Implementing a 1/3 Octave Filter

Feb 26, 2009 at 6:22pm

Implementing a 1/3 Octave Filter

Hi All,
I’m trying to simulate room reverberation using third-octave RT60 measurements (actual results are taken using RT30/20) and i was just wondering if anyone had much lucky implementing a third-octave (32 band) filter using MaxMSP?
I know it’s possible to replicate third-octaves using the FFT~ object, but FFT isn’t an option as for one it delays the signal chain etc.

Cheers all!

#42547
Feb 26, 2009 at 6:32pm

The [fffb~] object implements a chain of [reson~]-like band-pass filters. It might be worth a look.

lh

#152241
Feb 26, 2009 at 7:29pm

take a look at Elsea’s tutorial on the third octave filter:

ftp://arts.ucsc.edu/pub/ems/maxtutors/MSP.tutorials/ThirdOct.pdf

#152242
Feb 27, 2009 at 7:58pm
Visibleman wrote on Thu, 26 February 2009 10:22
Hi All,
I’m trying to simulate room reverberation using third-octave RT60 measurements (actual results are taken using RT30/20) and i was just wondering if anyone had much lucky implementing a third-octave (32 band) filter using MaxMSP?
I know it’s possible to replicate third-octaves using the FFT~ object, but FFT isn’t an option as for one it delays the signal chain etc.

Cheers all!

I made one using CNMAT’s peqbank~. Download the CNMAT Max/MSP/Jitter Depot and the peqbank~ object here:

http://cnmat.berkeley.edu/downloads

And search for:

+peqbank~_overview.maxpat

mz

#152243
Mar 1, 2009 at 10:57pm

Cheers all, but unfortunately FFT isn’t an option due to it being linear.

Mzed – I had a look at using the biquad as i have third-octave coefficients, so i’ll have a look into peqbank~. But not be a pain, have you got a direct link to the peqbank maxpatch? Had a look on here as well as CNMAT but couldn’t find a thing, cheers.

#152244
Mar 2, 2009 at 5:26am
Visibleman wrote on Sun, 01 March 2009 14:57
Cheers all, but unfortunately FFT isn’t an option due to it being linear.

Mzed – I had a look at using the biquad as i have third-octave coefficients, so i’ll have a look into peqbank~. But not be a pain, have you got a direct link to the peqbank maxpatch? Had a look on here as well as CNMAT but couldn’t find a thing, cheers.

Download this:

http://cnmat.berkeley.edu/files/maxdl/ANY/MMJ-DEPOT_2.0.5.zip

Search in there for “+peqbank~_overview.maxpat” in …/tutors/object-tutors/peqbank~/

I don’t know what platform you’re on, or I’d send a link to the object. Scroll down to peqbank~ here:

http://cnmat.berkeley.edu/downloads

and you’ll find links or either platform.

mz

#152245
Mar 2, 2009 at 12:57pm

Hi,
To make good 1/3oct filters for measurement, you should comply to IEC1260, this document tells you what the specs are. Hint: use three biquads in bandpass mode per filter, two with q=6, one with q=10 for flattop behaviour

greets,
Kessito

#152246
Mar 3, 2009 at 8:30pm

Hi all,

mzed – cheers for the link, been playing around with Peqbank all afternoon, but do you know if there is a performance hit using Peqbank~ over Biquad~?

Kessito – Yeah, IEC1260 is what i’m trying to design the filters to. But is the following patch about right?

– Pasted Max Patch, click to expand. –

Also, does anyone know where i might be able to get hold of 1/3 octave biquad coefficients? The coefficients i have currently are IIR Filter (ie: you have a/b[0] to a/b[6] instead of biquad a0, a1, a2, b1, b2 etc).

Cheers for the help so far, it’s appreciated greatly!

#152247
Mar 4, 2009 at 12:49pm
ComfortableInClouds wrote on Thu, 26 February 2009 12:29
take a look at Elsea’s tutorial on the third octave filter:

ftp://arts.ucsc.edu/pub/ems/maxtutors/MSP.tutorials/ThirdOct.pdf

Great !

Thanks !

#152248

You must be logged in to reply to this topic.