Skip to main content

I was mastering a song the other day, and (as happens quite often) I found that the Drawmer 1969 in my rack didn't sound quite as good as a certain plugin compressor I ABed it against. But rather than just go with the plugin and enjoy a faster-than-realtime render, I decided to experiment a little...

I routed the audio down 3 separate tracks in Reaper, then used 3 instances of Pro-Q in linear phase mode to create a cross-over network. I then loaded 3 of my plugin compressors, and dialled my sound in as I wanted it: the multiband version actually worked quite a lot better than the full-range compression on this mix.

Then, one by one I bypassed the plugin compressors for each band, and dropped in Reainsert with my 1969 routing instead, dialled in my compression, then bounced the stem. Voila: a hybrid multiband Drawmer 1969, with "perfect" digital linear phase cross-over filtering.

Sounded great actually! Only problem is, I now want another two 1969s so I can set it up in realtime duh

Comments

RemyRAD Mon, 11/28/2011 - 17:46

Nothing like spectral processing done right. I'm sure that sounded sweet. I found it interesting that you used FIR Crossover filters instead of IIR. Or, did I get that backwards again? Are you backwards also? Probably not since you are in the UK and I am in the FUK (Former UK).

LMAO, DDS gas on Tuesday WTF. OMG LOL from CR0W (Control Room On Wheels). I feel abbreviated.
Mx. RAD

IIRs Tue, 11/29/2011 - 18:35

RemyRAD, post: 379724 wrote: I found it interesting that you used FIR Crossover filters instead of IIR.

I think they were FIR filters, but as I understand it linear phase is not necessarily FIR, and FIR is not necessarily linear phase.

Anyway, I think linear phase cross-over filters are really important in multi-band compressors: I find the phase anomalies around the cross-over points in other designs are often a bigger problem than whatever it is I'm trying to fix in the first place.

The only other solution I am aware of is to avoid recombining the bands at all, and use dynamic shelving filters to create the gain differences between bands instead. In this approach the cross-over network only feeds the compressor sidechains: the resulting gain modulation signals are then matrixed into gain modulation signals for a set of dynamic shelving filters (one for each cross-over point) plus an overall gain stage.

However I'm only aware of two compressor plugins that use this approach, and I designed both of them. Not sure why: I reckon everyone else is missing a trick!

If you think about it, conventional multi-band compressor designs are kind of like using a sledgehammer for cracking nuts. If you're feeding amps for a PA system then obviously you need a network of steep cross-over filters to split each band off to its own amps and speakers. But a multi-band compressor splits the signal into X totally independant bands, only to then change the gain of each band a tiny bit (relatively speaking) and put it all back together again!

Of course there are other reasons why you might want to go the traditional route, for example you may want to saturate one or more bands, or use a specific compressor like I did above.

RemyRAD Tue, 11/29/2011 - 20:30

Aren't most crossovers that folks use utilizing Butterworth, Butter Pecan, Butterscotch something filters? Now I'm hungry but I'm not into suites unless they have lots of knobs & dials & other crunchy things.

I'd really love to hear what you actually achieved with an AB comparison? I think most crossovers utilize shelving filters that are generally not phase linear. And I was always under the impression that FIR filters were always linear where I I R were certainly nonlinear phase? But you are better educated and more knowledgeable than I am in those particular applications. Plus you're another one of those incredible UK guys that always get this stuff right. So I don't necessarily doubt that you aren't correct. I am both intrigued and quite interested in what you've done.

