Audacity has the highest quality for pitch corrections I heard so far. Also better than Adobe. However it sadly isn't possible to do a pitch correction with command line there and as I do most of my conversions with drag and drop cmd batchfiles I'm looking to replace audacity's pitch correct with a ffmpeg solution.
The best you could come up with in ffmpeg still isnt as precise as audacity is. For example a 4% (68 cent) downpitch of a 01:15:40:192 audio file comes out as 01:15:40:193 in audacity.
While in ffmpeg
-af asetrate=48000*0.96,aresample=48000:resampler=soxr:precision=33:osf=s16:dither_method=triangular,atempo=1/0.96 -ar 48000 -acodec pcm_s16le -f WAV %1%-downpitch.wav
comes out as 01:15:40:203 and
-af atempo=1/0.96,asetrate=48000*0.96,aresample=48000:resampler=soxr:precision=33:osf=s16:dither_method=triangular -ar 48000 -acodec pcm_s16le -f WAV %1%-downpitch-atempo-before-asetrate.wav
comes out as 01:15:40:200
To my ears the first command sounds a bit better than the second "tempo-before-asetrate" command, however not shure if the quality is really equal to audacity. I would have to test higher quality tracks to evaluate on that.
Looking forward to read about the experiences of the experts here and wish everybody a nice weekend
regards