On 2024-12-24 05:26, Jim Leonard via Hpr wrote:
Did post-processing of shows change at some point in the last few
months? Asking because there was a lot of audio dynamics compression
added to hpr4273 that wasn't present in what was submitted, and it
sounds fuzzy and overmodulated as a result. Normalization I can
understand, since people submit at different average volume levels...
but adding a set dynamic compression ratio to the source is not a good
idea because there is no set ratio for all inputs.
I also noticed at some point the opening changed; the opening few
notes are very quiet, and then suddenly get much louder when the intro
synthesized voice starts. (This is likely compression at work as
well.) I don't care how loud the music is; I'm just mentioning it
because, at some point months ago, the music was a constant level, so
something changed.
What changed, and when? And can hpr4273 be re-processed and
re-uploaded without added compression?
Yes we changed to produce stereo output. I remember saying it on the
community news but I can't find it back in the notes or transcripts.
I don't think that would make a huge difference but looking at the
diffs, I do see a change from
ffmpeg -y -i hpr${ep_num}.wav -af loudnorm=I=-16:LRA=11:TP=-1.5
hpr${ep_num}_norm.wav >> ${fname}_tmp.log 2>&1
to
ffmpeg -y -i hpr${ep_num}_sandwitch.wav -filter:a "dynaudnorm=p=0.9:s=5"
hpr${ep_num}_norm.wav >> ${fname}_tmp.log 2>&1
To be honest I have no idea why that change was made, but I usually only
make changes when something broke. Do you think that would account for
the change ?
So yes we can fix the, but there is no point reworking the audio until
we fix the scripts and confirm they are working.
With Dave stepping aside, the metadata preparation, show posting, and
media transcoding all need to be done by me.
I would like to automate as much as possible, while ensuring we keep the
manual quality control steps that prevent the majority of problems
getting through.
Therefore I will be reworking the flow to try and have the manual steps
be at the same time if possible.
I would also like to have a dashboard showing the waveform and spectrum
images, play links and transcripts, with links to all the files in all
the CCDN locations for to assist with issues like this.
*Call For Help*
To be honest I'm not the best to notice anything as subtle as this, not
that I don't check it, it's just that I don't spot this stuff.
Can audiophile people please subscribe to, and report issues with audio
in, the future feed <https://hackerpublicradio.org/rss-future.php>. Once
the file has hit the main feed it's already too late to replace the file
for anyone that downloaded it.
Anyone willing to help with the transformation scripts and documentation
should please create an account on our Git Repo
https://repo.anhonesthost.net/user/sign_up, and _email me 1:1 so I can
approve your account_.
They are due a cleanup anyway, so now is as good a time as any to start
working on this.
*Audio Workflow*
The principle of the Audio Workflow is dealing with random binary data
that can be submitted from any random person on the internet, in any
format they may feel the need to use.
We therefore we do not trust the audio, nor it's metadata.
We convert to raw PCM first, conduct a manual inspection of the audio,
and a visual inspection of the wave form.
Once it passes that the sandwich containing the intro, show and outro is
produced.
We then convert those and add the metadata (eg id3tags) from the show
submission database.
Then we distribute them.
*Required fixes*
Several items need to be fixed in the workflow.
* Fix audio dynamics compression (this issue)
* Transcripts are not working - due to conflicting dependencies
* Text to speech too complicated - due to conflicting dependencies
o Also requires a lot of manual attempts to get the text
correct(ish) - suggestions on how to improve this are welcome.
o Implement alternatives proposed by Archer72
* Some audio files (most recently flac) files come in without a
duration - why is this ?
* Ogg media is missing metadata tags - due to conflicting dependencies
* Anything else ?
Thanks everyone for Volunteering - And remember if you sign up before
2025 with the offer code "I<3HPR", we will double the remuneration for
new Contributors.
--
Regards,
Ken Fallon (PA7KEN,G5KEN)
https://kenfallon.com
https://hackerpublicradio.org/hosts/ken_fallon
_______________________________________________
Hpr mailing list
[email protected]
https://lists.hackerpublicradio.com/mailman/listinfo/hpr
