audacity_rescue.sh

#!/bin/bash
#Recover from an Audacity crash by reassembling the saved data files


if [ $# != 2 ]; then
    echo "ERROR: Not enough arguments"
    echo "$0 [SOURCE PATH] [DESTINATION]"
    exit 1
fi

DATA="$1/"
OUT=$2

if [ -e $OUT ]; then
    # Hope the path exists
    echo "ERROR: Output file exists ($OUT)"
    exit 1
fi

COUNT=1

find $DATA -name "*.au" -print0 | while read -d $'\0' FILE ; do
	#The offsets are probably all going to be the same, but best check it
	OFFSET=$(echo $(od -i -j4 -N4 -An < $FILE) ) # Use echo for easy trim
	if [ $COUNT -eq 1 ]; then
		# Write the header
		dd ibs=$OFFSET count=1 if=$FILE of=$OUT
	fi
	 echo "Adding $FILE (offset=$OFFSET)"
	 dd ibs=$OFFSET skip=1 conv=notrunc oflag=append if=$FILE of=$OUT
	 
	 let COUNT+=1
done

echo "Done"

Thanks for the script. Modified to use BleuLlama since I use a mac. Also used if (( ($COUNT % 2) == 1 )); then because I needed to recover a stereo recording and pairs of audio clips represents left and right pair. I only needed to recover at least one half of the channel.

#!/bin/bash
# Recover from an Audacity crash by reassembling the saved data files
# Modified for MacOS and skipping every second file due to stereo recording

if [ $# != 2 ]; then
		echo "ERROR: Not enough arguments"
		echo "$0 [SOURCE PATH] [DESTINATION]"
		exit 1
fi

DATA="$1/"
OUT=$2

if [ -e $OUT ]; then
		# Hope the path exists
		echo "ERROR: Output file exists ($OUT)"
		exit 1
fi

COUNT=1

find $DATA -name "*.au" -print0 | while read -d $'\0' FILE ; do
	#The offsets are probably all going to be the same, but best check it
	OFFSET=$(echo $(od -i -j4 -N4 -An < $FILE) ) # Use echo for easy trim
	if [ $COUNT -eq 1 ]; then
		# Write the header
		dd ibs=$OFFSET count=1 if=$FILE of=$OUT
	fi
	if (( ($COUNT % 2) == 1 )); then
		echo "Adding $FILE (offset=$OFFSET) $COUNT"
		rm -f OOO
		dd ibs=$OFFSET skip=1 conv=notrunc if=$FILE of=OOO
		cat OOO >> $OUT
	fi
	let COUNT+=1
done

echo "Done"

Hi all! Thanks for your work here. Do you know if this would work on a project with multiple tracks?

Suggestion: in cases where the files are not named in sequence according to time of creation, you can replace this line:

find $DATA -name "*.au" -print0 | while read -d $'\0' FILE ; do

with this:

find $DATA -name "*.au" -printf "%T@ %Tc %p\n" | sort -n | awk '{print $NF}' | while read -d $'\n' FILE ; do