Hi Henrik, all,
On Michael's camilladsp on RPI4 tutorial thread, I have posted about my tedious discovery that RPI3 was buggy with USB audio. Along my camilladsp journey I have made multiple other naive attempts (RPI3 with HDMI multimedia receivers; FireWire; Linux on MacMini to exploit its FireWire output; etc.) most of those have failed. I do not want to pollute this thread with such knowledge. I however believe some future readers may benefit from such knowledge. And I still secretly dream someone would end up explaining me how to activate FireWire of MacMini late 2012 with linux, so that I do not have to keep an obsolete iOS server on my home network, and so I can reuse cheap audio interface like RME FireFace400 or Motu Traveler MK1. Therefore my question: would you see a way for the community to share their noticeable working and failing camilladsp configurations ?
On Michael's camilladsp on RPI4 tutorial thread, I have posted about my tedious discovery that RPI3 was buggy with USB audio. Along my camilladsp journey I have made multiple other naive attempts (RPI3 with HDMI multimedia receivers; FireWire; Linux on MacMini to exploit its FireWire output; etc.) most of those have failed. I do not want to pollute this thread with such knowledge. I however believe some future readers may benefit from such knowledge. And I still secretly dream someone would end up explaining me how to activate FireWire of MacMini late 2012 with linux, so that I do not have to keep an obsolete iOS server on my home network, and so I can reuse cheap audio interface like RME FireFace400 or Motu Traveler MK1. Therefore my question: would you see a way for the community to share their noticeable working and failing camilladsp configurations ?
I haven't done it on linux but I have used a Mac mini 2012 with a variety of firewire interfaces (MOTU 896, Apogee Ensemble, RME Fireface 800) on Mac OS and CamillaDSP works really well.
Michael
There is a new gui release! The main new feature is import. You pick a config file, and then you get a prompt to select what parts of the config to import. can import selected parts of a config.
It supports camilladsp v1 and v2 (including the partial configs exported by REW), EqualizerAPO, and Convolver.
Get it from here:
https://github.com/HEnquist/camillagui-backend/releases/tag/v2.1.0
Updated setup scripts:
https://github.com/HEnquist/camilladsp-setupscripts/releases/tag/v2.1.0
It supports camilladsp v1 and v2 (including the partial configs exported by REW), EqualizerAPO, and Convolver.
Get it from here:
https://github.com/HEnquist/camillagui-backend/releases/tag/v2.1.0
Updated setup scripts:
https://github.com/HEnquist/camilladsp-setupscripts/releases/tag/v2.1.0
Hello,
Having realized (above) I can't use the RPI3 I have decided to do a fresh camilladsp install of my AMD64 / Intel celeron box - with Ubuntu server 22.04.
I have fist used Michael's instruction for RPI4 - installing "camilladsp-linux-amd64" instead of "camilladsp-linux-aarch64".
Attempting to run camilladsp gives following error
Now, I then have made the assumption that maybe the RPI4 tutorial would not be appropriate for my intel machine.
I therefore have attempted to follow the installation steps of @HenrikEnquist with "full_install_venv.sh" - following the readme of the page camilladsp-setupscripts.
Below the issue I have had, and the fix I have attempted:
First attempt:
Then I had following error:
I am at the end of my ideas how to progress. Although I know I have made it once on the previous desktop install of ubuntu on this same machine. But I do not remember what I've done.
Thank you for any suggestion how to progress here.
Merci !
Having realized (above) I can't use the RPI3 I have decided to do a fresh camilladsp install of my AMD64 / Intel celeron box - with Ubuntu server 22.04.
I have fist used Michael's instruction for RPI4 - installing "camilladsp-linux-amd64" instead of "camilladsp-linux-aarch64".
Attempting to run camilladsp gives following error
This is my first issue; I have no clue what it means, if you have a simple fix I'd be grateful.
Now, I then have made the assumption that maybe the RPI4 tutorial would not be appropriate for my intel machine.
I therefore have attempted to follow the installation steps of @HenrikEnquist with "full_install_venv.sh" - following the readme of the page camilladsp-setupscripts.