Oh I wish I had been born in the UK... that's where I really belonged. Just like Geoffrey Emmerich, where in the UK, if you weren't a great student (I certainly wasn't) and they didn't think you were college material, they would ask you what you wanted to do and shove you out the door at 16 right into a job. In the United States, I had to take it upon myself to do the same thing. Especially at the end of my junior year (while working at the Community College of Baltimore radio station, since I was 15 & the biggest recording studios south of NYC in Baltimore) I was working out the schedule with my counselor for my senior year. It was then my counselor told me I would have to come back a second year as a senior since I flunked gym class 2 years in a row for not taking my shower the last period of the day. I thought my counselor was joking with me? He wasn't. I told him that was insane but that didn't matter. I told him I was the only kid in his high school that was professionally working in my field since I entered high school! But that didn't matter. So I dropped out halfway through my junior year. I promptly went down town and took my GED. I found two incorrect multiple-choice questions no one had ever questioned about to the teacher. The teacher was also puzzled by this as no one had ever mentioned this to her in the years that she had been giving this GED test. I passed it the first time through before school had even let out. And I discovered it was so stupidly simple, I could have passed that after the seventh grade! I thought to myself about all the time I wasted learning nothing from stupid teachers that I could have better spent in my career field. So yeah, I should have been British. That was even more highly verified to me when Sir George Martin offered me a maintenance position at AIR Studios in Montserrat. But I prove my stupidity and turned him down as I wasn't interested in a career of pure maintenance. I'm a music person. But he did tell me that all of his engineers pretty much started off in maintenance. He told me it was important for his engineers to understand everything about the equipment, inside & out, first. And at 25 years of age, I already felt that I had proved my knowledge of the equipment after having built numerous control rooms and consoles from scratch, becoming an authorized Ampex, 3M, MCI, service technician & Scully Quality Control Manager & Final Test Technician. What really sealed the deal off was he told me he knew what I was making as an NBC Union engineer and he couldn't pay me that on British terms. He told me if I ever changed my mind to contact him. I change my mind the following week, over 25 years ago. I haven't gotten around to calling him back. People keep offering me studio & broadcast maintenance work and I tell them if I didn't take Sir George Martin's job offer what makes you think I would take yours? Either way, I still do maintenance for some of my closest friends & colleagues. You just can't not do it. Unfortunately, he's retired and I'm getting ready to. After I get a chance to visit Nashville, if I don't believe there is reasonable opportunity for me and/or my truck there, I'll sell everything and retire. And that doesn't necessarily mean that I will stop visiting & contributing to all of my professional friends & musical recording enthusiasts here at Recording.org. Although I do believe plenty of my professional colleagues here wouldn't mind if I just disappeared? Sometimes I get too silly for any kind of real comprehension about what I'm blathering about. I know that. I really can't stop injecting humor otherwise I might be injecting something bad for me. No no, I would never stick anything in my veins because I am too vain. By injecting I merely meant consuming things orally or is that aurally? Both I think? And when I retire, I can turn up my control room monitors even louder. Bottom line, I think I'm getting bored?

Thinking things through, under & around
Mx. Remy Ann David

IIRs Wed, 11/30/2011 - 10:31

RemyRAD, post: 379796 wrote: Aren't most crossovers that folks use utilizing Butterworth, Butter Pecan, Butterscotch something filters? Now I'm hungry but I'm not into suites unless they have lots of knobs & dials & other crunchy things.

Butterworth filters have a perfectly flat pass-band, which is obviously desirable. Most speakers systems use Linkwitz-Riley filters however, which can be created by cascading two butterworth filters in series. Lots of good info here: http://www.rane.com/pdf/ranenotes/Linkwitz_Riley_Crossovers_Primer.pdf

L-R cross-overs have a flat frequency response, and help to cure lobing problems caused by non-coincident speaker drivers, but they do cause phase shifts.

RemyRAD, post: 379796 wrote: I'd really love to hear what you actually achieved with an AB comparison? I think most crossovers utilize shelving filters that are generally not phase linear.

I'll have to get permission to post some clips...

Anyway, I think linear-phase filters inherently introduce latency, so can't really be used for PA system controllers. But in the plug-in world, with automatic latency compensation now pretty standard, it is becoming standard practise to implement linear phase cross-overs (or at least to offer them as an option).

I always switch Pro-Q to linear phase mode for this type of situation because I know that the minimum phase versions aren't optimal... I'm guessing they are basic Butterworth types, but I'm not sure. Now that you mention it, I would be interested to try this setup again using proper L-R filters instead.

