HI,
I have just tried to install the hifiberry dsp tools but the install hangs at this line
Building wheels for collected packages: lxml
Running setup.py bdist_wheel for lxml ... /
the install hang for about 30 minutes and I had to abort. Should I wait longer is this a problem with the script?
I installed it on a fresh install of raspbian (stretch) and I ensured that SPI was enabled. Any thoughts would be greatly appreciated.
Regards,
Rob Oldaker
Some libraries might be missing. Try
sudo apt-get install python3-lxml python-lxmlYou may need to install a few required packages first. You can do this with the following command:
sudo apt-get install libxml2-dev libxslt-dev python-dev
HI,
That worked after running those commands. Dsptoolkit has installed and it looks like the service is running OK as I can connect to my raspberry pi from SigmaStudio.
I am now trying to setup the dsp profile and get the ALSA volume control to work. I can write the EEPROM OK and the DSP program works.
I am confused about the instructions for saving the DSP profile (https://github.com/hifiberry/hifiberry-dsp/blob/master/doc/dspprofiles.md).
I can only manage to save the profile as a txt file (snippet below).
aram Address: 0xF050
Bytes: 2
Param Data:
0x1F, 0xFF
Block Write
Time: 9:38:52 - 835ms
IC: IC 1
Param Name: IC 1.POWER_ENABLE1 Register
Param Address: 0xF051
Bytes: 2
Param Data:
0x00, 0x1F
dsptoolkit will not load this file and the documentation seems to suggest it should be an XML file - how do I save it as XML? The right-click options are the same as the documentation having "Save as Text .." and "Save as Raw Data" - neither of these appear to save as XML.
I am using SigmaStudio 3.17
Thanks for your time,
Rob.
Hi,
you first have to write to the EEPROM and save the sequence to an XML file. Have a look at the section "Export the DSP profile".
Best regards,
Daniel
Hi Daniel,
I have worked this out now ( had to use the "Add sequence" command from the right-click menu).
I have created the .xml file and loaded into dsptoolkit without errors.
When I use command "dsptoolkit get-volume" I get the error message
"Profile doesn't support volume control"
Is it looking for a specific parameter name? I am just using the "2-way-eq" SigmaStudio project I downloaded a year or so ago.
Thanks,
Rob.
Hi,
the XML file is only the first step. With this, you can write the program to the DSP, but DSPToolkit don't know any parameters. For this it needs metadata. You can add these by hand or export the parameters from SigmaStudio and merge these semi-automatically into the XML file. Check the section "metadata" in the documentation. If you only need volume control, the fastest way might be adding it by hand.
Best regards,
Daniel
Hi Daniel,
I tried exporting the parameters and merging them in as per the documentation but it said "No parameters to merge".
I then tried adding this entry by hand to the profile and re-installing it:-
<beometa>
<metadata type="volumeControlRegister">118</metadata>
</beometa>
I chose 118 as the address since this was the address of the register that changed when changing the volume in SigmaStudio but it still says "Profile does not support volume control".
Rob.
Hi Rob,
you need to re-install the profile first. Just editing the file won't work as DSPToolkit won't notice the change.
Best regards,
Daniel
Hi Daniel,
I have re-installed the profile it but am still getting the error message "Profile does not support volume control" when using the command "dsptoolkit get-volume".
Rob.
Hi Rob,
check if the <beometa> section is WITHIN the <ROM> section in the XML file. If this is ok, please post the XML somewhere where I can have a look at it.
Best regards,
Daniel
Hi Daniel,
Here is link to it:-
https://drive.google.com/file/d/18ShBODjDu7Szj8lVsmbrNnwbr1I-6qzt/view?usp=sharing
Rob.
Looks like you found a bug. It seems, dsptoolkit has problems with files that have only one metadata record. I'm not sure, why this happens, but the workaround is to add another dummy metadata record like
<metadata type="x">111</metadata>
Best regards
Daniel
Hi Daniel,
That worked!
I can now set the volume using dsptoolkit which is good. I would also like to set the volume using ALSA. When I type
"amixer controls" I get the following which is encouraging
numid=1,iface=MIXER,name='DSPVolume'
but when I try to set the volume with:-
"amixer cset 1 50"
I am getting the error:-
Wrong control identifier: 1
I tried restarting the sigmatcp service and the output is shown below. Unfortunately it didn't make any difference.
pi@beo-create:~ $ sudo systemctl status sigmatcp.service
● sigmatcp.service - SigmaTCP Server for HiFiBerry DSP
Loaded: loaded (/lib/systemd/system/sigmatcp.service; enabled; vendor preset: enabled)
Active: active (running) since Thu 2019-01-10 15:52:29 GMT; 3min 14s ago
Main PID: 5070 (sigmatcpserver)
CGroup: /system.slice/sigmatcp.service
└─5070 /usr/bin/python3 /usr/local/bin/sigmatcpserver --alsa
Jan 10 15:52:29 beo-create systemd[1]: Started SigmaTCP Server for HiFiBerry DSP.
