nmf_to_wav
nmf_to_wav copied to clipboard
quarckster
Nice audio version 6.7
Hi,
First let me thank you for this project. It already helped me a lot!
Today I found an audio that couldn't be converted by this great code.
The Nice Player can play it, but the conveter couldn't convert.
The condition if headers["packet_type"] == 4 and headers["packet_subtype"] == 0: is never true.
Instead of the combination of (packet_type == 4 and packet_subtype == 0), the file shows only (packet_type == 4 and packet_subtype == 3).
The audio file is attached.
Can you give a light on this issue?
Thanks!
It requires to disassemble nice audio player where this file was recorded.
@zhongczlb - do you have this stood up somewhere where I can test out a file? I think I have a solution.
Hi @RyanStomel,
If @zhongczlb had the same problem than I did, you can test with the nice_audio_6.7.zip that I attached on my post.
Thank you for your help.
test out this file and see if it converted for you in your instance
It requires to disassemble nice audio player where this file was recorded.
If you want I can provide you with the player if you need, Drop me an email to [email protected]
I changed the email.... should be [email protected]
test out this file and see if it converted for you in your instance
Hi @RyanStomel sorry for the delay in replying..
I tested your nmf file in test_nice.zip and it didn't convert either.
Thanks
Unfortunately the problem is that the Python script does something it should not do to 4.1 NMF files. Firstly it extracts the audio stream but doesn't compensate for silence. When you have an NMF file which is compressed with g729 codec and it's for example 5h long, when 90% of the file is silence then the outcome file has only the voice packets. and it's only 10% the length of the original NMF. Secondly, NICE is using another NMF format since I think Engage 6.6. I contacted the author of the python script and he was willing to work on it after I pay him. So I stopped contacting him. If you want to use this tool to convert 1:1 NMF files from NIM 4.1 and below, then you can't guarantee file integrity since it doesn't cover the silence periods in the source file and thus you can't use it.
Hi,
First let me thank you for this project. It already helped me a lot!
Today I found an audio that couldn't be converted by this great code.
The Nice Player can play it, but the conveter couldn't convert.
The condition if headers["packet_type"] == 4 and headers["packet_subtype"] == 0: is never true.
Instead of the combination of (packet_type == 4 and packet_subtype == 0), the file shows only (packet_type == 4 and packet_subtype == 3).
The audio file is attached.
Can you give a light on this issue?
Thanks!
The script is ok, but ffmpeg dont realign uncompressed packets, and header was wrote isnt correct. The type and subtype of packets are 4 - 0/ 4 - 3 for audio and 4 -2/4 - 4 silences. Silences are suppressed from nmf file, but can calculate (dont need in my case).
Hi, just to be sure I get it, we can update the script to read this attached NMF? If yes, could you please indicate where to ?
Thanks :)
Hi,
in this link there is a modification of your code, however i need process NMF files more than 10 minutes, when i process these files i get multiple parts, https://github.com/mwittenbols/NMF-audio-to-WAV-convertor
How can i get only a part and not that the audio will be divided betwen several parts
Regards