Below the issue I have had, and the fix I have attempted:
First attempt:
I then have installed unar (sudo apt install unar)
Then I had following error:
I then have installed the python3 & pip related packages as on Michael's RPI4 thread - to which I've added "python3-venv" (cause it felt right ...).
Then I had following error:
I have then copied the script locally and replaced the two occurence of "python" by "python3" outside the comments, with which the script has progressed further. Next error:
I have tricked the script by putting a hard link to an actual duplication of the "requirements.txt" file which I had found in "~/camilladsp/gui -D/camillagui/". This has brought me to:
However .. the directory /home/daniel/camilladsp/gui is empty, it does not contain a "main.py" file. The script actually has failed.
I am at the end of my ideas how to progress. Although I know I have made it once on the previous desktop install of ubuntu on this same machine. But I do not remember what I've done.
Thank you for any suggestion how to progress here.
Merci !
You can install the package "python-is-python3" to get around the python3 vs python issue.
For the libpulse error, you need Pulse audio. It's described here: https://github.com/HEnquist/camilladsp?tab=readme-ov-file#additional-dependencies
It should not fail like that if unar is missing, that's something I'll have to look at. I have not tested the script in Ubuntu.
I can't say why the GUI dir ended up empty. Can you run the script again and copy the entire terminal output?
For the libpulse error, you need Pulse audio. It's described here: https://github.com/HEnquist/camilladsp?tab=readme-ov-file#additional-dependencies
It should not fail like that if unar is missing, that's something I'll have to look at. I have not tested the script in Ubuntu.
I can't say why the GUI dir ended up empty. Can you run the script again and copy the entire terminal output?
Hi Wirruna, thank you for asking.
I want to use stuff that is lying around. I like the idea of creating value out of things which others dump. In short, ecological concern. I am aware that this is not fully reasonable. But, DIY audio is not about reason, right ?
Done, indeed it helps.
Attached the terminal output after having removed the camilladsp directory, and added the "python-is-python3" package. Next error is
"ERROR: Could not open requirements file: [Errno 2] No such file or directory: 'requirements.txt'"
So I then attempt to circumvent this issue by adding a hard path to an actual duplication of the "requirements.txt" file which I had found in "~/camilladsp/gui -D/camillagui/". That is, I replace the line "$VENV_PATH/bin/python -m pip install -r requirements.txt" by "$VENV_PATH/bin/python -m pip install -r /home/daniel/install-script/requirements.txt".
At this time the scripts claims it succeeds. But the directory "/home/daniel/camilladsp/gui" is empty.
Attached the second file "requirements_hard_path.txt".
(Note that I have not yet installed the package "libpulse-dev").
Last edited:
Thank you Henrik. I confirm that "sudo apt-get install libpulse-dev" has solved my "error while loading shared libraries: libpulse-simple.so.0" issue.
I kind of vaguely remember I have asked you this question once already Henrik, sorry. Kindly note that, as presented in camilladsp's github readme, "libpulse-dev" appears to be needed only when building camilladsp, and if choosing pulse-backend. Also noteworthy: I do not seem to have had to install any pulseaudio related package when installing the desktop version of the same ubuntu on the same machine.
This however does not explain why this package is not needed when installing camilladsp with the same ubuntu server 22.04 version on the RPI4 (thread of Michael). Package "linux-modules-extra-raspi" maybe ?
With libpulse-dev (sudo apt-get install libpulse-dev) as above, I now have strictly carried on Michael's rpi4-camilladsp-tutorial and I am happy to report that my AMD64 machine is now running Ubuntu server 22.04 with both camilladsp & camillagui as services and web gui available. I have not installed alsa loopback since I do not need it. Detailed steps in the history file attached - but actually I"m just following Michael's tutorial. Thank you Henrik & Michael for this cool technology that you're making available !
Last edited:
The pulse audio libraries are normally installed as a dependency for a graphical desktop, so they will usually not be installed by default in a terminal-only system.
