search

jamoma / JamomaCore

Jamoma Frameworks for Audio and Control Structure
Other
36 stars 14 forks source link

2nd order linkwitzriley filters are phase inverted #280

Open Nilson opened 10 years ago

Nilson commented 10 years ago

The 2nd order Linkwitz-Riley Highpass and Lowpass filter are phase inverted to each other, which results in an unwanted sound coloration when summing their output. See patch and screen shot. This might be a bug in the original filter design, I don't know. 


----------begin_max5_patcher----------
752.3oc0WtsbZCCDF9ZySgFM8tRXrjwFnWk9bzISFisvnTaIWI4.jLgm8pC1
IjvISBmxMhQKqj92O1Uq34Ndvw74DID7Kve.ddO2wyyZxXvqdtGrHddRdrz5
FLgWTPXJXW22oHyUV6xRRhRTU.3S.ozISHBBKgz31DNSIoOQLthv87aLKxFa
L4+pEVUAkkST1CCsxpYwE1UC+sfFmCeycdkpw+lMoLVkLkxxtWn0jK3BQA8B
6B5GFpOJ.Jzbh.rdDbW8hno1smO9gavXnw1Kc5XF5dDIirp3pAIA99G.RPGS
jPYkUJPCXtdRRFXPPKAh+mAHtPlKS3kjkM5SxY7UiDYhfmmarDro.FuwvBs2
beLxEW8sQINHXk.qPKgjoBtCe9qGtngMp0clpEkD2tCgutK47rIBx+rx46Bb
pqB9ZvYvoGNLxL8QsVkzMK2S0yFoSaqU1eREZHx.r.rEhaoVAEsU9HoYL8oC
t63gjedoPRcpT6PR3oAIUEiIhCI9OzqOw6Hk.6Jdri691ST+sF9TS+B.bbLK
6shmxXgVaJh3dBKdbtqP7SvmBhTFmQVKmwTYVoevxBvOPWpBpZt02LNzROzv
MSuQ655liWgDiSkjkmxjIz95EisbneztxkFbRpjtLcpcA8WpWD5h0K5gdSn4
5Zzk.D3Vy4Bx4yJ0KpWNk82YT0SBZNYQOL31Wq2P999Maj1KRBuhoVEnsIsK
33cGFdzHK68ce3x7L+ProLu81Vq6QthbcBOklM8aFhqunqAwAt+X1VPbvAiX
6tXCzO7+ZsRwX+8bWxqDIMaZ88If2zRJQpnrXEkyVwmf24yTZZJgsZAYJUZZ
Skt89TsUMitpTSPKTi4oM57yqF4Dc1jSaxbhtpxbNepIpsYNWQx47QGTqzyv
ymdBaidFb1zSaprLu74bgm9sPOi9jxw0.Ktr7QhPVukVknehvCbgYZTW6TJy
M09hRnf7Hsw+AVKwBcuWktwakv0vd9vHXGy47Rm+C.cuWXE
-----------end_max5_patcher-----------

screen shot 2014-03-21 at 10 05 40 pm

tap commented 10 years ago

This is one of the caveats of a 2nd-order LR filter. There is a nice little summary of different crossover filter types in this document:

http://www.rane.com/note147.html

The 2nd-order LR does not feature a flat summed-response like the 4th-order. FWIW, the cross~ object in MSP uses 3rd-order Butterworth filters.

best, Tim

On Sat, Mar 22, 2014 at 12:11 AM, Nils Peters notifications@github.comwrote:

The 2nd order Linkwitz-Riley Highpass and Lowpass filter are phase inverted to each other, which results in an unwanted sound coloration when summing their output. See patch and screen shot. This might be a bug in the original filter design, I don't know.

----------begin_max5_patcher---------- 752.3oc0WtsbZCCDF9ZySgFM8tRXrjwFnWk9bzISFisvnTaIWI4.jLgm8pC1 IjvISBmxMhQKqj92O1Uq34Ndvw74DID7Kve.ddO2wyyZxXvqdtGrHddRdrz5 FLgWTPXJXW22oHyUV6xRRhRTU.3S.ozISHBBKgz31DNSIoOQLthv87aLKxFa L4+pEVUAkkST1CCsxpYwE1UC+sfFmCeycdkpw+lMoLVkLkxxtWn0jK3BQA8B 6B5GFpOJ.Jzbh.rdDbW8hno1smO9gavXnw1Kc5XF5dDIirp3pAIA99G.RPGS jPYkUJPCXtdRRFXPPKAh+mAHtPlKS3kjkM5SxY7UiDYhfmmarDro.FuwvBs2 beLxEW8sQINHXk.qPKgjoBtCe9qGtngMp0clpEkD2tCgutK47rIBx+rx46Bb pqB9ZvYvoGNLxL8QsVkzMK2S0yFoSaqU1eREZHx.r.rEhaoVAEsU9HoYL8oC t63gjedoPRcpT6PR3oAIUEiIhCI9OzqOw6Hk.6Jdri691ST+sF9TS+B.bbLK 6shmxXgVaJh3dBKdbtqP7SvmBhTFmQVKmwTYVoevxBvOPWpBpZt02LNzROzv MSuQ655liWgDiSkjkmxjIz95EisbneztxkFbRpjtLcpcA8WpWD5h0K5gdSn4 5Zzk.D3Vy4Bx4yJ0KpWNk82YT0SBZNYQOL31Wq2P999Maj1KRBuhoVEnsIsK 33cGFdzHK68ce3x7L+ProLu81Vq6QthbcBOklM8aFhqunqAwAt+X1VPbvAiX 6tXCzO7+ZsRwX+8bWxqDIMaZ88If2zRJQpnrXEkyVwmf24yTZZJgsZAYJUZZ Skt89TsUMitpTSPKTi4oM57yqF4Dc1jSaxbhtpxbNepIpsYNWQx47QGTqzyv ymdBaidFb1zSaprLu74bgm9sPOi9jxw0.Ktr7QhPVukVknehvCbgYZTW6TJy M09hRnf7Hsw+AVKwBcuWktwakv0vd9vHXGy47Rm+C.cuWXE -----------end_max5_patcher-----------

[image: screen shot 2014-03-21 at 10 05 40 pm]https://f.cloud.github.com/assets/113509/2489907/c725ec10-b17f-11e3-88ba-4bb5c36fed16.png

— Reply to this email directly or view it on GitHubhttps://github.com/jamoma/JamomaCore/issues/280 .

Nilson commented 10 years ago

thanks for the pointer Tim! The Bessel crossover looks interesting. Maybe we need a j.cross~ object to encapsulate the matching crossover filters.

lossius commented 9 years ago

Do that mean that this is not an issue, and could be closed? Other filter designs should rather be added as separate enhancement issues in the tracker.

I'm assigning to you @tap to close this issue if LinkwitzRiley 2nd order works as expected.

tap commented 9 years ago

​I would like to make sure this is documented, then we can close it.​

Tim

On Mon, Mar 30, 2015 at 4:30 AM, Trond Lossuis notifications@github.com wrote:

Do that mean that this is not an issue, and could be closed? Other filter designs should rethater be added as separate enhancement issues in the tracker.

I'm assigning to you @tap https://github.com/tap to close this issue if LinkwitzRiley 2nd order works as expected.

— Reply to this email directly or view it on GitHub https://github.com/jamoma/JamomaCore/issues/280#issuecomment-87611649.