CamillaDSP IIR format output?

[TNT]

The CamillaDSP - https ... github.com/HEnquist/camilladsp#introduction

has started to attract a lot of attention.

This is a request for support of the CamillaDSP text format for IIR PEQ - to be outputted by REWs EQ function.

The format is described here:

GitHub - HEnquist/camilladsp: A flexible cross-platform IIR and FIR engine for crossovers, room correction etc.

A flexible cross-platform IIR and FIR engine for crossovers, room correction etc. - HEnquist/camilladsp

github.com

It would be a great help.

Thanks!

 

[ddude003]

Although I have a couple of convolution engines at hand, I might be interested in trying CamillaDSP for FIR room correction on MacOS... Is there a Users Doc that describes straight forward steps for installation and an example filter chain for a vanilla 2 channel stereo FIR convolution?

 

[TNT]

I'm "new" here so wasn't allowed to post links... This is the available instructions but several use posted guides exists if you make a search.

GitHub - HEnquist/camilladsp: A flexible cross-platform IIR and FIR engine for crossovers, room correction etc.

A flexible cross-platform IIR and FIR engine for crossovers, room correction etc. - HEnquist/camilladsp

github.com

I run CDSP on a Mac mini and do use a combo of IIR and FIR.

If you don't fancy installing things etc perhaps use one of one of the packages that is delivered with CDSP integrated is a way to do it e.g, in Moode.

Good luck!

//

 

[John Mulcahy]

This is a request for support of the CamillaDSP text format for IIR PEQ - to be outputted by REWs EQ function.

I'll add it to the ToDo list.

 

[TNT]

Thank you so much John!

//

 

[John Mulcahy]

Support for CamillaDSP IIR filters with YAML file export/import has been added in the latest V5.20.14 early access build.

 

[TNT]

If I have installed and run V5.20.13 MacOS, may I just install this over it - the easy way upgrade as usual?

//

 

[John Mulcahy]

Yes.

 

[TNT]

Tested... :-)

filters:
filter_1:
parameters:
freq: 70.3
gain: -1.6
q: 3.521
type: Peaking
type: Biquad
filter_2:
parameters:
freq: 97.60000000000001
gain: 2.9000000000000004
q: 2.001
type: Peaking
type: Biquad

The many decimals and the odd 1 and 4 in the end - how come?

REW do not do tighter than +/-1 dB for auto EQ so the great precision seem a bit "unbalanced"?

//

 

[TNT]

Also tried to import an yml file I use but it gave parse error...

//

 

[John Mulcahy]

The random digits are snakyaml's representer getting a bit over-enthusiastic with doubles. I looked at replacing its numeric representer but it was a bit of a faff so I didn't bother

Your file is not valid YAML, the error messages are spectacularly unhelpful but the problem is an extra space in front of the label for the first filter.

 

[TNT]

OK - the yml file came out of CamillaDSP using its GUI to create it.

 

[ddude003]

Your file is not valid YAML, the error messages are spectacularly unhelpful but the problem is an extra space in front of the label for the first filter.

What the heck kind of parser worries about spaces... Come on CamillaDSP give me a break... OMeta anyone?

@TNT had said "If you don't fancy installing things etc perhaps use one of one of the packages that is delivered with CDSP integrated is a way to do it e.g, in Moode."

No worries, been using one version or another of Nix since the early 80's... AT&T Sys V + BSD enhancements anyone?

 

[John Mulcahy]

YAML uses spaces for nesting, like Python, so that's a feature of the configuration file format rather than CamillaDSP.

 

[TNT]

One could still argue that someting that one cant see has severe significance isn't so "UX" but maybe its not ment for humans to write this by "hand"?

It will be OK.

Now I'm just waiting for CamillaDSP to be able to import a text file of filters via the GUI so that one don't have to fiddle with an editor. Henrik, the designer of CDSP says it's on it's way. But it saves a lot of typing already ;-D thank you very much.

//

 

[ddude003]

It just seems silly that in this day and age we have off-side languages where strict formatting is for "human readability"... That's what parsers, pritty-printers, browsers and xtanslation utilities are for... If you don't like my style of writing a program then just whip out your favorite TXL or browser and change the "formatting" to your taste... Been smalltalking with a lisp since the late 70s... Not much new in hardware or software in the last 50 years... Just smaller and faster... And a lot of bad ideas codified by the likes of Micro$oft and others...

White space... Three too many, four not enough? Lol...

 

[TNT]

Using the Camilla support with good result. Looking at the text describing the option to output yml...

"Save filter settings to file"

"Setting" ?? Is this really the best word for it?

Next option is "... export... as text.." - the yml is also a text file...

Save vs. export... whats the difference really?

This section is not so intuitive and made me have to test most options in order to get what I wanted.

Maybe: 'Save Equaliser filters' / Load Equaliser filters.

And perhaps it is possible to move the Export "as text" and "as formatted text" to be an option in the save/load dialogue so the "Export" lines/options could be skipped all together?

//

 

[John Mulcahy]

I'll change it to "Save filter settings to YAML file".

 

[Wirrunna]

Could the Model field which currently is set to Filters be able to be overwritten so that the input is used instead of filter_1: , filter_2: . i.e. if it is set to Right Bass then the filter set is Right_Bass_1: , Right_Bass_2: etc.

 

[John Mulcahy]

Could the Model field which currently is set to Filters be able to be overwritten so that the input is used instead of filter_1: , filter_2: . i.e. if it is set to Right Bass then the filter set is Right_Bass_1: , Right_Bass_2: etc.

No, but REW can prompt for a label to use whenever CamillaDSP filters are saved. I have updated the V5.20.14 early access builds to add that.

 

[Wirrunna]

No, but REW can prompt for a label to use whenever CamillaDSP filters are saved. I have updated the V5.20.14 early access builds to add that.

Excellent, it does ask now.

I have attached a .mdat file that says it is 96 kHz file but when I select EQ and then pick CamillaDSP as Equaliser it says Equaliser: CamillaDSP(48kHz) - see attached screengrab, the REW measurement screen shows file details at bottom left and top right.

 

[TNT]

Is the Fs really relevant for IIR filters? For any equaliser?... or maybe some?

//

 

[John Mulcahy]

Excellent, it does ask now.

I have attached a .mdat file that says it is 96 kHz file but when I select EQ and then pick CamillaDSP as Equaliser it says Equaliser: CamillaDSP(48kHz) - see attached screengrab, the REW measurement screen shows file details at bottom left and top right.

It's actually a 6 kHz measurement. The Decimate IR Analysis preference allows REW to reduce the sample rate of measurements made over a limited frequency range to speed up processing and reduce storage. A 500 Hz measurement really doesn't need 96 kHz sampling, but you can turn that option off if you wish.

 

[John Mulcahy]

Is the Fs really relevant for IIR filters?

The filter shapes will be increasingly affected as frequency gets closer to half the sample rate.

 

[TNT]

OK - like a general choice and aspects of Fs for reproducing sound. But the filter parameters dont include a term of Fs - do they?

//