[RNLD] SG0120: technical issue

David Nathan dn2 at SOAS.AC.UK
Mon Aug 13 10:42:12 UTC 2012

Hi Jonathan

I understand you already contacted my colleague Dr Chatsiou at ELAR
who is currently trying to find someone who has skills with Flex etc
and might be able to help you.

The questions you raise are mainly related to the language
documentation activity rather than archiving per se. ELAR focusses on
preserving and disseminating what you deposit with us, the main
considerations being that the materials are in sustainable formats and
with sufficient metadata or explanation so that we can present your
deposit in such a way that the relevant users can find, understand and
use the materials as intended (ie as intended by you or by your
documentation project goals).

I'm not knowledgeable about Flex but I understand that it has various
suitable export formats, such as XML or "standard format", which would
be fine for archiving purposes (XML is probably better!). You might
want to check that any such export is explicit and transparent enough
so that it is understandable without the Flex software itself, and/or
is directly importable by other software such as ELAN (which you
suggest); otherwise, you could provide a suitable plain-text
explanation of the data and relationships. In other words, a
description of how you made the export would be helpful, as well as
description of the fields, and in particular what fields or values
link data in the Flex export with the Praat data. ELAR's online guide
on simple bundling is the other part of the solution - as you say, you
can bundle together a related Flex, Praat and audio file, with such
metadata or explanation as might help someone understand and
manipulate the data. If you wish, you can send drafts of your
descriptions of the data/exports and other metadata to ELAR and we can
provide feedback about it.

If someone can help integrate the Flex and Praat data, eg, using ELAN,
it would obviously be a bonus. None of our archive staff currently
have these skills, nor would it be a service that we would normally
offer. Note that you can update an ELAR deposit in the future if such
a solution emerges, and that, even if it does, it may well be useful
to have the original Flex and Praat data archived anyway.

I hope this help, best wishes


---------- Forwarded message ----------
From: Jonathan Brindle <jonathan.brindle at gmail.com>
Date: 10 August 2012 20:06
Subject: SG0120: technical issue


I would like to get some recommendations from the ELAR team, or anyone
who could help (feel free to forward this email), on how to archive
the work I have done so far. You wrote "You can even invent your own
system as long as you document it and use it consistently."

The issue I have is that I worked with Flex (for lexicon,
interlinearization, translation and comments), and all my segmentation
was done in Praat Textgrid. So I have all the necessary information,
but they aren't linked!  The Praat Textgrid contains dummy markers for
(i.e. A, B, C) representing paragraphs/text blocks in Flex. Do you
know people who manage to merge the two (i.e. annotate their text in
Flex with the Praat time information of the textgrid). I doubt, but if
it is possible. How do they do that?

I know I can use ELAN, but the project developed without the use of
ELAN, for different reasons, one being that doing the
interlinearization work in ELAN doesn't 'feed' the lexicon, which I
like to build in Flex. Do you know someone who could help in a
procedure which could bring an interlinearized Flex .xml file and a
praat textgrid together in ELAN?

But, crucially, do I have to do all this at this point (my project is
over and need to submit to the archive)? I would think that 'Yes', but
I would like to have your opinion on that. I am ready to merge them in
ELAN, but need to know how, but I could also document my "own system"
and bundle a flex .xml file, the .wav  file, and the Praat Textgrid.


Jonathan Brindle <jonathan.brindle at gmail.com>
Project sg0120

David Nathan
Endangered Languages Archive

More information about the Resource-network-linguistic-diversity mailing list