Real-time Parametric EQ Fiddling

[Sehnsucht]

I'm using moOde purely to access the parametric eq, which makes a lot of difference to how my headphones sound!  Something I'm not able to do though, it seems, is make changes to the eq settings and hear the response immediately.   I have to make a change blind, save, then test, and the currently playing song starts from the beginning.   Is there no way to keep the current song playing and apply the change with save & test, or better still let me move the sliders and hear the results immediately?  Some parts of the song are better than others to detect differences.

[Tim Curtis]

DSP's including Crossfeed, Polarity inversion, Parameteric EQ, GraphicEQ require MPD to be restarted in order to pick up any changes to the settings so not really possible to do anything "real-time".

CamillaDSP Pipeline editor can perform the changes without restarting MPD so you might want to give it a try since it can do Parametric EQ.

[jonners]

Also there are autoeq files available for thousands of headphone models here: https://github.com/jaakkopasanen/AutoEq/...er/results
They can easily be loaded into Camilla in Moode.

[Sehnsucht]

Also there are autoeq files available for thousands of headphone models here: https://github.com/jaakkopasanen/AutoEq/...er/results
They can easily be loaded into Camilla in Moode.

Yes, I've been using the settings from there because he seems to have based those settings on Oratory1990's settings but the Oratory's require high/low shelf filters as well as peak, whereas the parametric eq in moOde only supports peak.  Do you know how to convert from one to the other?  

Camilla looks powerful but also scary and I have no idea how it works.  But you've inspired me to give it a try.    When I'm playing with camilla or parametric eq do I have to worry about the distinction between 44.1kHz and 48kHz?    On AutoEq's parametric eq notes he writes:

Parameters produced by AutoEq are equal with EqualizerAPO using 48 kHz sampling rate. When using other equalizers or sampling rates, it's always highly recommended to check that the frequency response of the equalizer matches the parametric eq curve in the graphs.

I listen to 44.1kHz music only although I'm not sure he's talking about that.

[jonners]

@Sehnsucht
Just follow the guide in the first post of the following thread and enable a 'Quick Convolution Filter': http://moodeaudio.org/forum/showthread.php?tid=3617
Camilla in Moode 7.3 should accept .wav files so there is no need to convert them.

[Sehnsucht]

@Sehnsucht
Just follow the guide in the first post of the following  thread and enable a 'Quick Convolution Filter':  http://moodeaudio.org/forum/showthread.php?tid=3617
Camilla in Moode 7.3 should accept .wav files so there is no need to convert them.

Sadly, I'm not getting anywhere.  The first post says essentially to:
download the .wav
upload to convolution file
hit convert
apply to l & r in Quick Convolution filter
get green tick
click save

I've uploaded the wav.
There's no convert anywhere.  You say it should accept .wav so ignore that step.
Apply the same .wav (i just uploaded) to ir left/right in the quick convolution filter section. I set ir type to wave. 
I click save.  I now see "version : segmentation fault" above the save button.  (Rebooting and setting configuration in General to off doesn't remove this message. Maybe it's always there?). 
Do i need to set pipeline configuration?
Do I need to set Configuration to quick convolution filter, seeing as that's what i'm doing after uploading the .wav?  I don't hear any music and there's a red cross under camilladsp on the audio config page.  But there's always a red cross under ir type in quick convolution filter.

This is on the latest moOde on a pi zero w.  Parametric eq was working fine before.  I've tried all 3 available .wav files for the hd 595 headphones from the autoeq github page.  There's nothing very exciting in /varlog/syslog or daemon.log, or /var/log/mpd/log.

Some relevant (?) bits from system info:
moOde release = 7.3.0 2021-06-25
RaspiOS = 10.6
Linux kernel = 5.4.77+ #1371
Platform = Pi-Zero W 1.1 512MB
Architecture = armv6l (32-bit)
Parametric EQ = Off
Graphic EQ = Off
CamillaDSP = __quick_convolution__.yml

[jonners]

Perhaps someone with more knowledge than I have can help. I'm wondering if the RPi Zero has enough processing power for this job?
Yes you do have to set Configuration to Quick convolution filter, but you shouldn't need to touch the pipeline config.

[Tim Curtis]

In Audio Config, the CamillaDSP SET button (and thus the option itself) would be disabled on a Zero. It's only supported on Pi-2B v1.2, Pi-3/4 and Allo USBridge SIG.

[Sehnsucht]

In Audio Config, the CamillaDSP SET button (and thus the option itself) would be disabled on a Zero. It's only supported on Pi-2B v1.2, Pi-3/4 and Allo USBridge SIG.

Ah!  That might explain it.  Shame the "CamillaDSP" button isn't disabled!!  I'll have to try again on my Pi 4.

[Tim Curtis]

That could be a bug. I'll add to the TODO list to investigate.