RemyRAD, post: 379796 wrote: And I was always under the impression that FIR filters were always linear where I I R were certainly nonlinear phase?

FIR filters can be anything you want: linear phase, minimum phase, maximum phase, whatever; so long as the filter kernel is long enough you can create any filter response. FIR can even do reverb: the only difference between an FIR filter and a convolution reverb is the length of the kernel / impulse. (This leads to another interesting insight: when dealing with a small space like, say a phone booth, or an acoustic guitar body, the dividing line between reverb and EQ starts to get rather blurred)

IIR filters are usually not linear phase, but I think it is possible to linearize the phase response using tricks like processing twice, once forwards, then once backwards: this will subsitute the warped phase response for an overall delay, ie: latency. My maths starts to let me down when getting into filter design however, so I may well be wrong... I shouldn't have skipped so many maths lessons when I was at school.

smoke

RemyRAD, post: 379796 wrote: Although I do believe plenty of my professional colleagues here wouldn't mind if I just disappeared?

Not me: you're the best thing about this forum thumb

RemyRAD Wed, 11/30/2011 - 11:01

Thanks IIRs, that was very kind of you and certainly quite flattering. I actually love to hear what my cousins across the pond in the UK always have to say. I feel such a kinship with you guys because I never quite felt that I'm like my fellow United States counterparts/colleagues. Fellow engineers from the United States generally come off as being almost too extreme to me. While I find most of you United Kingdom folks to be much more less neurotic & extreme while being quite a bit more erudite with a methodical calm about you all. I mean you are a different culture from us even though we are so close to being alike. I suppose I'm just as extreme so I'm probably the pot calling the kettle black or at least smoking the pot like Bob Dylan turning on the Beatles. Or so the story has gone I've heard. And I really think it was you guys that really changed the way everybody in the world now records things. And of course some of the most desirable equipment came from across your street. But how can you guys be so good and still have lost that war in 1776? Maybe too methodical in your style? But it's all good. It worked out great in the end and we got lovely friends across the pond. But why didn't you guys come up with some better camouflage instead of those bright red coats? The police here still love stopping sporty little cars that are bright red. So I guess it's a hangover from times gone by?

I didn't drink anything last night so no hangover here. But I'd love to get together and share a pint with ya if the opportunity ever arises. Maybe we should just do it via Skype?
Mx. Remy Ann David

IIRs Wed, 11/30/2011 - 13:31

We brits have been lucky to have the BBC I think. I wonder if it will still exist once our current tory government finishes raping the country and grinding us all back to serfdom?

RemyRAD, post: 379835 wrote: I didn't drink anything last night so no hangover here. But I'd love to get together and share a pint with ya if the opportunity ever arises. Maybe we should just do it via Skype?
Mx. Remy Ann David

yeah, I was sorry not to make it across the pond on that last tour I did. But US bands have to jump through a lot of hoops to get visas for UK crew, so they ended up using US engineers for that leg of the tour. Of course I run a studio these days, so I can't afford things like holidays anymore...

You should come and visit the UK sometime. Then I could buy you a pint of proper Real Ale instead of that watery rubbish that passes for beer over your way. (y)

RemyRAD Thu, 12/01/2011 - 01:16

Yeah, lousy imitation of real beer. I really loved my time driving through Germany, Austria, Switzerland, southeast France. Americans need to know what real beer, real wine & unpasteurized cheese tastes like! The autobahns were a blast with the Mercedes 190 E I rented. Didn't get a chance to patch myself across the Channel. Back in the mid-1980s. I was suffering because I was so close yet so far. I let my passport expire and now I have to go to this whole blasted process all over again. I'll get there, someday. Certainly not to seek employment because there's no way I could compete with guys like yourself. Besides I'm too old! This over 50 stuff bites the big banana.

Strange? Bananas were the first solid food my mother fed me. I'm reverting back to childhood because of senility.
Mx. Remy Ann David