Guest

— Forum Scope —

— Match —

— Forum Options —

Minimum search word length is 3 characters - maximum search word length is 84 characters

Multiple frequency bands
January 12, 2015
6:17 pm
JD Young
Leiden, The Netherlands
Forum Posts: 80
Member Since:
November 5, 2014
Offline

Hi Will, how are you?

I’m re-reading your plug-in book at the moment and trying to make sense of the filter theory:) I’m building a mastering plug-in for which I need 6 frequency bands, mainly in the low end of the frequency spectrum. I have tried the second order Linkwitz-Riley filters from your book, and also 6dB one pole filters, but after adding the separate bands again it sounds different than the input signal. I’ve tried with separate LP and HP filters, and with constructing the HP’s as in HP = Input minus LP, but no succes…

Might it be that these crossovers are just not very accurate in the low end, or do you have a tip for a filter that might do the job?

All the best,

JD

January 12, 2015
8:19 pm
W Pirkle
Forum Posts: 247
Member Since:
January 29, 2017
Offline

The LW filters actually work well for this the only thing you have to remember is to invert the phase of "one side" of the bandpass, as in the 3-band Spectral Compressor from the FX book. Have you experimented with that one?

You can also use the Analyzer's button to take the actual frequency response of the bank of filters after adding back together - you should get a perfectly flat response, and you can verify this with the Spectral Compressor example.

Will

January 12, 2015
8:28 pm
JD Young
Leiden, The Netherlands
Forum Posts: 80
Member Since:
November 5, 2014
Offline

I totally forgot about that! thanks

January 13, 2015
1:36 am
JD Young
Leiden, The Netherlands
Forum Posts: 80
Member Since:
November 5, 2014
Offline

Hi Will,

Besides from the LW filters and inverting the phase, I might have found another solution. I applied 5 simple LPF’s – low to high frequency - and just subtracted them from each other.

Band1 = LP1
Band2 = LP2-LP1
Band3 = LP3-LP2
Band4 = LP4-LP3
Band5 = LP5-LP4;
Band6 = InputSample-LP5

This also sounds exactly the same after recombining the bands, while taking less CPU. Is there a disadvantage to this approach I’m not hearing/seeing? (in this case it's just for subtly narrowing the stereo image in 5 steps)

JD

March 18, 2015
4:09 pm
JD Young
Leiden, The Netherlands
Forum Posts: 80
Member Since:
November 5, 2014
Offline

Hi Will and Tom

I’m still messing around with different multiband setups, using the Linkwitz-Riley filters from the FX book. For some purposes they perform really well, but (in cases with more than two bands) when the crossover frequencies become closer to each other, the frequency response becomes less flat. I’ve compared this to some commercial plug-ins I have used for a while, and they don’t seem to have the same issue. I was wondering what’s their secret? Or what I possibly did wrong…

I searched the web for more information about this issue, but I can’t seem to find a clear solution. Some people are talking about linear phase FIR filters, or all pass filters to compensate for the phase differences. Can you maybe point me in the right direction?

All the best,

JD

April 17, 2015
4:58 pm
W Pirkle
Forum Posts: 247
Member Since:
January 29, 2017
Offline

Hi JD

Sorry for the late reply here - a good place to start would be the filterbanks used for MPEG encoding and/or QMF (quadrature mirror filters). I'll try to post more information after the semester ends.

- Will

April 18, 2015
10:26 pm
JD Young
Leiden, The Netherlands
Forum Posts: 80
Member Since:
November 5, 2014
Offline

Google is my friend I'll look into it. Tnx Will!

October 19, 2015
1:59 am
JD Young
Leiden, The Netherlands
Forum Posts: 80
Member Since:
November 5, 2014
Offline

Hi Will,

I gave building a multiband plug-in another go today, and I finally found the problem I've had before with the Linkwitz-Riley filters. In my first attempt a while ago, I did not use them hierarchically, but I used them all on the source signal. This resutled in a far from flat frequency response with more than one split. This time I split the signal in two bands, and then I split the two bands again in different sub-bands. After adjusting the phase everything seems to work fine! There are slight amplitude variations if the crossover frequencies get really close, but these difference are negligible.

In the meantime however, I did experiment with the subtractive filter approach, with an absolutely flat response as a result. I’m going to compare the two, to see what approach would be best for plug-ins for mastering.

Cheers! JD

Forum Timezone: America/New_York

Most Users Ever Online: 36

Currently Online:
4 Guest(s)

1 Guest(s)

Top Posters:

Skyler: 48

Derek: 46

Peter: 41

Frodson: 40

clau_ste: 39

Gwen: 32

JimmyM: 30

TheSmile: 27

EZB: 24

lppier: 23

Member Stats:

Guest Posters: 1

Members: 522

Moderators: 1

Forum Stats:

Groups: 11

Forums: 31

Topics: 545

Posts: 2134