Voxer texts....

What have you tried so far? What software have you used to perform an extraction more generally? Do you have any files in the Application's folder?

Hey everyone,

I bumped at the Voxer App as well. But I'm not interested in the text messages, but in voice recording from the Push to Talk app. It seems that is not a typical audio codec. At least VLC doesn't recognize it. On their homepage http//voxer.com/legal/proprietarynotices they are listening

Voxer® Proprietary Notices and Third Party Licensor Notices, Disclaimers and License Requirements

I've found there the Skype audio codec SILK listed. But after doing research on Silk I've found out that it's not the SILK codec but something else. It has to have #!SILK\n as header information. I didn't find this header on my files.

Does anybody have an idea how to play that voice record files?

What header did you have on your files?

there is the folder Documents\Messages\. There are more than 100 sub-folders with message_ as prefix i.e
message_1341678560945_0866591922_3abba19 or
message_1341670699412_0287941589_2b3b730

i've figured out
1341670699412 is unix-time formated, confirmed with the creation timestamp of the folder
0287941589 don't know yet
2b3b730 there are several folders with this suffix. could be parts of one conversation which belong together

inside this folder are two or three files and a subfolder named parts
files
"END" size 0
"journal" with this content 175897800
"header.json" with this content (XYZ, ABC and MySubject are replacements for real names)
{"content_type""audio","create_time""1341670699.412435",
"from""user.XYZ.1340641899195_62153783",
"message_id""1341670699412_0287941589_2b3b730",
"subject""MySubject","talk_mode""ptt",
"thread_id""HL_user.ABC.1331775136058_73026264_user.XYZ.1340641899195_62153783","to"["user.ABC.1331775136058_73026264"]}

the subfolder "parts" has always only one file named 0. the size of 0 varies from 10kb to 150kb, mostly less than 100kb
here are the header of some the files "0". the values in brackets differ
04 [31 64 0A] 20 20 20 20 20 20
04 [31 62 0A] 20 20 20 20 20 20
04 [31 35 0A] 20 20 20 20 20 20
04 [66 0A 20] 20 20 20 20 20 20

thats all I have figured out for now. I appreciate more given hints.

The most relevant part seems to me the "header.json".
From the (very, VERY little) I know JSON
http//www.json.org/
is a sort of XML , a kind of "descriptive" language (and "container")
http//www.w3schools.com/json/default.asp
http//en.wikipedia.org/wiki/JSON

I would try to find a JSON parser capable of correctly reading the "textual" part, first thing.

See if this fits
http//tomeko.net/software/JSONedit/index.php?lang=en

jaclaz

If you want a free json viewer try Allan Hay's site http//www.ash368.com/#/os1/4547935104

H

for me is the header.json less important. I'm actually interested in the voicerecord files and how to put them together and what codec it is. As I mentioned before, the audiofiles are probably splited in several folders with the same suffix.

for me is the header.json less important. I'm actually interested in the voicerecord files and how to put them together and what codec it is. As I mentioned before, the audiofiles are probably splited in several folders with the same suffix.

Yep ) , but maybe the "structure" is useful or needed.

See if this works ?
https://github.com/Voxer/voxer2mp3

This small tool converts Voxer audio to mp3 format. Voxer uses a minor variation of the open-source SILK super wideband audio codec from Skype.

maybe the issue with the previous attempt is the "minor variation".

jaclaz