stxx
What are you using as interface??  If you are trying to monitor through sonar and not the interface's mixer, latency woill likely be an issue.   Unless youre using thunderbolt.

Thunderbolt can reduce latency by maybe a millisecond or two, but that's about it. The better USB2/3 interfaces can hit sub-5ms round-trip latency easily so long as the PC isn't working too hard and Thunderbolt might reduce that to 4ms. In real world terms round trip latency under 10ms (the equivalent of your guitar amp being 10 feet from you) is usually good enough.

There's always going to be latency where analogue->digital or digital->analogue conversion is concerned because the chips need time to do their stuff. Things get a bit more complicated when MIDI controlled hardware is involved because the hardware needs time to respond to the MIDI, and if it's a digital synth then that has its own convertors and the latency they introduce as well.

More latency can come from a synth having a greater than zero minimum attack time on the amplitude envelope to avoid loud clicks as the volume shifts from zero to more than zero in less time than a cycle of the oscillators, though that affects software synths as much as it does hardware. It's the same sort of issue you get when shortening an audio clip without making sure the resulting clip starts at a sample level of zero.

None of which is insurmountable, but Thunderbolt interfaces offer at best only a tiny improvement in latency, which may well not be noticeable.

I still have a Roland A-33 midi controller (full keyboard with no audio output, just midi output).  In a lot of ways that piece is really handy.  It ends problems that can pop up with both a VST and a keyboard voicing midi.

If there's a voice in my "real" keyboard I want to use, I can always play the midi file back through that voice and capture it in an audio track.  It's also interesting to record multiple audio tracks with different patches from a single midi track and hear how each voice sounds in a mix.  I've put keyboard lines into a "Vocal Ooh's" voice with outstanding results.
 

The problem that brought me here was a latency issue.  That was resolved by inserting a VST into my project.  Now I've started on learning how to sue the VST.
 
I'm running into the same thing that has frustrated me every step of trying to use Sonar.  The first thing I found was from a Cakewalk site.  "Step 1:  In your Browser, click the Synth tab"  I import a Midi file - no Browser.  I open a .wrk file - No browser.  I open a .cwp file - No Browser.  I create a new project - there's the browser, but there's no Synth tab.
 
So I select a track, use Insert->Synth from the menu.  That works, but the VST does nothing to the track.  I've tried a bass synth, a piano synth, and a synthesizer synth.  I have discovered how to add a VST track that apparently uses whatever patch is assigned to it from the VST, but I've not found how I can change the patch in the track.  I can't even find a patch selection when I use the TTS-1 GM patch.
 
I've spent about two hours on trying to figure out how to use VST's and don't have much to show for it.  I found a tutorial about recording a Sonar song start to finish.  I'm 15 minutes into part 2.  So far the "instructor:" has hit about a dozen computer keyboard keys to display or hide the things he wants to look at and he's managed to have a really bad drum line that has an incredibly mechanical sound.  And that video is more informative than the ones I find on Cakewalk pages.
 
Trying to learn Sonar has made me feel like I'm just an idiot.  the stuff from Cakewalk is so far over my head I can't even figure out what questions to ask.  Everything I'm finding seems to be aimed at studio engineers.  When I have to Google two words in every sentence I read, I tend to stop reading.

For the OP, there was a 9 hour tutorial set done by Karl Rose (fastbikerboy in the forums here) during the SONAR X2 days that is still 90% applicable to SONAR as it exists today. That was made free to everyone and posted to YouTube a few years ago. Karl does a stellar job of walking through options as he goes.
 
The YouTube set does not have a Chapter listing, but that can be found in this post on Steam (use this as your guide for the video set). The actual video set (called "SWA Complete SONAR X2") can be found here (50 Chapters total).
 
For you, Chapters 19 and 20 (MIDI Synths) are good ones to look at. Karl works at inserting TTS-1 in the beginning of Chapter 19 and goes from there.
 
Please do not feel discouraged, and do not hesitate to post here. We have ALL suffered the learning curve. Sometimes it is as simple as finding the proper term to research, other times is more of a walk-through required; but this community is one of the best there is, so please ask as needed.
 

Garageband is a toy compared to SONAR or any competitive DAW.  If you want demo quality at best for your music, then maybe GB is good enough.  If you want to make even half serious recordings, GB is still in elementry school while SONAR (et al) is getting a masters and the learning curve is potentially formidable if you lack patience

Here's a little tip that might get you going.
 
If you go to preferances/ midi / devices. uncheck all output options. 
Now when you open a GM midi file it will automaticly open with the TTS-1 and all the tracks assigned to the proper instruments and patches. 
 
OK hopefully you have found the browser to the right. It's just minimized in some views. don't forget, Sonar is a Windows based program so a lot of views are minimized , resized, floated etc. spend some time exploring the edges where you'll find all sorts of hidden views for workspaces. All software works this way.
Print off thekeyboard  shortcut page,
https://www.cakewalk.com/Documentation?product=SONAR%20X3&language=3&help=Menus1.181.html
 
 
 
 

Merlin I highly recommend a separate recording interface. You can get them for 200.00. I can give recommendations. Usb as the main connect into computer. If you record one or two tracks at a time you only need one or two inputs. 
 This makes it all much easier since the onboard computer audio isn't the best for recording with regard to latency.
 
I also use a hardware synth. I send midi to Sonar through either the usb out on my keyboard or through my midi equipped interface. I can also send audio from my hardware synth into my computer recording interface. In this way you have two sets of tracks both midi to play the soft synths and audio from the keyboard.
 
So a hookup with you using an outboard interface would look like this:
midi from keyboard>midi into interface or directly into computer.midi can be either a 5 pin connection or a usb from keyboard.
Audio outs from keyboard or line outs> into interface line in>then goes internally to sonar>select interface inputs as seen on track inputs for the track you intend to record in Sonar .Set a separate midi track to track midi if you want the audio and the midi versions. Once you capture the midi you can tie it to any soft synth you have. Simply drag it onto a soft synth track or send the output to that synth input.
 
The way some get around latency issues with midi and slow built in sound drivers is they use audio out from synth for tracking, maybe use headphone jack directly on keyboard while at the same time tracking midi. You don't need to hear the soft synth for tracking. It should play in time after you have tracked the midi makes sense? :)
 
In some cases the audio from the keyboard might sound better than the TTS-1. The only real reason you need midi is for editing purposes. If you're a good keyboard player, maybe you don't need correction :)
 
Having an interface will eliminate the need for a mixer, although I keep mine around and have it tied to two of the inputs on my interface. This way I can feed anything into the mixer on a whim, guitars, shakers, etc, 
 
Hope this helps some. 
 
Cancer survivor too. Congrats!