for just peak removal of a bass part, would just use a limiter and not compress the bass yet....so the mix doesn't get too dense allready.
I find the old VC64 (Vintage Channel VST) compressor the nicest, neutral and it has the basic controls you need, Nomad CP2S is nice too.
I only do NY (Parallel) compression and workflow is to set a low ratio 1:2, and fader all down....then dial in untill I come to point where the compression only makes it louder and no longer adds dynamics.
then play with ratio first, then fine-tune treshold and last I tweak attack/release
 
good luck.
 

Such helpful info.
Im also a newb here trying to strive for a better sound.
I'll definitely try out these tips you guys provide!

I like the idea of copying a track, flipping the phase and adding compression to the unflipped track to get an idea of what is really happening to the signal as you compress it.
 
I'm discovering lately that working with the attack and threshold is very important to the outcome of the effect, sometimes a very small adjustment can make the difference between hearing more detail and squashing a signal too much. I do rely on the visuals of a meter as well as my ears when making these adjustments. For a signal with momentary peaks like a piano with hard hitting tines or a guitar with an aggressive attack I like to just tickle the meter since my goal is to level only those peaks. I listen and watch those attacks on the meter in terms of -db until I get it the best I can. It's all a balance between the signal input level..or how much of the signal  you want to feed to the compressor and what you have determined to do to the signal. Determine ahead of time what your goals are. Don't let the compressor make that decision. Tell the compressor what you want it to do.
 
I have to ask this- Are you using the monitor button on your track? If not....you won't hear the processed sound through the PC and it won't sound like anything is happening.The processing will also produce a very small lag time. On a decent system it probably won't be very noticable, on a system with high latency you will get a delay. Some guitarists prefer a hardware compressor for this reason.
 
 

The reason why the null method woks well and is interesting is because when the music is present it is much harder to actually what any compressor is doing. The null method removes the music and only leaves the effect of the compressor behind.
 
If it snaps in when compression takes place you are attacking the signal too quickly and hence effecting the transients of your sounds.
 
If you read the thread I created on how to set up a compressor you will find I set the Attack first.  It is actually one of the most important parameters because it has so much influence on the leading edge of the sound.  Release is also an important parameter and usually comes second.  You can even do this with a high ratio and low threshold just so you can hear it more so.  Then once you get those two things set up you then go back and focus on ratio and threshold.
 
Also remember that higher ratios produce a smaller sound and lower ratios produce a much bigger sound and hence you can then turn it down.  (maximum illusion, minimum voltage)  Try and get the compression you want with the lowest ratio you can.  Even a ratio of 2:1 is a lot.  People tend to use too high a ratio and hence then wonder why their mixes (or tracks) sound small.
 
Editing parts in an editor before mixing will make a big difference too.  As softer parts can be raised a bit and loud bits pulled down so when the compressor is finally inserted over a track then it ends up doing a little work only with a low ratio and it will sound very transparent.
 
The free Tokyo Dawn Labs compressors are excellent too and can be set to very transparent.

https://soundcloud.com/aaudiomystiks/infatuation-or-love-sonata-no5
Ok there is a lot of disinformation here, stuff that will only confuse you.  There are two or three people who have responded, who are professionals...Jeff, Rain (even though Rain and I don't get on) they will give you sound advice.  Any one suggesting multi-band for a newb is silly, how can you set a multi-band if you have no clue about compression in the first place.  Boost 11 is designed for limiting, and won't particularly help squash the track in a good way.  Here is my advice and it is gleaned from many years of research, although I am not a profession.
 
First...do you need to compress?  Huh this sounds like a silly question but it is not!  What I want you to do is first go into preferences and in audio preferences change the pan law to minus 3.  This will change your mixes to the point you go wow.  I'm asking you to do this because if you are going to learn to mix the first thing you need to do is make Sonar work like an analogue desk.  Then place the console emulator on all the tracks and buss's.  Make sure you place the buss console emulator on the buss's and the track emulator on the tracks.  This can be used to compress your tracks ever so slightly, shape the tone and fatten bass, even without touching a compressor.  I hope you have producer.  If not, still change pan law.  Pan Law is important, it will give the stereo image a boost, by cutting 3db from the centre and create a wider stereo image. http://www.cakewalk.com/Documentation?product=SONAR%20X2&language=3&help=Playback.27.html
http://en.wikipedia.org/wiki/Pan_law
 
After this change the track meters to RMS-24, as Jeff will chime in these meters aren't entirely accurate and he will start talking about VU meters but if you know the RMS meters, then they can be used to predict wether you need to compress.
 
Ok now we can talk about compression, check the bass meter and if the meter is hovering around -18 RMS, then you probably don't need to compress or very little, if it is more than -18, don't reach for the fader, use the trim at the top of the track to get the bass to sit at around the above target.  Here is why I want you to think about Sonar differently.  Sonar and most DAW's are designed to work like an analog desk, huh?  All the rules about summing still apply, hence the reason I told you to change the pan law.  Using the fader only decreases the resolution in digital.  So by using the trim you keep the summing rules and the resolution rules.  If needed only change the fader a few DB and never up.
 
Now if you still think you need to compress, here is  couple of basic settings, although I am struggling to remember them off the top of my head.  I have a manual, given to me by professional's from audio school and the basic principles work.  For bass, use a ratio of 4-6 no more.  10 is limiting and yes you can limit bass with a 10 ratio, the 1176 which is a famous compressor used in the 70's was used this way.  But for a newb, it is too much.  The PC76 is an emulation of this plug and if you have the pro channel then perhaps learn how to use it. http://www.cakewalk.com/Documentation?product=SONAR X2&language=3&help=ProChannel.4.html#1200380
 
However if you don't and all you have is the sonitus one, then learn how to use it.  Here is another quick lesson, all compressors are different, the sonitus one is just a plain vanilla compressor.  I'm using H-Comp by Waves for mastering...very smooth and transparent, check it out and if it comes up for sale at 50 bucks...get it but it has different controls and reacts much differently to sonitus.  For starters it has no gain make-up, it's auto.  It has a punch control which acts like a second attack and it has 4 analogue modes.
 
Ok sorry I digressed, ratio 4-6, attack 10-50 ms, release 100-500...this is the setting I can't remember off the top of my head and it varies again with different compressors.  make-up depends on how much gain reduction, anywhere up to 6db.  These settings are pretty stock standard but will work with the sonitus.  Also set the knee to hard for bass.
 
My advice is first set your levels for all your tracks, get everything sitting at around that -18 RMS mark.  Do this using the trim, then if needed use the fader marginally to balance.  If you have the pro channel use the tape and console emulator's first to see if you can compress this way.  If you have the console emulator, even the tube module is excellent for basic bass compression.
 
Here is the latest mix I've been working on: https://soundcloud.com/aaudiomystiks/infatuation-or-love-sonata-no5 and even though it's a virtual orchestra, the principles still apply.  There is still a little mud on the bottom end, which I am going to fix right now.  EQ here is another misunderstood principle which I suspect you will have questions about.  Again if you have the prochannel the quadcurve is excellent and with a little know how it can make mixes really punch through.
 
Ben