Last Active: Nov 29, 2024
Threads: 227
Posts: 6,334
Reputation:
26
Maybe this could be on the Music forums but since focus is on finding the app to one particular job, I'll post here. Sorry for any inconvenience.
I got a YT video with a recording of a 1949 vinyl. It's an old Japanese music on what may be a Polidor 78 RPM. But it's full of crackling and hissing.
I'm not an audio expert and an mostly automated app would be ideal. I want to extract the MP3 file and clean it some, 128 Kbps mono will suffice.
Any suggestions, hints, or starting points will be appreciated.
Last Active: Sep 12, 2021
Threads: 28
Posts: 2,900
Reputation:
36
I would try and maybe list the artist and album, that way if that audio needs to be cleaned it is done from a lossless source.
Last Active: Nov 29, 2024
Threads: 227
Posts: 6,334
Reputation:
26
That's a good idea, but sadly it's all in post-war Japanese and few people can read that in my location. Maybe I can find someone on the Internet.
Last Active: Mar 08, 2019
Threads: 415
Posts: 3,499
Reputation:
47
Audacity and Click Repair would be good places to start dueda. But using youtube as a source you might as well not bother trying to clean it up. It's like putting a tie on a turd.
Last Active: Nov 29, 2024
Threads: 227
Posts: 6,334
Reputation:
26
That's true, but Audacity is mostly manual work, I'm not at all good with audio techniques. Gonna check on them. Also maybe I find other sources.
Last Active: Jan 14, 2020
Threads: 2
Posts: 28
Reputation:
2
Dec 28, 2018, 12:50 pm
(This post was last modified: Dec 28, 2018, 12:51 pm by wwary4. Edited 1 time in total.)
perhaps deep neural networks could 'enhance' audio. such tools should be available soon if they aren't already.
check out tensorflow.
Last Active: Nov 29, 2024
Threads: 227
Posts: 6,334
Reputation:
26
Thanks guys! Successfully used Audacity and Click Repair, they have automated tools to do the job but I'd to fiddle a bit with parameters, nothing hard for a layman with damaged hearing.
Results were good with my hardware, but nasty artifacts were still audible with one phone (Samsung Galaxy J7 2017 model). I suspect prolonged use on max volume damaged the phone's loudspeaker, thus rendering the residual glitches very audible on it, while almost inaudible on better hardware.
TensorFlow, on the other side, is a tool for building AI applications, I guess it would take me huge effort and resources I can't even scratch to do it.
Last Active: Mar 08, 2019
Threads: 415
Posts: 3,499
Reputation:
47
Nice, and you also learned something. Glad it worked out.
Last Active: Jan 14, 2020
Threads: 2
Posts: 28
Reputation:
2
IIRC, Sound Forge also had some options for that, but its not free.
Well, there are some tools for tensorflow, but most of them are for images/video. that will be best solution for such tasks.
Last Active: Sep 12, 2021
Threads: 28
Posts: 2,900
Reputation:
36
If you have Adobe Audition you might want to try and use that.
|
