I've used compression on and off over a period of many years, and one thing that really stands out is the invention and refinement of the technique of parallel compression--which this article does not mention. Whether it is pop or classical, sound effects or dialogue, I find parallel compression superior, particularly in combination with multiband separation.

@majoraxis Thanks for sharing, will try this one out.

Also agree with DrDave re parallel comp; best way to kill the life in dialogue is to hammer the transients.

@DrDave Any sound examples and tutorial?

@ DrDave - Wikipedia agrees with both of us on what Parallel Compression is:

"Parallel compression, a form of upward compression, is achieved by mixing an unprocessed 'dry', or lightly compressed signal with a heavily compressed version of the same signal."

http://en.wikipedia.org/wiki/Parallel_compression

Using the Dry Mix knob on this plugin can lead to better sounding "compressed" tracks, regardless of whether we agree on what to officially call the technique. ; )

Fabfilter has excellent, very in-depth tutorials, too.

@majoraxis please feel free to use wikipedia for a source for audio information. After all, what could be better?

For those less inclined to trust wikipedia, parallel compression is when you run two tracks in parallel. Mixing is just mixing. If you are clever with aux busses, you can run two tracks in parallel, one through the main bus, one through the aux bus. It isn't the best way, it doesn't sound the best, and it doesn't work as well, but if you don't want to cut and paste one track next to the other one (easiest and best way) you can run it through a bus. You will have to make sure that your latencies tap out exactly if you use the buss, on two tracks you can line them up very easily and set the delay and so on by copying the track settings, plus the latency going into the effects processor will be identical, along with the pan and any automation or track effects.

NB: You want the two tracks to phase out exactly if one is inverted (before you add the compressor/effects on one track), that's the key. Then you have zero latency issues.

Copy the tracks, with all the settings, apply the multiband splitter, use a limiter (recommend waves L2 but Samplitude has a very good one built in) and you are good to go in a few seconds, and can tweak all the settings perfectly on the second track, The splitter isolates the band you want to work on.

The L2 uses the Gerzon psychoacoustic shaping for a natural sound.

If you use the in/out bus of the cheesy plugin and think it sounds the same, by all means do it that way! If you haven't tried it the other way, take a minute to try it and see the endless possibilities.

And for those of us more inclined to Instagram presets: http://www.propellerheads.se/products/reason/index.cfm?fuseaction=get_article&article=devices_audiomatic

I just love the SSL mixer clone in reason, parallel compression works great with it!

VMS Multiband compressor: just love the sound and control.

@majoraxis, that may be, my recollection is different. It could very well be that the definition has changed since I was a student. Stuff changes all the time.

Sound on Sound: I don't regard SOS as a reputable source, it is a popular magazine, and you are free to use it. I think the idea that SOS sets itself up as some sort of authority on digital audio is good business, but their articles are often full of all kinds of garbage. I still enjoy the magazine. One of the most hilarious things I have ever seen is their analysis of "correct" microphone positions. Magazine writers have long, extensive resumes that consist of articles they have written for the magazines. They aren't teaching at Detmold.

The guy who developed sidechain compression from the original 1977 article is Hulse. Strangely, his name is omitted from the Wikipedia and SOS articles. I'm sure it was a complete accident that they left his name out. Get it straight from him, or use any of the several versions floating around. http://homepages.paradise.net.nz/rhulse/Side%20Chain/sidechain.htm

If you have technical training in audio, here is the math: http://homepages.paradise.net.nz/rhulse/Side%20Chain/sidecmat.htm

Note that there are other engineers who dispute the math, and even dispute that the effect is real. That's show biz.

@thepalais excellent point .Too much multiband can tweak out the sound. You only split the bands if you need say, more bass, for example. It is way, way more transparent because you compress only one portion of one of the two tracks. Big difference. If you are just compressing the one track, you don't need the splitter, but in practice most material will benefit from this.

Phase issue with the multi splitter: if you are using say 20 percent of the compressed track with the 100 percent of the noncompressed track, the splitter will not have a significant effect on the overall phase. There is a threshold for these phase effects, so factor that in.

Anyone who hasn't tried the Gerzon shaping with parallel compression, I recommend trying it. Old but good. There is no point in using a cheap plugin for this particular effect.

Just have fun, try it every way people suggest, or invent your own.