Make us a FIR, 8 pole, Butterworth High Pass Filter, please

[jeffca]

AUTO-ADMIN: Non-MP3, WAV, OGG, SoundCloud, YouTube, Vimeo, Twitter and Facebook links in this post have been protected automatically. Once the member reaches 5 posts the links will function as normal.

OK, you guys are some kind of coding geniuses so for one of you it shouldn't be hard (OK, it might be difficult). I need a finite impulse response, 48db/octave, butterworth high pass filter. Actually, everybody who does computer audio needs one, but they just don't know it.

At present, I'm giving a lot of work to ROM-plers like SampleTank and BFD. Until Elemental Audio released their tremendous, free Inspector RTA plug-in, I'd no idea that these sample synths produced so much bass and sub-bass garbage. To be honest, though, any source whether audio, soft synth or final mix can benefit from high pass filtering.

Using up to 3 instances of Cubase SX's Q filter per track really tightens up the low end of my mixes, but this is a rather inelegant way to solve the problem and less effective than than a FIR 48db/octave, butterworth high pass filter.

So why do I ask you to make all of us ( and more important to me: me!) this piece of specialized code?

• Finite Impulse Response filters have linear phase response and add no distortion to the signal. For this application, I don't want a filter with "character" (like just about every filter plug-in available).

• To my knowledge, Butterworth filters have the smallest "knee" to their filter slopes. This means that you can more precisely and effectively tune the filter for maximum noise attenuation with minimum effect to the main signal.

• Most filters, both commercial and otherwise are 12db/octave (2 pole). Very few of them say they are linear phase designs. I believe that FIR filters eat up more CPU cycles and, apparently, are harder to design.

• Nobody else seems interested in coding this type of filter because it's not going to be a big money maker and, as I already stated, a lot of people don't realize how much they need it. Once you see and hear what a really good high-pass filter does, believe me, you won't want to mix without it.

• Most of the really suberb filters are part of mastering EQ's, are for PowerCore & ProTools and they cost a lot of money.

So I'm begging you SmartElectronix code heads to do every aspiring DAW musician and engineer a big favor and make this very neccessary tool. I'd be willing to do beta testing and also pay a reasonable fee to guarantee that you make me an OS X VST version.

Jeff Henning
jeffca@comcast.net (mailto:jeffca@comcast.net)

[Jeremy_NSL]

I see your point, jeff.. but do you realize most of the 'useless subbass' on most well-sampled, or synthesized tracks is -70 or less db in amplitude? In a mix its often below the ATH (Absolute Threshold of Hearing). For certain sounds that have noticeable sub-bass/bass garbage I just use any decent parametric EQ and roll them off at 100hz or so.

Of course, the plug you mention would still be useful, but only if it could be done with low cpu overhead.. A fancy FIR filter that takes 10% or more of my cpu per instance is essentially useless..

[jeffca]

(I'll confess right now that I'm a total audio fanatic. No, I'm not one this those chumps who has to have Stereophile's new Amp of the Year every year. I love listening to and making spectacular sounding music in a variety of genres.)

Floyd, surely, how you use a tool like this would depend upon your system and what you're mixing. I'm running a G5 and have a PowerCore board so, for me, CPU drainage isn't as much of an issue as if I was still using an older machine.

In a worst case scenario, a filter plug like this one could be assigned to a group bus to conserve cycles. In any event, just one instance on a bus for signals above 75 to 100 hz and one on the master mix bus, set between 20 to 30 hz, will clean up the low end.

At present, I'm running around 12 to 18 tracks with SampleTank, BFD and Absynth. Turning the bass and kick drum off, the mixes have a total 50 to 60db noise floor below 125hz that undulates with each signal.

While you are correct that in this range, a signal like that is almost imperceptable the the human ear (I'm well versed in the Fletcher-Munsen equal loudness curves), that's not the big reason I need a steep, FIR hi-pass.

Left unchecked, that low end grunge modulates the real bass information and creates side bands that extend higher into octaves where your ear is more sensitive. Since the projects I'm on at present are R&B/hip hop stuff, it all gets pretty crunched in a 5 band compressor and the mid bass gets a bit congested.

Cassettes and LP's have about a 50db noise floor in the low end (that might be optimistic). Remember how stunning the bass on well recorded CD's sounded in comparison when they first arrived? That's what I'm talking about.

On my projects, the difference isn't that stark, but there is a difference that the artists and my producing partner could hear since I started filtering out that gunk .

Whether this type of filter or the improvement it can make is important to you or anyone else who'd read this thread, I can't say, but it does matter to me. And, as I qualified at the beginning of this post, I am an audio nut.

Jeff

[Bram]

hey jeff,


maybe you'd be better off asking some of the "regular" VST developers out there... Frankly your plugins might be "interesting" and usefulll, but coding them is everything but interesting.

Sorry man, we make plugins for FUN, and, while your idea(s) sound(s) nice for a musician, where's the fun in coding them?


- bram

[Bram]

oh, and, on a technical note:
FIR filters don't have poles. IIR filters do.
So, an 8-pole "FIR" butterworth doesn't exist.

Nor does a butterworth FIR filter exist: bufferworth filters are by definition IIR...

cheerio!

- bram

[jeffca]

Well, I just had to ask.

Thanks for the tech info. Nobody ever bothers mentioning that when they talk about this kind of heady stuff in MIX or EQ, but you knew what I meant.

Guess I'll just have to bite the bullet and shell out $500 for the Sony Oxford EQ. Aaaahh, crap!

Jeff

PS - I just downloaded a bunch of your OS X plug-ins and I'm looking forward to using them.

[antti]

Using up to 3 instances of Cubase SX's Q filter per track really tightens up the low end of my mixes, but this is a rather inelegant way to solve the problem and less effective than than a FIR 48db/octave, butterworth high pass filter.
So why do I ask you to make all of us ( and more important to me: me!) this piece of specialized code?

• Finite Impulse Response filters have linear phase response and add no distortion to the signal. For this application, I don't want a filter with "character" (like just about every filter plug-in available).

IIRs don't add distortion to the signal either. Nor do they necessarily have "character".

• To my knowledge, Butterworth filters have the smallest "knee" to their filter slopes. This means that you can more precisely and effectively tune the filter for maximum noise attenuation with minimum effect to the main signal.

No. For tuning as far frequency response goes, chebyshev/elliptic filters are superior to butterworth. Unfortunately their phase response is quite horrible.

I believe that FIR filters eat up more CPU cycles and, apparently, are harder to design.

I designed a linear phase FIR highpass filter with stopband attenuation 40dB (0-10Hz) and passband ripple 0.1dB (20Hz-22kHz). The filter order was 9206. Which means the cpu usage would be rougly equal to 2000+ order IIR filter.

[scuzzphut]

or you could just turn down the bass knob

[KiloWatts]

this post is kinda silly