🙌🙌🙌🙌🙌

Wow Sadie, amazing!!!!!

So fucking good, it was inspiring!! I want to see it again in 4DX

Ahh Travis, thank you brother!!! So glad you're listening, man!!

Mostly I'm just glad I no longer have to do it MANUALLY like I was for the first few eps of Gutter. Added an easy 6-8 hours of EXTREMELY tedious work to every edit.

It takes 30-45 min depending on the length of the file and I basically can't touch my computer during it since the mouse cursor needs to be exactly where the script moves it to on giving functions without interruption.

It's a hassle but the difference it makes is night and day.

I need to run it once manually every time though to get the proper Learn and Render processing times, then modify the script with those times (plus 5 seconds of buffer), as well as the total length of the track and the length of the selection (rounded down usually).

Ideally, iZotope would've accounted for this and made a version of the process that progresses chronologically through the track but they didn't. Thus the script.

The script makes a selection 1/10th the track length, "Learns" the bleed profile, "Renders" it *twice*, then moves on and repeats.

The latter. I initially tested it on small pieces of test audio and it worked stunningly well, so I went ahead and ran it on a full test track (took ~45 min). The result was basically non-existent.

Turns out it starts losing efficacy on selections longer than 5 minutes, dramatically so after 10.

*brushing hands together* easy peasy!

But of course, as mentioned by the Gutter account, a crucial first step is proper sound treatment and separating the voices with sound blankets or big acoustic panels, rear-to-rear.

It should be noted that they're recording through the same audio interface so the tracks are perfectly in sync otherwise it would be unsalvageable

This is all to avoid the muddy acoustic profile of dynamic mics that I loathe so much compared to a bright clear capture of our Rode NT-1 condenser mics

Then after that, I put a noise gate on both tracks and do a lot of manual cutting. I also crossgate our speech leveling which helps the remaining mic bleed not get lifted.

It still bleeds pretty significantly but I use iZotope RX 10's mic bleed reduction process on every 5-10 minute chunk of audio with a custom python script a friend helped me write. It's the process for our in-person recordings as well but when we're remote I only have to do it on M&E's tracks.

Yessssss!!!

this one's so so good............................