Coyote on Nostr: I am still :ablobcatsweatsip: ing binkle It's best if you encode lossy stuff yourself ...
I am still :ablobcatsweatsip: ing (npub1y6t…9wl9) binkle (npub1kdy…2hnf) It's best if you encode lossy stuff yourself when testing so you know the audio isn't going through some filters that'd give false positives, but speaking on lossless, I can only tell the difference forensically; higher quality lossy codecs are indistinguishable to my ear (while also being a 10th the size).
A fun hands-on experiment is to take a known lossless recording, encode an MP3 of it (MP3 specifically because it tries to match the original waveform rather than being perceptually the same), flip the waveform, and listen to the interference. You won't hear anything during quiet parts and the loud parts will sound like your speakers blew out. Here's an example from the Spirited Away Suite, 8:28 to 8:39. The first file is the lossless interference with Audacity's "220-260 kbps (Best Quality)" variable bit rate MP3 and the second is the original lossless.
https://media.singing.dog/e5a6e32147301791961895910a753bd415969ed47bd7f44bf0f33b2e505f49b5.flac
https://media.singing.dog/86fc88d190d16ad642c58019e3558c85a6f59fd4bd6a0590204fdff1138ba28e.flac
A fun hands-on experiment is to take a known lossless recording, encode an MP3 of it (MP3 specifically because it tries to match the original waveform rather than being perceptually the same), flip the waveform, and listen to the interference. You won't hear anything during quiet parts and the loud parts will sound like your speakers blew out. Here's an example from the Spirited Away Suite, 8:28 to 8:39. The first file is the lossless interference with Audacity's "220-260 kbps (Best Quality)" variable bit rate MP3 and the second is the original lossless.
https://media.singing.dog/e5a6e32147301791961895910a753bd415969ed47bd7f44bf0f33b2e505f49b5.flac
https://media.singing.dog/86fc88d190d16ad642c58019e3558c85a6f59fd4bd6a0590204fdff1138ba28e.flac