't a way except for fairly unacceptable work arounds. I get what you want to do. I've been doing this (recording)one way or another for over 45 years.
Remember if your sound source is a VSTi then the MIDI you play either live or recorded can't possibly play different audio.
So while the guitarist is playing audio you're not. You're playing a MIDI performance. Now if that MIDI performance can create the subtleties and nuances of your live playing of it then the recording of it can't be any different. The only way you might be able to do that is to record the audio itself, and seeing it's MIDI triggered in what way could it be different whether recorded live or bounced or frozen.

This topic pops up now and again and always runs the same gambit. My personal experience with (wanting) this has been doing screen captures and wanting to record both a vox track and a synth's output, which required physical loopback cabling to achieve. I found out after the fact that my Saffire has a built-in loopback feature, but is still the same deal. Right now it "is what it is" and only a feature request can change it. I agree that it would be nice to have internal to SONAR rather than rely on an external means (internally done would allow for better safeguards on feedback loops IMO).
 

mudgel
Remember if your sound source is a VSTi then the MIDI you play either live or recorded can't possibly play different audio. So while the guitarist is playing audio you're not. You're playing a MIDI performance. Now if that MIDI performance can create the subtleties and nuances of your live playing of it then the recording of it can't be any different. The only way you might be able to do that is to record the audio itself, and seeing it's MIDI triggered in what way could it be different whether recorded live or bounced or frozen.

A quick point to the above... I mentioned this to Noel a year ago when I was in Boston... Z3TA+2 is a good example where oscillators can be run "free"... so each performance (even identically done), can/will yield a different audio output... this is a situation where freezing a synth or capturing a "MIDI performance" falls short. Being able to perform takes of a synth output the same as any other audio take would prevent losing a potential take that is desired for the final result.
 
[I saw the comments regarding "randomness" above... and to me, the "free oscillator" situation is what first came to mind.]

This has been an interesting conversation.  My keyboard skills are poor enough that I need to record midi, so I can fix stuff later on, so I would never use this direct recording capability.  That said, it appears it would be a useful feature to some and would also differentiate Sonar in the marketplace.  Hopefully the feature request will get enough votes to encourage the bakers to implement it.

I'm in the fortunate position of having an RME device in which the TotalMix software makes it easy to route audio out and in again without having to be processed by any ADDA.

I'm following this discussion with interest.  I probably don't completely grasp what the difficulties are, but I'm wondering if Silk Tone's Sidechain Mixer plugin might not be useful here ?  Could it let you get around "But as soon as the synth audio is selected as the input, the record enable for the track disappears!" (timg11) ?
 
 
 

timg11

The Grumbleweed
bounce out the synth to your desktop (or wherever) and drag it back in Sonar.

I'm not sure what you mean by "bounce to desktop", but I played around with the bounce to tracks method more and finally got it to work. I discovered these facts:
 
1) You have to uncheck Fast Bounce (checked by default), and check Audible Bounce.  This is not mentioned in the 9-step procedure in the help file.
2) Presets don't fully "preset". They save some settings but not others. Notably, the Destination track and the Source Buses/Tracks are not part of the preset, and your settings are not remembered, even between sequential uses in the same project. They have to be manually set every time.
 
I still don't understand why the direct recording of the synth is not allowed.
 

it seems like the point of this thread got lost between talking about jimi hendrix and playing live. please clarify: you are basically looking for a way to "print" your VST instruments to "tape", to you use the old terminology. once you are happy with your midi parts, you can print it and then delete the midi and remove the synth to free up resources, correct?
 
that's what i used to do in DP. i'd have 6 stereo tracks with their input set to Reason instruments, and then arm and record them. the audio from the 6 reason instruments would then bounce/mix-down/print to their 6 stereo tracks and then i wouldn't have to have Reason open and i could move onto the next phase of recording (vox/guitar/etc).
 
unclicking 'fast bounce' is the easiest way to do it, but it also prevents you from doing anything during the bounce other than listen to the music. in DP, this action wasn't considered a bounce; you were just recording the output of your instrument to a track, thus you could also record vocals or tamborine or whatever while the other tracks were recording output from your synth.

mudgel

Earwax

mudgel
It's not hard to do, that is to record your VSTi in real time, you just have to be prepared to deal with the issues that it brings up. Send your monitor source out of your audio device and loop it back onto another track making sure that that track is out of your monitor path. Voila, you have a live recoding of your performance. If you are going to add further material you'll have to nudge the recorded VSTi audio tracks.

