\documentclass{report} \author{xengineering} \title{soundbox documentation} \usepackage{graphicx} \graphicspath{ {./diagrams/} } \usepackage{hyperref} \usepackage{parskip} \begin{document} \maketitle \newpage \tableofcontents \newpage \listoffigures \newpage \chapter{Introduction} \label{chap:introduction} \texttt{soundbox} is a device to connect classic audio systems to the network. \section{Versioning} This device is versioned with Semantic Versioning\footnote{\href{https://semver.org}{https://semver.org}}. The resulting version numbers have the format \texttt{..} like e.g. \texttt{2.0.3}. While Semantic Versioning is usually targeted at software only it is here used for the whole device. This includes mechanical, electronic and software aspects as shown in figure \ref{diagram:device-semver}. \begin{figure}[h] \centering \includegraphics[width=\textwidth]{device-semver.pdf} \caption{API for a device versioned with Semantic Versioning} \label{diagram:device-semver} \end{figure} This versioning makes sure that users do not have to care about the device internals at all. Devices can be seen as atomic from a user's perspective. This level of granularity is choosen because users are not expected to disassemble devices or to care about the software internals. %\section{Licensing} \chapter{User guide} In addition to \autoref{chap:introduction} `\nameref{chap:introduction}` the following sections document everything required to use \texttt{soundbox} devices. %\section{Device setup} \section{Playing audio} \label{playing-audio} For a simple audio streaming test the following commands from the FFmpeg streaming guide\footnote{\href{https://trac.ffmpeg.org/wiki/StreamingGuide\#StreamingasimpleRTPaudiostreamfromFFmpeg}{https://trac.ffmpeg.org/wiki/StreamingGuide\#StreamingasimpleRTPaudiostreamfromFFmpeg}} can be used. The following command has to be executed on the soundbox to start listening on port \texttt{5316} of a local multicast IPv4 address for incoming audio streams. Since multicast is used, this can be executed on multiple soundbox devices to stream to all of them. \begin{verbatim} ffplay udp://224.0.0.99:5316 \end{verbatim} As soon as the soundbox is listening the stream can be send from another computer in the same network with this command: \begin{verbatim} ffmpeg -re -f lavfi -i aevalsrc="sin(400*2*PI*t)" -ar 8000 -f mulaw \ -f matroska udp://224.0.0.99:5316 \end{verbatim} It will send a simple sine wave with 400~Hz to the soundbox device. Alternatively a MP3 file can be sent like this: \begin{verbatim} ffmpeg -re -i -acodec copy -ar 11025 \ -f matroska udp://224.0.0.99:5316 \end{verbatim} These commands allow to test the \texttt{soundbox} device setup. \chapter{Production} The given chapter contains documentation about how to produce a \texttt{soundbox} device. %\section{Printing mechanical parts} \section{Alpine Linux installation on a Raspberry Pi} Alpine Linux for the Raspberry Pi can be downloaded from the Alpine Linux\footnote{\href{https://alpinelinux.org/}{https://alpinelinux.org/}} download page. This section describes the installation procedure for the variant targeting the AArch64 architecture provided as a compressed tar archive. It can be downloaded and verified against a checksum with the \texttt{wget} and \texttt{sha256sum} utilities. \begin{verbatim} wget https://dl-cdn.alpinelinux.org/alpine/v3.19/releases\ /aarch64/alpine-rpi-3.19.0-aarch64.tar.gz wget https://dl-cdn.alpinelinux.org/alpine/v3.19/releases\ /aarch64/alpine-rpi-3.19.0-aarch64.tar.gz.sha256 sha256sum -c alpine-rpi-3.19.0-aarch64.tar.gz.sha256 \end{verbatim} The image has to be flashed to a SD card which will be inserted into the Raspberry Pi. This SD card has to be put into a Linux PC first. In Linux every SD card is represented as a block device like \texttt{sda} or \texttt{sdb} with a file path like \texttt{/dev/sda}. Possibly existing partitions on the SD card are represented with that path and a number as suffix like in \texttt{/dev/sda1}. The program \texttt{lsblk} gives an overview of the currently connected block devices. The correct device name for the SD card can be identified by its size. \begin{verbatim} $ lsblk NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINTS sda 8:0 1 59,5G 0 disk `-sda1 8:1 1 59,5G 0 part \end{verbatim} The SD card has to be formatted with a single bootable FAT32 partition. This can be achieved with the command line program \texttt{parted}. First a \texttt{msdos} partition table has to be created followed by a partition taking the full size of the SD card. The boot flag of the partition is enabled and a FAT32 file system is created with \texttt{mkfs.vfat}. All of these