The Benchmark database of phonetic alignments was in fact the first private illustration to me showing that we could not link all data to phoible. It is useful to add it. Writing a crawling script to get data from the online app is simple by iterating over urls and searching for alignments in <code> blocks, but all files are available for direct download as well, so it seems best to just make a little export. To illustrate a particular sound, one could link to a given URL where it occurs and the alignment is shown:
This would illustrate the sound in action, but not capture all instances, yet this may be enough. If I remember properly, we are talking about some 300 sounds here.
The Benchmark database of phonetic alignments was in fact the first private illustration to me showing that we could not link all data to phoible. It is useful to add it. Writing a crawling script to get data from the online app is simple by iterating over urls and searching for alignments in
<code>blocks, but all files are available for direct download as well, so it seems best to just make a little export. To illustrate a particular sound, one could link to a given URL where it occurs and the alignment is shown:This would illustrate the sound in action, but not capture all instances, yet this may be enough. If I remember properly, we are talking about some 300 sounds here.