Das Werkstatt

Posted: **Fri Nov 11, 2011 9:19 pm**

[PROBLEM]
The Austrian radio channel "Ö1" sends confusing title metadata within their live streams.

[SOLUTION]
After reading the source code of their streaming website, it was clear that the URL http://oe1.orf.at/programm/konsole/live contains the currently running title (and a few of the subsequent ones), encoded in JSON format.

One can now simply use plain-vanilla PHP, and transform the JSON data into something that, for example, streamripper can read:

Code: Select all

#!/usr/bin/php
<?php
$PAUSE=3;           // Wait n seconds before refreshing the metadata
$METADATA_URL="http://oe1.orf.at/programm/konsole/live";

while(true)
{
    // Fetch stream metadata in JSON format from OE1 website:
    $json_string = file_get_contents($METADATA_URL);

    // Parse JSON data into PHP object:
    $json_data = json_decode($json_string);
    $current_track = $json_data->list[0];

    printf("TITLE=(%s) %s\nARTIST=OE1 [ID:%d]\n.\n",
        trim($current_track->time),
        $current_track->title,
        $current_track->id
    );

    sleep($PAUSE);
}
?>

NOTE: It's important that the script continuously refreshes the metadata itself - therefore the infinite "while" loop, and the $PAUSE sleep.

If you save the PHP code above in a file called "metadata-oe1.php", you can call streamripper like this:

Code: Select all

streamripper http://mp3stream3.apasf.apa.at:8000/listen.pls -d /srv/streamripper/OE1 -E /path/to/metadata-oe1.php

Posted: **Sun Jun 03, 2012 2:09 pm**

Here is a short BASH script used for recording different streams:

Code: Select all

#!/bin/bash

# ---------------
STREAM_OE1="http://mp3stream3.apasf.apa.at:8000/listen.pls"
METADATA_OE1="/path/to/metadata-oe1.php"

STREAM_BLA="http://wherever.bla/stream.m3u"
# ---------------

DIR_OUTPUT="/srv/streamripper"

function start_rip
{
    local SOURCE="$1"
    local METADATA=$(eval "echo \$METADATA_$SOURCE")
    local STREAM_URL=$(eval "echo \$STREAM_$SOURCE")
    local DIR_OUTPUT="$DIR_OUTPUT/$SOURCE"

    if [ -z "$STREAM_URL" ]; then
        echo ""
        echo "No URL defined for source '$SOURCE'. Exiting."
        echo ""
        return
    fi

    echo ""
    echo "Recording '$SOURCE'..."
    echo " - URL: $STREAM_URL"

    local CMD="streamripper $STREAM_URL -s -d \"$DIR_OUTPUT\""
    if [ -n "$METADATA" ]; then
        echo " - Metadata: $METADATA"
        CMD="$CMD -E $METADATA"
    fi
    echo "$CMD"

    read -p "Press any key to continue..."
    eval "$CMD"
}

start_rip "$1"

This script can be called with different "callsigns" of streams (example: "OE1", "BLA") and will create subfolders with these callsigns in "/srv/streamripper". Metadata handling scripts can optionally be provided per-stream.

Posted: **Wed Jan 08, 2014 5:08 pm**

After upgrading to Ubuntu 12.04.3 LTS, I had weird problems with my metadata-fetching code:

PHP Warning: file_get_contents(http://oe1.orf.at/programm/konsole/live): failed to open stream: Connection timed out in metadata-oe1.php on line 9
PHP Notice: Trying to get property of non-object in metadata-oe1.php on line 15

Seems like something (either on my side or theirs) changed which causes this connection timeout.

I've now added checks to avoid accessing the $json_data variable unless it's been set:

Code: Select all

#!/usr/bin/php
<?php
$PAUSE=3;           // Wait n seconds before refreshing the metadata
$METADATA_URL="http://oe1.orf.at/programm/konsole/live";

while(true)
{
    // Fetch stream metadata in JSON format from OE1 website:
    $json_string = file_get_contents($METADATA_URL);

    if (isset($json_string) && !empty($json_string))
    {
echo "JSON string: ".$json_string; //DELME

        // Parse JSON data into PHP object:
        $json_data = json_decode($json_string);
        if (!empty($json_data))
        {  
            $current_track = $json_data->list[0];

            printf("TITLE=(%s) %s\nARTIST=OE1 [ID:%d]\n.\n",
                trim($current_track->time),
                $current_track->title,
                $current_track->id
            );
        }
    }

    sleep($PAUSE);
}
?>

Posted: **Wed Jan 08, 2014 5:11 pm**

I've also encountered character encoding issues in the files produced by streamripper.
Luckily, streamripper has commandline arguments which allow to define the encoding to be used for metadata:

--codeset-filesys: Tells streamripper what codeset to use for the file names when it writes to your hard drive.
--codeset-id3: Tells streamripper what codeset to use for the id3 information.
--codeset-metadata: Tells streamripper what codeset is being used for metadata in the stream coming from the network.

I've set all of them now to "UTF-8" and now my filenames with umlauts are perfectly fine

Thanks to a thread about streamripper encoding on ubuntuusers.de

Das Werkstatt

Fetch and prepare ö1 Metadata for streamripper

Fetch and prepare ö1 Metadata for streamripper

Clean script for stream recording

Connection timed out...

Metadata encoding