I have imported old time-aligned files from Toolbox where a person still used a cassette, which had two sides. For each side, there are about four Toolbox files, like this:
Side A: File1 Time 0-400, File2 Time 400-800, File 3 Time 800-1200 etc.
In order to keep the time stamps for all files, I needed to merge them into one ELAN file to correspond to the same audio file. If I split the audio file into many parts, any time stamp after the first file will be useless because it start from 0 again.
However, since each file has their own “ref” (i.e. speaker), ELAN creates a new tier set per speaker. This looks nice, but it is less optimal for a multi-file search because the morpheme tier exists four times in each file, with a different parent each time.
My question is whether ELAN allows me to split or save just an “excerpt” of the ELAN file including cutting the audio at the right time. I can save a part of the annotation but ELAN does not copy the audio information. The other option would be to change all “ref” information to a general “speaker” and maybe put the speaker information in a note tier, but before importing into ELAN. Does that make sense?
The option File->Save Selection as .eaf… has an additional option/checkbox “Clip media using the clip media script” and this sounds like what you need. You would have to look into how to setup the function File->Export As->Media clip using script…
Nevertheless, it’s not quite clear to me what your current situation is and what the situation you want to achieve. It sounds like you have 4 Toolbox files per cassette side, but you got them into one eaf file that you then want to split? Well, this probably isn’t the case, but it’s not clear to me why you need to split the annotation and audio files.
Anyway, in the multiple file search you can also choose to search in all tiers of the same type. So, even if you have several ref or morpheme tiers with different names, it is still possible to e.g. execute a query in all the morpheme tiers.
Apologies if I got the situation wrong.
After installing ffmpeg.exe and specifying the folder for it in the media-clip.txt file, I got rid of the error message. But when I open the newly created, cropped ELAN file, the program asks me for the clipped media file, which looks different from my original (of course) but this had never been created before. I had hoped that ELAN does create the new clipped media file automatically in the folder where the original, long media also is placed. Am I getting this wrong?
As applied in this function (i.e. in combination with the saving of a selection of the eaf), the clipped media files are stored in the same folder as the new .eaf file. When opening the new .eaf file, the media clips should be found automatically by ELAN. If ELAN prompts you to locate the media files then it might be no media clips have been produced? You can check this in the file explorer (are the clip files present in that folder or not) and you could check the logs (View->View Log…) immediately after exporting. The clips have the same name as the original with _begintime_endtime suffixes.
When I just tried with ffmpeg I noticed an error message in the log “Unknown decoder ‘copy’”. It seems newer versions of ffmpeg don’t support the “-vcodec copy” and “-acodec copy” parameters anymore. Removing both altogether seems to have the same effect, namely: use the same encoding for video and audio as the original. Specifying an existing coding also worked, e.g. “-vcodec h264”.
It is probably best to first try the command on the command line, to see if it works (entering the full path for the input (-i) and output file and start time (-ss) and duration (-t) values in seconds).
