photog.social is one of the many independent Mastodon servers you can use to participate in the fediverse.
A place for your photos and banter. Photog first is our motto Please refer to the site rules before posting.

Administered by:

Server stats:

250
active users

#proaudio

1 post1 participant0 posts today
Continued thread

Features I particularly like for a desktop #ProAudio interface:
- USB C w/ separate USB C power (for situations where the host can't provide adequate power, e.g. a phone)
- physical on/off
- per channel switches for 48v, line/instrument, and live monitoring
- per channel knob for monitor mix
- big fat volume knob
- gain boost button and separate volume for headphones
- pleasantly stiff analog knobs
- separate level meters for both inputs and outputs
- all of the buttons light up when engaged

Continued thread

But it also occurs to me that this is a use case where I could see value in machine learning tools

to recreate the cadence and art of the delivery cleanly from existing sources would make these much more accessible to modern audiences.

Reading King is not the same as hearing him speak. And I'm not normally one to listen to speeches.

(reposted to fix borked threading)

2/2 #ProAudio

I have a 16-disc (!) set of sermons and speeches by Dr. Martin Luther King, Jr. that recently I got from a library fundraising sale for like a dollar.

The audio quality in some of the early stuff especially is very poor, owing in part to the circumstances of recording and in part to the limitations of available technology.

A lot could be done with modern software to clean this up, so I'm going to give it a swing as an exercise.

1/2

After spending part of the last week trying to figure out a good workflow for noise reduction in problematic spoken word audio

(and documenting it here venya.soundslike.pro/blog/2025)

I am now playing on hard mode by trying to fix up audio from an outdoor wedding recorded on a not-close Zoom H4N under a local airport flight path

venya.soundslike.proVenya does sound - Meeting ACX (Audible) standards with Adobe Audition (and Reaper)
Continued thread

When I switch to the melodic range spectrogram layer, you can see WHERE the ref mix is louder. The low end is much more powerful throughout.

But there are also some short higher pitches (xylophone, I think?) arpeggiating throughout that are much louder. It's PRETTIER, but I didn't actually like how bright they were in the ref mix.

Continued thread

Reference mix is on top; our mix is on bottom. These are normalized to -1 dB peak, because I am specifically evaluating the mix. (If I just wanted to compare how they sounded and which I preferred, then it would make sense to normalize to perceived loudness, the default for #ffmpeg-normalize.)

Our mix has more dynamic range, prob too much. That fits _my_ tastes, but it's not what the market prefers. You can see some limiting in the ref mix, but not excessive.

I need to compare a reference mix to the one that my team did in the AES student mixing competition last month. #SonicVisualiser is a great tool for this, but first I need to normalize the audio tracks so they are at roughly the same perceived loudness. ffmpeg-normalize is a quick CLI tool for this purpose.

github.com/slhck/ffmpeg-normal

(I had to install several things to get to the point I could use this, but "you need this" first messages were helpful and clear. Yay, #Debian.) #ProAudio #MusicDev

Audio Normalization for Python/ffmpeg. Contribute to slhck/ffmpeg-normalize development by creating an account on GitHub.
GitHubexamplesAudio Normalization for Python/ffmpeg. Contribute to slhck/ffmpeg-normalize development by creating an account on GitHub.

My new sticker order arrived. I am trying out TheStickyBrand.com, which had a promo to get 100 of these (about 3"x2") vinyl holographic stickers for $25.

I like the way they turned out. There are some other designs I'd like to have lots of. I will order more. These are destined to be given to my audio nerd friends (and former instructors, one of whom inspired the design).

Which is available here, btw: archive.org/details/bs-meter/B