No one is saying what you describe hasn't been tried. Those who like to record "live" have been doing it (or some variant) for years. A substantially more direct method is what is being suggested. Some get the concept, and some don't. Some prefer the methodology, and some don't. I'm okay with that. People work differently. But to suggest (and I'm not saying you are), that one method is superior to the other, or that people who prefer one method over the other are somehow brain-damaged, is ludicrous. 

I merely suggested a way to do what you want. I certainly didn't equate the desire for your way of wanting to do it with anything of a derogatory nature neither do I say one way was superior. At this time there just isn't a way except for fairly unacceptable work arounds. I get what you want to do. I've been doing this (recording)one way or another for over 45 years.
Remember if your sound source is a VSTi then the MIDI you play either live or recorded can't possibly play different audio.
So while the guitarist is playing audio you're not. You're playing a MIDI performance. Now if that MIDI performance can create the subtleties and nuances of your live playing of it then the recording of it can't be any different. The only way you might be able to do that is to record the audio itself, and seeing it's MIDI triggered in what way could it be different whether recorded live or bounced or frozen.

Mudgel,
I know you weren't suggesting one way was superior to the other. Nor did I interpret your post as being derogatory in any way. That's why I placed the comment in that post
"...and I'm not saying you are..". I appreciated your input. I will respectfully disagree with the concept of the MIDI data not playing back different audio. I've been doing this for decades myself. And trust me, it has happened. Not only with the sound (mettelus just above and others in previous posts have cited examples), but even the gestures as part of the performance. While recording a duo performance with, say, a guitarist, try improvising in a piece with different time signatures and tempi while recording the MIDI data along with the guitarist's audio data. Then play it all back.
 

Earwax
Because the guitarists in the bands cited in my examples are not recording gestures. They are recording sounds. They are recording sounds at the same time as the keyboardist and drummer are recording their sounds and some (depending on the VSTi and what performance tweaks they are using) of their gestures.  The live performance breaks down even in duo or trio recording without the ability to record VST/VSTi directly into Sonar. 

But what the guitarist hears while recording into SONAR is what will be heard on playback. SONAR is recording the sound of the guitar. Monitoring the sound playing through a VST in real time will produce the same end result as playing the track with the recorded guitar sound back through the VST.
 
Again, this doesn't obviate what you want, but I don't think what you want is needed in this particular instance. You'll hear the same thing during real time monitoring as will on track playback.

gunboatdiplomacy

timg11

The Grumbleweed
bounce out the synth to your desktop (or wherever) and drag it back in Sonar.

I'm not sure what you mean by "bounce to desktop", but I played around with the bounce to tracks method more and finally got it to work. I discovered these facts:
 
1) You have to uncheck Fast Bounce (checked by default), and check Audible Bounce.  This is not mentioned in the 9-step procedure in the help file.
2) Presets don't fully "preset". They save some settings but not others. Notably, the Destination track and the Source Buses/Tracks are not part of the preset, and your settings are not remembered, even between sequential uses in the same project. They have to be manually set every time.
 
I still don't understand why the direct recording of the synth is not allowed.
 

... in DP, this action wasn't considered a bounce; you were just recording the output of your instrument to a track, thus you could also record vocals or tamborine or whatever while the other tracks were recording output from your synth.

Therein lies the problem. You can't do this in Sonar.

I believe Scook has posted a work around for this many times.
IIRC it went something like this...
 
Place a note past the end of the song on the MIDI or instrument track
Enable "Echo On" on the MIDI or instrument track
Select the track(s) associated with the synth
Bounce the track(s) with "Fast Bounce" disabled, "Audible Bounce" and "Live Input" enabled
Play synth while the track(s) are being bounced.
 
This should allow you to "record" the VSTi while playing live.