Page 1 of 1

Converting a library of audiobooks for blind people

Posted: Sun Mar 04, 2018 3:28 pm
by kbaerwald
Hi

I try to setup a reasonable process for converting a library of audiobooks for blind people to a condensed mp3-based library. The typical audiobook contains 10-15 single mp3 files previously converted from the original analog Compact Cassettes. My idea is to convert all mp3 files to wav, merge the wav files to one wav file containing the whole audiobook (with cue marks) and load it into Goldwave for further processing. This is to trim the technical (spoken) information at the beginning and end of that file, remove the worst analog noise and save the resulting file as flac for further tagging and processing.

The first problem I ran into was the size of the merged wav file - up to 6GB (16bit data). I can merge into such a file with Goldwave merge function but it does not load due to time length problems. If I tell the merger to save a flac file instead it takes nearly 20 min for a typical audiobook - much too long. I have a Core i7 4720HQ system with 1TB SSD and 16GB memory and expected this step to be much faster. Furthermore - after the merge Goldwave does not load the generated flac file into the workspace automatically: it just merges and then I have to load the merged file manually which again takes time as Goldwave has to decode the flac again.

Before I cannot solve these problems I will not continue with my other tasks.

Does anybody here has a good idea how to avoid these obstacles? Maybe they are selfmade :( ? Should I change my work flow? Thanks for any hints

Klaus

Re: Converting a library of audiobooks for blind people

Posted: Sun Mar 04, 2018 6:14 pm
by DougDbug
The first problem I ran into was the size of the merged wav file - up to 6GB (16bit data).
WAV files are limited to 2GB or 4GB depending on the software you're using.*

Since you have a solid state drive I'm not sure what you can do to speed things up. Sometimes it helps if you use a different drive for temporary storage so you can read from one drive while writing to another. (With regular hard drives the read/write head doesn't have to jump around between reading & writing locations.)

...Usually you spend more time editing/listening and you might just have to live with a 20 minute coffee break once in awhile.

You may be able to automate some of this with Batch Processing, The Effect Chain Editor, and the File Merger. But the more you automate the more waiting you'll have to do... (I can't help you because I'm always doing something slightly-different and it's not worthwhile for me to "program" or "script' anything.)








* There is a 32-bit file-size field in the file header which limits you to 2GB for a signed integer or 4GB for an unsigned integer. I believe the original WAV spec said 2GB maximum, but I don't know why since the size can't be negative.

Re: Converting a library of audiobooks for blind people

Posted: Mon Mar 05, 2018 2:45 pm
by BillK
I am really perplexed as th WHY you would want to convert MP3 VOICE files to WAV. ?!?!? You did say Audio Books, correct? Mostly Voice? Voice files don't need that much "fidelity". I have taken audio CDs and converted them to MP3s using GoldWave many times. I do that to place all the CDs onto ONE CD, so my wife can listen to books in her car - WITHOUT having to fumble changing the CDs out periodically. I have placed as many as 35 CDs onto one CD by reducing the storage down to 48 Kbps. YOU may not wish to go that low, but most people will not be able to tell the difference.

Re: Converting a library of audiobooks for blind people

Posted: Mon Mar 05, 2018 4:04 pm
by DougDbug
I am really perplexed as th WHY you would want to convert MP3 VOICE files to WAV. ?!?!?
From what I understand the main goal is editing & noise reduction and WAV is just part of the processing sequence (since the MP3 will have to be decompressed anyway).

In the end, the files were going to be FLACs. (Although, it would probably be nice to have MP3s also.)

kbaerwald,
mp3DirectCut can do basic editing on MP3s, and it can go a LOT faster because you are working on smaller files and you don't have to wait for decompression/recompression. However, it cannot do noise reduction or EQ or anything like that.

Re: Converting a library of audiobooks for blind people

Posted: Tue Mar 06, 2018 10:31 am
by kbaerwald
Thanks for all the answers!

Indeed - wav would have been my working (intermediate) format - I will go with flac now though compression/decompression takes some time. The main issue is to remove noise (audio cassette), cut off beginning and end (they contain technical information to handle the cassette for a blind person) and to align loudness of 10-15 recordings per audiobook with various quality so that they all can be merged into one flac file for further handling. That could be mp3 for standard players or m4b for Ipods. I want to do these steps for roughly 40-50 audiobooks.

Looks like I will do it the standard way.
  • Will align (Autogain?) and convert single mp3 records to flac with Goldwave batch processing
  • Then merge all flacs into one file (9-12 hours long with cues), jump from cue point to cue point and remove the technical information
  • Reduce noise and detect new cue points (for later mp3 file sequence)
  • Save the flac for later handling.
Klaus

@DougDbug: thanks - will look into mp3DirectCut.

Addtl remark: it really helps that I can run several instances of Goldwave at the same time!!

Re: Converting a library of audiobooks for blind people

Posted: Tue Apr 10, 2018 9:36 am
by GoldWave Inc.
You could save the files in "Extended Audio (*.xac)" format. That is an uncompressed format that can store very large files. There is no decompression time to open a file (the file opens instantly). It's a good intermediate format if you have lots of storage space and want to save time (and preserve quality).