Analoge 2Buss Compression and VSL

[jammusique]

Who amoug all of you are using an analoge 2 buss compressor on the main stereo out. And what units or type of compressors do you have good or bad things to say about?

[Dom]

I often use a Drawmer 1969 (soft-knee tube compressor) , with slow attack and fast release, to take off about 2-3 dB max, then into a Manley Massive Passive Eq to add some air. However more recently, I tried the stock Digidesign Compressor at a very low ratio, then into Sonnox Eq, and I may prefer this plugin combo to the analogue one, even if simply for the recall.

[jammusique]

Coincendentally, I am considering a 1968. How would you describle what the Drawmer does to the sound?

[Dom]

Apart from the obvious bit of gain reduction, which gives a bit of glue, the 1969/1968 muddies up the sound a bit as well, which can be a good thing sometimes, but often I prefer the clarity of the plug-in compressor.

The 1969 works better on a pop/dance/rock track, or a filmscore type track with more constant dynamics, eg with some drums or constant percussion. This will allow you to push the output stage constantly into the red a touch, then the unit adds some nice saturation.

I have considered the Safesound Audio Toolbox compressor, which allows very low ratios (most other compressors have a minimum ratio of 1:1.5) so you can set a low threshold and get a bit of 'glue" during the quiet bits as well, and without the whole thing getting overcompressed.

[jammusique]

Thanks for your info.

I do some cinematic, and some jazzy pop, and am thinking that I would like to have a 1968 and also a more tranparent buss compressor (SSL type or Mastering type).

[Dietz]

My most recent favorite:

-> [URL]http://www.virtual-music.at/webseiten_d/guelph/guelph201.htm[/URL] (... sorry, there seems to be no English page yet)

The perfect combination of SSL-like tightness with API-like features (sidechain filters!), switchable (!) transformer-inputs for some Focusrite-like harmonics and dry bass, plus built-in parallel compression.

Handbuilt on order! 8-)

[Dom]

@Dietz said:

http://www.virtual-music.at/webseiten_d/guelph/guelph201.htm

This looks certainly like a better alternative to the SSL compressor, with the added features, and even costs slightly less. Minimum ratio of 2:1 is still a bit high, but perhaps with a parallel blend of 50-50 this could be great.

[jammusique]

Hey that Guelph does look interesting. I didn't see a price there on that page?

Yes SSL's not having a high pass filter can sometimes muddy up the bottom end.

Here is what might be a simular product:  http://www.mercenary.com/fcs-p3s.html

And here is what the maker has to say about it:

When I started this design, back in 1994, I was hoping to get the character of the dbx 160 types of compressors but with soft knee response and without the noise and artifacts of those comps, especially the artificial sounding "grind" that they have during compression recovery that is a problem when using them for electric bass processing. I succeeded.

With higher ratios, you can squeeze the crap out of a signal and still retain it's sonic character and the stereo image. 

An engineer supplied us with waveform pics from a Pro Tools session of several compressors, the dbx 165, a Manley unit and our design... It's compression "signature" is very close to the dbx and was preferred by this particular studio. So much so that they put 12 of them in their remote truck in a custom case to save space. They will all be used on live vocal feeds to the truck.

In modes that use the peak detector like AR and especially in feedback mode the distortion gets more prominent as compared to RMS mode which is extremely clean. You can get a very nice smashed/crunch if you want in these modes... Good for adding that "smack" to a drum bus. I personally really like using AR/feedback on drum tracks/buses.

Using NLC mode, exploding/ dramatic drum sounds are easy to implement. Which brings up another comparison... The Valley 610 compressor. On the P3S, using NLC mode and extreme amounts of compression, around 30-50 dB of gain reduction, you can actually erase drum tracks leaving only ambience intact. The 610 is likely the only other unit that can do this.

In vocal recording, especially live work, you can adjust NLC mode to only act on very high amplitude peaks with very high amounts of GR and then "get out of the way" seamlessly.

On electric bass, when I used to use dbx compressors, I didn't like the distortion while the compressor was recovering from deep GR... This isn't the kind of distortion that is sought after, but a grinding artificial sounding artifact that takes the clarity away from a bass line.

The FCS compressors don't have that artifact, which is good since I like tube amp distortion whan I want distorted bass lines. I also like to use a slightly slower attack on bass to preserve the attack tones/noises that RMS mode suppresses.

