My number one hated stage in transcription work is segmenting. I would sit there fuming while manually segmenting the recordings I made before I could even start transcribing. It was frustrating because it seemed like something that a machine could to a relatively good approximation of instead of me sitting there for hours doing it for each file!
Luckily, it turns out that between Praat and ELAN, you can very easily have a decent approximation of segmentation done for you. Not perfect, but it saves HEAPS of time. If you have a ton of recordings to segment into units before you need to transcribe, this is the process for you!
Thank you to T. Mark Ellison for helping out with this heaps.
Praat stage
First load the sound file that you want to segment into Praat (Open > Read from File). Create a Praat Textgrid file based on silences.:
This next part is our best setting after a few trials:
The resulting text grid should look something like this:
The *** is where Praat has segmented for sound. It’s not perfect, but it gives a pretty good shot at things, and you can adjust the boundaries manually in Elan. Save this text grid now.
ELAN stage
Import your Praat text grid:
Cheers Hedvig for letting me know that if you tick the “exclude silences” box, you can have ELAN automatically remove any empty segments from the Praat text file:
And you will have your segmented Praat text grid as a layer in Elan looking something like this!
The longest file we tried it on was a 1 hour recording of Samoan (cheers Hedvig Skirgård from Humans Who Read Grammars for providing the file!). It took about 8 minutes for Praat to segment. A 10 minute recording is done in no time.
Now be on your merry way setting up your tiers and transcribing to your hearts content 🙂
Yammering On 
I tried this out with a recording done with one person in a quiet room, using a headset mic and it seemed to work well. I compared it with the inbuild Elan silence recogniser ‘Silence recogniser MPI-PL’. That worked well too but creates annotations for the silences too – which are a bit odd. If you really didn’t like them you’d have to work out how to delete them later.
LikeLike
Thanks for that Eri (and Mark). I like to segment my files while listening to them, there’s always something that’s revealed when doing that, and most of my files don’t exceed 10 minutes really – but thanks for sharing this, I’ll give it a go 🙂
LikeLike
Heya Marie, thanks for the comment! For shorter files, I agree, it’s sometimes just better and more informative to segment yourself. Let me know if you find this helpful for longer files!
LikeLike
Hi Eri,
I just tried this with some of my data and it works fine!
The only problem I have is when I try to modify the segments in ELAN…
Is it possible? and if so, how?
The way I did it just created a separate tear named “silences” that corresponds with the ref. tear but I am unable to make any changes to the selection.
I would really appreciate some help.
Thanks!
LikeLike
Hi Ev, sorry for the late reply! I just came back from the field on the weekend and now catching up with online things 🙂 In the event you haven’t found a solution yet: you should be able to modify the segments if you are in segmentation mode, not in annotation mode. You can drag the arms of the segments, or just move entire segments to wherever you want to.
LikeLike
Thank you so much! Great trick. Do you happen to know how to import the textgrid to an existing tier? For some reason I can’t change the attributes of the new tier that is created for the textgrid (automatically called “silences”), which is a bit annoying.
You know what my biggest ELAN-annoyance was? All the double-clicking for making segments in other tiers than Words (which I could do with Tier > Tokenize) when I did glossing and translations. I now found that you can get segments in the Gloss and Translation tiers by clicking Tier > Create Annotations on Dependent Tiers.
Also, if you go to View > Shortcuts > tab Annotation Mode, you can make easy shortcuts for “Go to next annotation and start editing”.
(Maybe I’m the last linguist to find out these things, but you never now.)
LikeLike
Hi Eline, thanks for the comment. It’s pretty amazing how many little tricks in Elan you only find out through speaking to other people (and yes, does make one feel like they’re late to the party haha).
Unfortunately I don’t know if there is a way to import the textgrid onto an existing tier. In the even that you’ve figured out a solution to this problem between posting the original comment and now… do please let me kno! I want in on the trick 😀
LikeLike
I have been using this program for transcribing and the information is very helpful in getting things started. However, the initial Parameters for the Intensity Analysis seems to have a slight issue with breaking up the sound file too much to single words at a time or even only parts of words at a time and then requires a great deal of boundary movement to match the transcription to the sound file. We have played with the numbers and discovered that changing the Minimum Sounding Interval to 0.2 instead of leaving it at 0.1 seems to work better, giving it a slightly bigger range to fit more of the word and more words in all together. It makes it longer sounding portions, but it does help keep the words in tact.
LikeLike
Hi Amanda, thanks for the comment! This is super interesting, since I find the opposite problem with the basic parameters set out here – clauses upon clauses end up getting segmented as one huge chunk with no breaks! (What language were you segmenting?) I’ll have a play around with the settings too and see what I end up finding for Nambo!
LikeLike
I tried doing this but as soon as I open a file I get an error message that says the file is too long or something like this. What could I do?
LikeLike
Wow sorry, this message somehow got lost in the quagmire. I’ve never had such a message come up, even for the 1 hour long recording. So maybe your recording is much longer than that. I don’t have a solution to this problem, I’m sorry. If you manage to find out what can be done, do let me know.
LikeLike