The aarch64 binary is built without pulse audio support. The idea is that this version is likely to be used on a headless raspberry pi or such, without and graphical desktop installed. While the amd64 build is more likely to be running on a full desktop system. Not always true obviously, but seems to mostly work out. Maybe there should be separate versions with and without pulse for both architectures.
Thank you Henrik. All clear now. Pulsaudio is required on camilladsp's public amd64 build, not on its aarch64 build.
And the desktop version of e.g. Ubuntu comes with Pulseaudio activated, not its server version. Makes sense & matches my observations indeed. Merci !
And the desktop version of e.g. Ubuntu comes with Pulseaudio activated, not its server version. Makes sense & matches my observations indeed. Merci !
Also the header -dev libraries are not typically installed by default in debian-based distributions.
Ok there is a typo in the unar command. I'll fix asap.
I'm running into a problem with the new GUI backend.
I'm running CDSP on PiCorePlayer using LykkeDK's samplerate changer, but I don't think that's the problem, I think there's a deeper issue with the python install. When trying to start camillagui I get the following error:
From looking through the code it seems that "from jsonschema import Draft202012Validator" in settings.py is a new addition for the 2.1.0 release, and if I roll back to 2.0.1 then camillagui works again. I'm using the standard python3.8 package released for PiCorePlayer, but it looks like this doesn't support this particular validator. Here's what's present in my /usr/local/lib/python3.8/site-packages/jsonschema/init.py:
Is there a solution? Do I need to update the python installation somehow?
I'm running CDSP on PiCorePlayer using LykkeDK's samplerate changer, but I don't think that's the problem, I think there's a deeper issue with the python install. When trying to start camillagui I get the following error:
Code:
Traceback (most recent call last):
File "/mnt/mmcblk0p2/tce/Camilla_Extensions/camillagui/main.py", line 9, in <module>
from backend.routes import setup_routes, setup_static_routes
File "/mnt/mmcblk0p2/tce/Camilla_Extensions/camillagui/backend/routes.py", line 1, in <module>
from .settings import BASEPATH
File "/mnt/mmcblk0p2/tce/Camilla_Extensions/camillagui/backend/settings.py", line 7, in <module>
from jsonschema import Draft202012Validator
ImportError: cannot import name 'Draft202012Validator' from 'jsonschema' (/usr/local/lib/python3.8/site-packages/jsonschema/__init__.py)
Code:
"""
An implementation of JSON Schema for Python
The main functionality is provided by the validator classes for each of the
supported JSON Schema versions.
Most commonly, `validate` is the quickest way to simply validate a given
instance under a schema, and will create a validator for you.
"""
from jsonschema.exceptions import (
ErrorTree, FormatError, RefResolutionError, SchemaError, ValidationError
)
from jsonschema._format import (
FormatChecker,
draft3_format_checker,
draft4_format_checker,
draft6_format_checker,
draft7_format_checker,
)
from jsonschema._types import TypeChecker
from jsonschema.validators import (
Draft3Validator,
Draft4Validator,
Draft6Validator,
Draft7Validator,
RefResolver,
validate,
)
try:
from importlib import metadata
except ImportError: # for Python<3.8
import importlib_metadata as metadata
__version__ = metadata.version("jsonschema")Is there a solution? Do I need to update the python installation somehow?
Looks like you have a very old version of jsonschema. What version is the one in the pcp repos?
I wish PCP would upgrade to a more recent python version, 3.8 is getting quite old and will reach eol later this year.
I wish PCP would upgrade to a more recent python version, 3.8 is getting quite old and will reach eol later this year.
After doing some digging I see I'm using the jsonschema implementation from LykkeDK's package. From the filename I guess this is version 3.2.0. I found a git which provides up-to-date code which says it's compatible with python3.8, but (predictably) my simplistic attempt to graft this in failed ( "No module named 'attrs' " ). This is complicated by the fact that TinyCore uses its own idiosyncratic installation environment, so I've probably missed something important (you can't just 'pip install'). I've posted in Jesper's thread to see if he can give any help.