1st REW Measurements for Camilla + Idiot = PLEASE HELP!! Obvs

[Deleted member 23424]

Come on now, we all knew this would happen.

By following the Julian Krausse video on YouTube that I've seen recommended on here several times I've got the measurements side of things done OK. They certainly look like what I was expecting anyway.

But my three attempts to transfer this into Camilla (Moode baked in) thus far just produce an unlistenable distorted mess.

I created the below text file to use as the House Curve. This is the text provided underneath said video and is supposed to represent the HK curve for speakers.

Here is said House Curve overlaid the average curve of my measurements.

Match Response To Target (with the equaliser set to Generic) produces this:

I then click Export Filter Settings as text and save (file name HELP)


CAMILLA

In Moode's webUI, all other EQs set to off, I've uploaded HELP as a convolution file.

IR RIGHT and IR LEFT both set to HELP.
IR Type is a guess at S24LE (but I've tried all options to no avail)

HELP! I think everything is ok up until either the point of Convolution File creation OR installation.

No doubt I've missed a glaringly obvious thing out, if anyone has any ideas it would be very much appreciated.

Cheers

 

[Deleted member 23424]

Have also tried other House Curve files, same outcome.

As soon as I turn Camilla off the sound is ok again.

 

[Deleted member 23424]

This is the contents of the Convolution File 'HELP'.

Which makes me wonder what the point of me using Camilla is? Couldn't I just add these values to the PEQ instead?

 

[Deleted member 23424]

Said JK video

 

[XpanD]

Are you sure you're using the right file in the right place? The file you're exporting is indeed just a list of PEQ filters. The "save filter coefficients as file" button also spits out a text file, I have a feeling that may be the one you need.

 

[Deleted member 23424]

Are you sure you're using the right file in the right place? The file you're exporting is indeed just a list of PEQ filters. The "save filter coefficients as file" button also spits out a text file, I have a feeling that may be the one you need.

I looked at that earlier, but it only gives two options, neither of which seem appropriate.

This is the contents of the mini DSP version

 

[Deleted member 23424]

Tried using said miniDSP file created as per above, caused an odd ringing noise and distortion.

 

[XpanD]

Had a look at the video, I'm guessing CamillaDSP requires a slightly different approach. From its Github page description:

REW can automatically generate a set of filters for correcting the response. These can then be exported as an .xml-file. This file can then be translated to CamillaDSP filters using the translate_rew_xml.py Python script. This will generate filters and pipeline steps that can be pasted into a CamillaDSP config file. This script currently supports only Peaking filters.

To generate said .xml file... (taken from here)

REW automatically creates the correctly formatted .xml file if we select 'rePhase' as the equalizer when select 'save settings to file' is selected.

Was hoping it'd take one of the two coefficient export options, but info on the format it actually uses seems pretty spotty. Hope this helps any!

 

[Deleted member 23424]

Had a look at the video, I'm guessing CamillaDSP requires a slightly different approach. From its Github page description:



To generate said .xml file... (taken from here)



Was hoping it'd take one of the two coefficient export options, but info on the format it actually uses seems pretty spotty. Hope this helps any!

Cheers

Hmmmm. I guess I need to work out what Python is and how to use it.

This has not been the enjoyable afternoon of tinkering I'd intended it to be.

 

[Deleted member 23424]

These are the alternative EQ options built into REW. Might one of these be compatible with Camilla?

 

[mdsimon2]

miniDSP uses a slightly odd biquad configuration from the rest of the world. If you switch the signs on A1 and A2 when entering in to CamillaDSP it should give you the expected response. You should be able to view the response in the Filters tab, it will also tell you if the biquad is stable.

Another thing to keep in mind is that biquads are only valid for a single sample rate, so if you generate them at 48 kHz you need to make sure CamillaDSP is always running at 48 kHz. I learned today that the default in Moode is to switch CamilaDSP sample rate based on source file.

Michael

 

[vco1]

This has not been the enjoyable afternoon of tinkering I'd intended it to be.

I'm with you. CamillaDSP surely looks like a wonderful tool. If you fully know what you're doing, that is. I have it installed on an RPi4. But that's about it. No clue where to start. The github page is of not much use either.
I might go for a miniDSP Dirac solution in the end. Way more expensive, but at least that seems to be kind of plug and play.