operations require \texttt{root} privileges. \textbf{Warning:} These operations might cause data loss if the wrong device is specified. It has to be asserted that not the wrong block device (e.g. the PC disk) is chosen. \begin{verbatim} parted /dev/ mklabel msdos parted /dev/ mkpart primary fat32 0% 100% parted /dev/ toggle 1 boot mkfs.vfat /dev/1 \end{verbatim} After the file system is created the downloaded Alpine Linux tar archive has to be extracted to it. A temporary mount point can be created with \texttt{mktemp}. After mounting with \texttt{mount} the archive can be extracted with \texttt{tar}. Finally the file system is unmounted and it is waited for all buffered write operations to the SD card with \texttt{sync}. Also these command require \texttt{root} privileges. \begin{verbatim} mountpoint=$(mktemp -d) mount /dev/1 "$mountpoint" tar -C "$mountpoint" -xf alpine-rpi-3.19.0-aarch64.tar.gz umount /dev/1 sync \end{verbatim} Afterwards the SD card can be inserted into the Raspberry Pi. To do initial configuration it should be connected to a screen via HDMI. Also a keyboard has to be connected via USB. A mouse is not required. Finally the Pi should be powered up by connecting a USB power supply to the corresponding USB power input of the Raspberry Pi. Text from the boot process should be displayed on the screen. After logging in with the username \texttt{root} and no password basic system configuration on Alpine Linux can be done interactively with \texttt{setup-alpine}. The Alpine Linux installation guide\footnote{\href{https://wiki.alpinelinux.org/wiki/Installation}{https://wiki.alpinelinux.org/wiki/Installation}} contains further details. When a diskless installation is chosen the configuration changes have to be made persistent with the \texttt{lbu} command. Otherwise they will be lost on the next reboot. \begin{verbatim} setup-alpine lbu commit -d \end{verbatim} With this setup the Alpine Linux installation is completed. \section{Configure Alpine Linux for soundbox} soundbox requires packages from the Alpine Linux community package repository. Since this is not enabled by default it has to be done manually. This can be done by editing the \texttt{/etc/apk/repositories} file with an editor like \texttt{vi}. The hash sign \texttt{\#} in front of the URL to the community repository has to be removed. Instead also \texttt{sed} can remove this sign on a fresh installation quite easily since the community repository is the only one with such a sign in front of it. \begin{verbatim} sed -i 's/^#//g' /etc/apk/repositories \end{verbatim} With an enabled community repository all required packages for soundbox can be installed. The list with required packages can be written to the \texttt{/etc/apk/world} file with \texttt{echo}. Running \texttt{apk add} after it will install all of these packages including their dependencies and removes everything else. This makes sure that always exactly the same set of packages is used for a soundbox device. \begin{verbatim} echo 'alpine-base alsa-utils alsaconf busybox-mdev-openrc chrony ffplay iw openssh openssl wpa_supplicant' > /etc/apk/world apk add \end{verbatim} The only user on a soundbox device is \texttt{root}. Even that super user has to be added to the \texttt{audio} group to be able to play sound via ALSA. This can be achieved with \texttt{addgroup}. \begin{verbatim} addgroup root audio \end{verbatim} soundbox uses the HiFiBerry shield to output high-quality audio. This shield needs to be enabled as an audio output by specifying a Devicetree overlay in the Raspberry Pi configuration file with \texttt{echo}. Since this file is stored on the SD card only which is mounted as read-only it has to be remounted with the read-write option first. \begin{verbatim} mount -o remount,rw /media/mmcblk0p1 echo 'dtoverlay=hifiberry-dac' >> /media/mmcblk0p1/config.txt \end{verbatim} If the builtin audio of e.g. a Raspberry Pi 4 should be used another Devicetree option has to be inserted instead: \begin{verbatim} echo 'dtparam=audio=on' >> /media/mmcblk0p1/config.txt \end{verbatim} Since a diskless Alpine Linux installation is preferred for soundbox the changes have to be made persistent with \texttt{lbu}. Rebooting after it makes sure the SD card is again mounted in read-only mode and all Kernel-affecting changes can be applied. \begin{verbatim} lbu commit -d reboot \end{verbatim} After this reboot the soundbox device should be able to play audio like described in section~\ref{playing-audio}. %\section{Final assembly} %\chapter{Device internals} %\section{Mechanical design} %\section{Electronics} %\section{Operating system} %\section{Software} \end{document}