So as you know WAVEFORMATEXTENSIBLE uses specific order of channels and RIFF WAV container must adhere to it, each channel should go after the other in that order and channel mask shows which channel names to skip and which not. So... In ffmpeg it shows correctly FL+FR+FC+LFE+BL+BR+SL+SR+TFL+TFR+TBL+TBR (so backs are before sides) yet in Audition interface Ls, Rs go first and that introduces a bug. Ideally you must change Ls to Lsd and Rs to Rsd so that in interface it goes (as in Microsoft standard) Lsd, Rsd, Ls, Rs. Or if you think this is not cool, reorder the Ls with Lsd contents. This is very bad since some Atmos masters are like this (if you do not use automation, so no need in dynamic objects). Dolby Media Encoder, BTW, also looks affected. FFmpeg extracts correctly with ffmpeg -i cnahcna8.wav -af "pan=mono|FC=BL" mono5.wav Sample zip attached (rename from docx to zip). Export thankfully not affected. See the screenhot:
... View more