Conquer Giant APE and Giant FLAC

I have already written posts on splitting and converting from either album-length APE files or album-length FLAC files into track-length FLAC files.

(My post on converting APE files; my post on splitting APE files; my post on splitting FLAC files. You will want to go over these other articles if you have not done any of this before as there are some dependencies which you will not likely have installed.)

That was rather satisfying, but I then found that I would locate entire discographies where every album was a single APE or FLAC.  I did not enjoy the prospect of running each command on each pair of files.

That’s why scripting languages exist.  It’s a good thing I bought that BASH book.

Anyway, this script is designed to recursively locate APE, FLAC, and CUE files under a user-supplied root directory; then it runs shntool and cuetag on each APE/FLAC and CUE pair.  (You can read my previous posts linked above for more details about how these two commands work.)

It even cleans up after itself.  Since it deletes not only the temporary files it creates but also all the APE, FLAC, and CUE files it uses, you will want to pay close attention when the script tells you it is about to clean up all those files.  If you tell it to proceed (by hitting ENTER) without checking its work all those files will be deleted.  As such you should work from copies and you should check that it did what you wanted it to do before hitting ENTER when it asks about cleaning up all files.

Feel free to suggest any improvements.  I hope you find that it works well and improves your lives.

One important thing to note is that it is not an intelligent script.  It cannot determine for you whether a given APE or FLAC file is actually album length.  It assumes that you are offering it a folder (with or without sub-folders) which contains only album-length APE or album-length FLAC files with an accompanying CUE file (one matching pair per containing folder).

Sometimes a cue file will be formatted incorrectly or have some other issue.  Sometimes they can be fixed by opening them in a text editor.  Probably they can always be fixed in a text editor, but it might be difficult to determine what’s wrong with the CUE file.  If a CUE files has issues you will see errors on the command line (if you are staring at your monitor) and the album-length file will not be split/tagged.  You’ll have to work on that pair separately, but since you are working from copies it won’t matter if they get deleted.

If you throw something unusual at it, it’s hard to imagine how it might behave.  It’s likely not dangerous but you should always work from copies and preserve the originals until you are satisfied.

# covert APE/FLAC albums to FLAC tracks recursively
# by JamesIsIn from
# Do something nice today.

## gather information

# obtain directory in which to work

printf "\nHello.\n\n"
read -e -p "Please provide the root folder under which I will work recursively: "
if [ -d "$REPLY" ]; then
printf "\nI have confirmed this is a directory.\n\n"
printf "\nI cannot confirm this is a directory.\n\n"
echo "$REPLY"
exit 1

# number, performer, and track names

printf "There are two formatting options for file naming (assuming the cue contains correct information):\n\n"
printf "01 – Track Name.flac\n"
printf "01 – Performer Name – Track Name.flac.\n\n"
read -p "Do you want to include the performer name in the file names (this is only recommended for multiple-artist albums)? [Y or N]: "
if [ "$REPLY" = "Y" ]; then
printf "\nOk, I'll include performer names: 01 – Performer Name – Track Name.flac\n\n"
trackname="%n – %p – %t"
elif [ "$REPLY" = "y" ]; then
printf "\nOk, I'll include performer names: 01 – Performer Name – Track Name.flac\n\n"
trackname="%n – %p – %t"
else [ "$REPLY" = "[Nn]" ];
printf "\nOk, I won't include performer names: 01 – Track Name.flac\n\n"
trackname="%n – %t"

printf "\n\nLet's get to work.\n\n\n"

## do the work

# parse files into respective arrays recursively

find "$directory" -type f -name \*.[Aa][Pp][Ee] -o -name \*.[Ff][Ll][Aa][Cc] > /tmp/whyme

declare -a albumfind
let i=0
while read albumline; do
done < /tmp/whyme

find "$directory" -type f -name \*.[Cc][Uu][Ee] > /tmp/whyme2

declare -a cuefind
let i=0
while read cueline; do
done < /tmp/whyme2

# follow array counts

printf "\n\nCounts are as follows. There are "${#albumfind[@]}" albums and "${#cuefind[@]}" cues. Your albums ought to equal your cues. If these counts are not as expected, please inspect the contents of the folder you provided before proceding or re-run the script and select a different folder.\n\n"
read -p "Press <ENTER> to coninue. "

