I need to use the program called ogg123
for converting input ogg file to wav. My ogg is on S3 - so I need to download it first and transcode after that. But I think that will be faster to transcode sound on the fly - without writing to the disk source ogg file - using process substitution.
From bash I want to do this:
ogg123 -d wav <(./test.py) -f out.wav
In Python, for binary output, I tried:
os.write
sys.stdout.buffer.write
fp = os.fdopen(sys.stdout.fileno(), 'wb'); fp.write
For example:
#!/usr/bin/env python3
import os
import boto3
s3_client = boto3.client('s3')
os.write(1,
s3_client.get_object(
Bucket='my-bucket-name',
Key='file.ogg'
)['Body'].read()
)
./test.py
in all cases print visually valid data. For example:
xxd <(./test.py) | head
print:
00000000: 4f67 6753 0002 0000 0000 0000 0000 0000 OggS............
00000010: 0000 0000 0000 5664 269e 011e 0176 6f72 ......Vd&....vor
00000020: 6269 7300 0000 0001 2256 0000 0000 0000 bis....."V......
00000030: 058a 0000 0000 0000 a901 4f67 6753 0000 ..........OggS..
00000040: 0000 0000 0000 0000 0000 0000 0100 0000 ................
00000050: 5096 ab47 0e16 ffff ffff ffff ffff ffff P..G............
00000060: ffff c503 766f 7262 6973 0600 0000 6666 ....vorbis....ff
00000070: 6d70 6567 0000 0000 0105 766f 7262 6973 mpeg......vorbis
00000080: 2242 4356 0100 4000 0018 4210 2a05 ad63 "BCV..@...B.*..c
00000090: 8e3a c815 218c 19a2 a042 ca29 c71d 42d0 .:..!....B.)..B.
But all these variants generates the same error from ogg123:
Error opening /dev/fd/63 using the oggvorbis module. The file may be corrupted.
I tried the same file locally - file is correct:
$ ogg123 -d wav file.ogg -f out.wav
Audio Device: WAV file output
Playing: file.ogg
Ogg Vorbis stream: 1 channel, 22050 Hz
Done.
What can you recommend?
Thank you for help, I found that I can fetch through S3 using pure ogg123:
ogg123 -d wav https://s3.amazonaws.com/mybucket/file.ogg -f out.wav