 

[Deleted member 23424]

miniDSP uses a slightly odd biquad configuration from the rest of the world. If you switch the signs on A1 and A2 when entering in to CamillaDSP it should give you the expected response. You should be able to view the response in the Filters tab, it will also tell you if the biquad is stable.

Another thing to keep in mind is that biquads are only valid for a single sample rate, so if you generate them at 48 kHz you need to make sure CamillaDSP is always running at 48 kHz. I learned today that the default in Moode is to switch CamilaDSP sample rate based on source file.

Michael

Oh lord. This sounds like an absolute waste of time then. I'm beginning to think I might be better looking at alternative streaming/DSP options. On initial look MOODE seemed my best choice due to it being all under one roof but Qobuz bit rates aren't consistent therefore rendering this entire exercise pointless?

 

[mdsimon2]

I would try enabling resampling in CamillaDSP and set the sample rate to whatever you generate your biquads at. I think it should be setup such that it just changes the capture rate but then resamples to your specified playback rate but I am not sure how the specific rate switching implementation works.

I just setup a RPi4 with Moode today, I can play around with it tonight and try to see what it is doing.

Michael

 

[Deleted member 23424]

I'm with you. CamillaDSP surely looks like a wonderful tool. If you fully know what you're doing, that is. I have it installed on an RPi4. But that's about it. No clue where to start. The github page is of not much use either.
I might go for a miniDSP Dirac solution in the end. Way more expensive, but at least that seems to be kind of plug and play.

I don't mind getting in over my head, steep learning curves and all that. But this one seems to not only be fiddly as funk but ultimately, given the above bit rate revelation, it's beginning to look utterly pointless (for my use).

Might just use REW again but include dips too and use the resultant info in the PEQ. At least that'll work for all music.

 

[Deleted member 23424]

I would try enabling resampling in CamillaDSP and set the sample rate to whatever you generate your biquads at. I think it should be setup such that it just changes the capture rate but then resamples to your specified playback rate but I am not sure how the specific rate switching implementation works.

I just setup a RPi4 with Moode today, I can play around with it tonight and try to see what it is doing.

Michael

I've never even heard of a biquad, let alone generated one.

I think I'd be best going down the PEQ route I mentioned above before I end up starting a thread asking for assistance to reassemble my brain

 

[scrubb]

My take is that when computer hacking audio nerds create something which requires code level tinkering without any clear documentation or simple instructions, it will remain in the realm of computer hacking audio nerds and not be seen as valuable to anyone else.

 

[mdsimon2]

I've never even heard of a biquad, let alone generated one.

I think I'd be best going down the PEQ route I mentioned above before I end up starting a thread asking for assistance to reassemble my brain

Oh for sure, the PEQ route is definitely easier and makes more sense for this use case. When you are done entering the PEQs you should be able to view the combined output in the Pipeline tab and make sure it is doing what you think it should.

Michael

 

[mdsimon2]

My take is that when computer hacking audio nerds create something which requires code level tinkering without any clear documentation or simple instructions, it will remain in the realm of computer hacking audio nerds and not be seen as valuable to anyone else.

Personally I think the Moode implementation of CamillaDSP is unfortunate, it tries to automate too much and hide setup from the user which makes it more difficult to troubleshoot. It almost needs more of a setup questionnaire so you know exactly what it is doing.

CamillaDSP is exceptionally well documented and using the stock GUI is just as easy as using a miniDSP. Those of us with a smidge of technical ability will gladly use it over more expensive less powerful solutions, YMMV.

Michael

 

[Deleted member 23424]

Personally I think the Moode implementation of CamillaDSP is unfortunate, it tries to automate too much and hide setup from the user which makes it more difficult to troubleshoot. It almost needs more of a setup questionnaire so you know exactly what it is doing.

CamillaDSP is exceptionally well documented and using the stock GUI is just as easy as using a miniDSP. Those of us with a smidge of technical ability will gladly use it over more expensive less powerful solutions, YMMV.

Michael

Oh I've no doubt Camilla is a wonderful thing, alas my experience of being this far down amongst the wires is far less wonderful. I'm not sure if it's Camilla or Moode but I can't even see the wood from the trees on this one