# split apes and flacs; hide album files

for (( i=0 ; i < ${#albumfind[@]} ; i++ )) ; do
# path is cue iteration less file name
shnsplit -d "$albumfolder" -o flac -f "${cuefind[i]}" -t "$trackname" "${albumfind[i]}" ;
mv "${albumfind[i]}" "${albumfind[i]}".hideme ;

# remove 00 files

find "$directory" -type f -name 00\*.[Aa][Pp][Ee] -o -name 00\*.[Ff][Ll][Aa][Cc] > /tmp/whyme3

declare -a deadfiles
let i=0
while read deadline; do
done < /tmp/whyme3

for (( i=0 ; i < ${#deadfiles[@]} ; i++ )) ; do
mv "${deadfiles[i]}" "${deadfiles[i]}".hideme ;

# tag files

for (( i=0 ; i < ${#cuefind[@]} ; i++ )) ; do
# path is cue iteration less file name
cuetag "${cuefind[i]}" "$albumfolder"\/*.flac
mv "${cuefind[i]}" "${cuefind[i]}".hideme ;

## clean-up

printf "\n\nI have completed all operations to the best of my ability.\nI will now clean up all files.\n\nPlease ensure you are satisfied with the state of things before you continue."
printf "\n\nPlease note: I am about to delete all the album-length files I believe I have used."
read -p "Press <ENTER> to coninue (Ctrl-c will abort clean-up). "

# remove temp files, hideme files, and unset all variables

find "$directory" -type f -name *.hideme > /tmp/whyme4

declare -a hidemes
let i=0
while read hidelines; do
done < /tmp/whyme4

for (( i=0 ; i < ${#hidemes[@]} ; i++ )) ; do
rm "${hidemes[i]}" ;

rm /tmp/whyme*



Have fun with that.


7 thoughts on “Conquer Giant APE and Giant FLAC

    1. I trust you worked from copies as advised. If you will post some of your output I might be able to tell you what’s wrong with your cue/flac combintations. Often it’s a simple syntax problem in one or more of the files.

  1. Hey, James, thanks for your reply to my blog at linuxquestions :). Thanks for the equation ( ff –> nnn) , too, such things are difficult to find when you happen to need them right away.

    I also analysed and answered your post and the answer may be of interest for you, as you’re trying to create an automated script.
    It’s all about the possible difference in formats used by, for example, WVPACK (by which many rips are created) and shntools, which we then use to split the *.wv rip created in WvPack and using the CUE sheets created by the same WvPack.
    Regarding your script, then, with a *.wv file you can, in all logic, assume its being created by WvPack — hence, the CUE format must be expected to be in “”, as per WvPack documentation, and not “hh:mm:ss.ff”, as shntools will assume.
    Sure, this brilliant guess needs being tested in real life, which I can’t do right now, sorry :(. As I was satisfied by my results, the possible inaccuracy thus created may have been quite trifling for human ear. Yet, for the sake of purity and mathematical exactitude it must be tested and verified, right?

    Thanks for the equation anyway, and for the script!! Really I had hard times trying to find any references online, except every now and then “good advise” to install certain stuff in MS Windows etc…

    1. The CD format (which is what shntools currently expects) is hh:mm:ss:ff (colon not dot). That makes it a little easier to identify (and hopefully to script).

      I hope to write a script to just convert cue files into the other format (change the colon to a dot and change the frames to thousandths of a second). I’m just not very good with sed and so am working very slowly to write the code to parse and convert all the correct time stamps in a cue file.

      Once that is done, the script above will work fine for them as well.

Leave a Reply

Your email address will not be published. Required fields are marked *