Slice using computer?

Oh hey.

http://www.sonicvisualiser.org/doc/reference/2.1/en/index.html#layerexport

This looks interesting! Have you (or anyone else) tried using it?

Have not but will download and give it a shot. Still very surprised that nobody’s bothered going down this road.

I asked recently the same question of 1010 Bitbox and discovered that the format is in a proprietary XML. The information’s easily accessible there, but not to any music software!

I’ve been very surprised over the years at the lack of standards regarding markers and metadata in audio files and the difficulty of sharing them with the files.

I produce long form audio documentaries for radio and markers are incredibly useful to find snippets of audio dialogue in sometimes hours long audio files. It’s almost impossible to do when using different softwares and so paper edits still reign. There’s a cross platform software developed explicitly for radio production (which I shan’t name) which doesn’t even retain timecode information in files once cuts are made in them. That is to say, as above, one has to go back through the transcript to find the timecode at which a cut was made! I share jj.kidder’s very surprisedness!

I think it was Reaper which was the only software which made markers available for Morphagene.

It occurred to me that researchers that work with speech (linguistics, speech pathologies, speech recognition, and so on) have pretty sophisticated needs when it comes to labeling audio. It might be worth looking around there since the barriers to sharing are lower in the scientific communities.

For example, here is a program designed for labeling speech (it even draws pitch contours) but works on any kind of sounds:

http://www.fon.hum.uva.nl/praat/

The output is a text file.

http://www.fon.hum.uva.nl/praat/manual/TextGrid_file_formats.html

I tried it out. It only crashed once. Haha.

Hold on. Audacity has something called “Label Tracks”.

http://manual.audacityteam.org/man/label_tracks.html

Just tried quickly adding 3 labels to an mp3 file:

Then exporting the label track yielded a text file with the following 3 lines:

4.478823	6.422177	test
9.413120	11.083190	test2
14.180410	16.366683	test3

Looks like we might have a serious candidate…

2 Likes

Yeah, Praat was one I was thinking of, but it’s written in Java, is it not?

i.e for a format to load into the 301? The issue often with markers is to ensure that they stay in the same place on the audio file, no?

I still have a copy of ReCycle somewhere but haven’t used it since, well… the days of Reason! That seems like an awful long time ago… it really is the biggest tragedy that Propellorheads went so hardline with their copy protection and encryption as this would have been a world dominator by now if they had just chilled out a little bit and been more open. They certainly lost a customer here precisely because of this.

I just took a peek at a .slc file from the ER-301 and the information is only really machine readable i.e. not stored as XML.

Found this round up of the different slice file formats

http://www.chickensys.com/translator/documentation/formatinfo/slicefile.html

I tried to look at Translator, but i couldn’t work out if it handles slicing or not, it’s a bit weird and the website a bit broken.

Vice seems popular, but I’ve no idea what export looks like!

Oh wow - just seen all this activity now… maybe what I wrote is no longer relevant!!

1 Like

Audacity looks like win!!

Nice find, I had no idea it did that :smiley:

Best thing is, I already have it installed!!

Here is a huge software suite from McGill for music research:

The timecodes in the Audacity label track are in seconds from the beginning of the file. I don’t see any problem with loosing time sync with that. Do you?

I used that in the abovementioned translation to a different file type for the ABC here in Australia to move from Logic to Steinberg’s WaveLab and it worked very well…and was surprisingly current with format versions despite the archaic website…can’t be sure about slicing 'tho

1 Like

That’s right. It is literally just binary-packed C structs at the moment (the fastest read times). I knew I would have to revisit this in the future so I made it quick and dirty thus easy to throw away and redo.

You’re WAY better qualified than I (I’ve not yet my hands on a 301!) I’m just thinking of relativity in terms of trimming, i.e. if you move a start point the text information isn’t dynamic, so it may shift the position of the marker unexpectedly?

Did I explain that sufficiently?

Actually, I wasn’t planning on using the text data at all, except maybe for display purposes. Slice # numbers would be recalculated each time in the ER-301 after loading and sorting the annotations.

So this really is blue sky development, that’s awesome!!

I really approve of taking this to the next level and making an open source slicing protocol… I realise this is a huge amount of work though!

I am seriously getting feelings that an open source version of ReCycle would be the bomb :robot:

I wonder how difficult it would be!

yes, well that would be ideal…

In some of those cases i mentioned the information for markers held in timelines rather than in the audio file itself and so quickly became useless.

…and of course Audacity is freeware!

I recently updated it from a 1.x version to a 2.x version as i was searching for something to decode DTMF codes and discovered that there. It also does silence detection in a great way…tools in it for free that I have paid fairly significant money for over the years.

1 Like