Then, there is "Glue"...

In a stereo mix, at low-medium ratios, details are lifted out of the mix enhancing density of the program material... Subtle string overtones and vocal nuances come to life that were buried in the mix before.

This "Glue" will be obvious in any mode that the P3S offers, for stereo program use RMS will likely be the winner here. Btw, you can use the P3S as a mono compressor by just driving a single input with no compromise whatsoever.

The types of music I use the P3S on in program compression modes are Rock, Jazz, Fusion, Hip Hop, R&B, Prog Rock, Electronic, Acoustic Folk, World/Ethnic, Classical (including modern experimental classical)

The program material always sounds richer, better with the P3S engaged.Especially with the output transformers! The same is true of submixes and drum buses.

Best regards, Roger

[Dietz]

@Another User said:

Yes SSL's not having a high pass filter can sometimes muddy up the bottom end. [...]

That's a common misunderstanding. Cutting the sidechain frequency-wise means that the compressor will react _less_ to the filtered frequency range. IOW: Using the sidechan's HP will cause more bass to pass through, and thus less overall pumping.

[jammusique]

Yes I got that sidechain thing backwards. Without it, of course it will set of the compressor often more. 

That's alway an issue with anything with drums (kick) and bass (which is a large part of the overall music production). Orchestral music probably has other needs.

So Dietz you've tried and liked the Guelph?

I know that mixers often like to use parallel compression on drum busses, but I wonder it there's any use for such in other applications?

[Dietz]

For pure, classcial orchestral bus compression there might be other options, too (... ideally, we all would have a Fairchild 670, or at least one of its recent decsendants 8-) ....), but in general, I'm _very_ fond of the Guelph 201.

The main difference between typical pop- and typical orchestral bus compression is that you don't want the same rigid L/R-link of the compression for an orchestra. The Guelph's "Dual"-Mode takes care for that.

Provided I have some final limiter on the bus after the compressor, I use parallel compression on the main bus quite a bit, actually.

[jammusique]

Very interesting, thanks for the info...

[jammusique]

Hi Dietz, I've been talking to a few engineers about this stereo/dual mono question. Do you personally use dual mono systematically for orchestra (live and/or VSL) , or only on certain occassions? What might be some of the pros and cons for both?

[Dietz]

Uh-oh --- that's a severe misunderstanding. [:S]

By no means I meant to say that one should use two independent mono-compressors for orchestral mixes! I was talking about the way a stereo compressor "looks" at its incoming sidechain signal. Most of the time, the control signals are summed to a mono-sidechain from both left and right, thus making the compressor reducing the gain on the one side, even if the level peak only appeared in the other channel. This is perfectly fine for rock/pop most of the time.

The more the left and the right channel carry independant information, the more it is important to differentiate between the RMS and the peaks of the sidechain signal, though. RMS (i.e. average loudness) should be taken into account to the same amount for both channels, while indivdual peaks are only compressed on the side they appear. This is what some hi-end bus compressors do: The Maselec MLA-2 calls this "ImageLink", for example; the API 2500 has free control between the strength of this linkage, and its called "Dual" on the Guelph. There a most certainly quite a few other examples, too.

This description might oversimplify the matter, but you should get the idea.

... things get even more interesting when doing M/S rather than L/R-processing, but that's already a different topic.

Hope this makes the whole story a bit clearer.  :-) Sorry for any confusion.

[Dom]

Interesting Dietz. So the "Dual" mode on the Guelph will not just unlink the sidechain signal of the two channels completely, but will keep RMS compression linked and still leave peak reduction unlinked?

[Dietz]

That's how I understood it, yes. :-) If I'm not mistaken, that's where the "Dual" comes from - two detection units. But I'm only a user of this unit, not the designer; I could be completely wrong. I just can tell you that it sounds great.

[Dietz]

This kept bothering me, so I talked to Alexander Guelfenburg yesterday (the guy who built the Guelph 201). Some clarification is due now :-)

There _are_ two dector units in DUAL-mode, but not for RMS/peak differentiation (... I mixed that up with other devices). While under normal circumstances the compressor's sidechain is made up of the mono-sum of both channels (logical AND), the louder signal is used in DUAL-mode (logical OR). This is more or less how the human ear workd perception-wise.

Clear as mud! 8-)