Jan 10 15:52:30 beo-create sigmatcpserver[5070]: INFO:root:running as root, data will be stored in /var/lib/hifiberry
Jan 10 15:52:30 beo-create sigmatcpserver[5070]: INFO:root:initializiong ALSA mixer control
Jan 10 15:52:30 beo-create sigmatcpserver[5070]: INFO:root:reading profile /var/lib/hifiberry/dspprogram.xml
Jan 10 15:52:30 beo-create sigmatcpserver[5070]: INFO:root:caching program memory checksum
Jan 10 15:52:30 beo-create sigmatcpserver[5070]: INFO:root:checksum memory: b'\xc5W3\xcew\x08\x08\x91t\xce\\R\xf6^V\x13'
Jan 10 15:52:30 beo-create sigmatcpserver[5070]: INFO:root:DSP profile doesn't have a checksum, might be different from
Jan 10 15:52:30 beo-create sigmatcpserver[5070]: INFO:root:Announcing via zeroconf
Jan 10 15:52:31 beo-create sigmatcpserver[5070]: INFO:root:Starting TCP server
lines 1-16/16 (END)
Never tried the "cset" command.
Try somthing like
amixer sset DSPVolume 60%
Best regards,
Daniel
That worked!
Thanks for all of your help.
Regards,
Rob.
You're welcome. Are you doing some interesting stuff with your DSP program?
Best regards,
Daniel
Hi Daniel,
I am not really doing a lot with the DSP. I am developing audio software (called MyFiMusic) and am currently testing it on the different HiFi Berry boards I have (DAC +, amp and BeoCreate) - I wanted to get ALSA volume to work on the BeoCreate so I have a standard interface to the volume across all boards.
I am very impressed with the sound quality of the BeoCreate out of the box (using the "2-way-eq" project) - I am driving a pair of unmodified BeoVox cx100 speakers with just with 60watt amps. Its significantly better than the analogue NAD amp I was using before and also uses a lot less power. When I get time I would like to see what they sound like with the DSP software optimsed for them and powering each driver separately from each amp.
My software is capable of setting up the raspberry pi for the DAC+ and AMP software without little intervention from the users (the user simply selects the board and the software sets up the correct "dtoverlay" flag for them) . I don't need to tell you that the BeoCreate is a bit more complicated as it involves installing SigmaStudio, burning the EEProm and setting up the dsptoolkit which will all have to be done manually by the user! The user is rewarded with a fantastic sounding amp at the end of it, however.
Thanks again for your help,
Rob.
Hi Rob,
if you don't do any specific processing on the DSP, it might be easier just to use one of the standard profiles. Dealing with SigmaStudio is really only needed if you want to implement specific processing that isn't supplied by our default profiles.
Is your software already available?
Best regards,
Daniel
Hi Daniel,
I didn't realise you supplied standard profiles (I assume the board will come pre-programmed with a default DSP program?) - do you have a link with more info so I can try them out? I thought if a user wanted to use third party software they had burn the EEprom etc.
The software is available - there is also a website but it needs updating (www.myfimusic.co.uk)
Install instructions:-
1. download installer from website
curl http://myfimusic.co.uk/InstallMyFiMusicLite.sh > InstallMyFiMusicLite.sh
2. run the installer
bash InstallMyFiMusicLite.sh
2. Use raspi-config and ensure user pi is logged in by default.
3. reboot
Access MyFiMusic from a browser at "http://<ip address of rPi>:8080"
Setup the sound card, wifi hotspot etc. by going to settings page. Ensure it is setup as a "server & player". You can add further channels to give access to Spotify, podcasts, radio stations, BBC iPlayer etc.
I also have an android app that is currently under test and I would be happy to let you try out (it is a wrapper around the standard android web browser but also allows uses notifications and the volume keys to provide easier access).
Regards,
Rob.
Hi Rob,
your can find some default profiles here:
https://github.com/hifiberry/hifiberry-dsp/tree/master/sample_files/xml
You don't even have to download them, dsptoolkit can do this, e.g.
dsptoolkit install-profile https://raw.githubusercontent.com/hifiberry/hifiberry-dsp/master/sample_files/xml/4way-inputswitcher.xml
Best regards
Daniel
The latest default profiles are based on this one:
https://github.com/hifiberry/hifiberry-dsp/blob/master/sample_files/dspproj/4way-iir-delay-inputswitcher.dspproj
It has a bit more features, but this won't have an impact on sound quality.
Best regards,
Daniel
Thanks for this. I am still right in thinking that users will still need to burn the EEProm using SigmaStudio - or will the boards come with this DSP program as the default ( i.e. pre-burnt on to the EEProm).
Rob.
Hi,
the boards come with the default profile installed. But even if you want to re-install it or install another profile, you don't need SigmaStudio. You can install all the provided DSPProfile XML files just with the dsptoolkit command line utility.
SigmaStudio is only needed if you want to implement some special features that aren't available in an exiting profile.
Best regards,
Daniel
Hi Daniel.
Its a bit clearer now! The DSP profile contains the DSP program which is something I didn't realise before.
I have just tried to install the DSP profile you mentioned earlier and it all works fine.
When installed on a new system will dsptoolkit know that the default profile has been installed initially or will it always need to install a profile before ALSA volume gets sets up?
Rob.
Hi,
you need to (re-)install a profile before dsptoolkit can handle ALSA volume. Otherwise dsptoolkit won't know the profile that is installed on the DSP.
Best regards,
Daniel
Hi Daniel,
Thanks for the clarification. I have just been listening to the new profile (4way-iir-delay-inputswitcher) and noticed it has some audible drop-outs/clicks that happen every so often (2 a minute or so). They are not there after reverting to the original profile (based on "2-way-eq").
Regards,
Rob.
Hi,
the new profile has an auto-mute function that mutes the output if the input level is very low. It also mutes if the player stops playback. So this might happen during track changes.
While it can be tunes a bit, the input-switcher profile doesn't support this at the moment.
Best regards,
Daniel