Thank you for your donation!


Cloudsmith graciously provides open-source package management and distribution for our project.


Problem: How to use Camilladsp Quick Convolution
#1
I tried creating a loudness contour filter and it worked well.  I decided to try the Quick Convolution feature.  I used the file creation procedure I've used in Hifiberry OS with the DAC+DSP hat that worked well.  

I created a sweep tone file in REW.  I played it back through Moode and the amplifier I'm using. I recorded it in my primary listening position using a Behringer calibration mic run through a Scarlett audio interface into Audacity.  The record level was set at about 70 dB.  I trimmed down the recording to include only the actual audio, eliminating the parts at the beginning and end where I was working the controls. 

I uploaded the file to Camilladsp in the Quick Convolution mode.  I used the mono file as both left and right channels.  I used the existing quick convolution pipline as a template.  The configuration checked out.  I tried playing back some audio.  I got a huge amount of distortion.  Baroque chamber music with a classic rock sound.  I tried lowering the mastergain filter by 30 dB.  It reduced the distortion some. 

The convolution filters that came with Moode worked fine in the sense of producing undistorted audio.  I'm sure I'm doing something wrong.  I hope it's simple, stupid, and easy to correct.  The documentation for Camilladsp doesn't directly address how to import an IR file for room correction, at least the list of instructions for dummies that I hoped to find.
Reply
#2
The quick convolution uses an IR file as input.
Is the uploaded file an audio recording of the sweep or an actual ir file? If it is the first you need to create an IR file with REW.
Personally I prefer moving mic with noise above a sweep on a single point.
Reply
#3
(01-03-2022, 09:27 PM)bitlab Wrote: Personally I prefer moving mic with noise above a sweep on a single point.

Confused  explain ?
----------
bob
Reply
#4
(01-04-2022, 05:17 AM)DRONE7 Wrote:
(01-03-2022, 09:27 PM)bitlab Wrote: Personally I prefer moving mic with noise above a sweep on a single point.

Confused  explain ?
 I also use the method of playing pink noise and holding and moving the microphone around the listening area for a while. REW will average the response over time.
Reply
#5
(01-04-2022, 05:17 AM)DRONE7 Wrote:
(01-03-2022, 09:27 PM)bitlab Wrote: Personally I prefer moving mic with noise above a sweep on a single point.

Confused  explain ?

I found it to be more stable (same response when moving your body/head in the listen position) then using sweep on one or more fixed positions. It also fast to measure and easy to do with REW.

A nice introduction to the subject is the following video:
https://www.youtube.com/watch?v=6RiuwqzjqlQ
And a discussion at https://www.audiosciencereview.com/forum...elp.12641/.
Or some more background info at https://www.ohl.to/audio/downloads/MMM-m...rement.pdf.
Reply
#6
That's probably it.  I uploaded the recording of the sweep.  That's the way it works in HifiberryOS.  I'll try REW and see how it works.  I had a feeling it was the file.


(01-03-2022, 09:27 PM)bitlab Wrote: The quick convolution uses an IR file as input.
Is the uploaded file an audio recording of the sweep or an actual ir file? If it is the first you need to create an IR file with REW.
Personally I prefer moving mic with noise above a sweep on a single point.
Reply


Forum Jump: