Andrea on Warpcast

Content pfp

https://warpcast.com/~/channel/nerd

0 reply

0 recast

0 reaction

Andrea pfp

Using ChatGPT to compare the frequency responses of the two tracks. Song: Dio - Stand Up And Shout 1) Audio Fidelity mastered By Steve Hoffman 2) 2022 Joe Barresi Remix

3 replies

1 recast

3 reactions

Desh Saxena pfp

This is sick, I kinda wanna save this to show my audio engineer friends

1 reply

0 recast

1 reaction

Andrea pfp

I also did another test to check the hum noise in a guitar recording, where I suspected the problem was around 140 Hz. Chatgpt: "There is indeed a fairly obvious peak around 140 Hz, confirming the presence of a hum at that frequency. There are also weaker harmonics visible in the multiples. You heard right!"

1 reply

0 recast

1 reaction

Desh Saxena pfp

Interesting, did chatgpt offer any solutions to combat the hum? 🤔

1 reply

0 recast

0 reaction

Andrea pfp

Yes, chatgpt can also help you eliminate any residual harmonics or strengthen the cleansing by apply correction on the file uploaded.

1 reply

0 recast

1 reaction

Desh Saxena pfp

Oh damn, that's really useful. This is the type of stuff I like when it comes to ai in music, when it makes our lives easier

1 reply

0 recast

1 reaction

Andrea pfp

However, it's better to upload the audio at a lower level, since once ChatGPT makes edits, it could introduce clipping just like any other audio tool. Not sure if you can request a peak limiter.

1 reply

0 recast

1 reaction

Desh Saxena pfp

Yeah seems reasonable, I wonder if there are any apps developing music centric ai tools out there that focus on mixing and audio editing, might be cool to test out in future projects

1 reply

0 recast

0 reaction

Andrea pfp

Yes there're mastering tools with AI but I remember wasn't free. I've tried one for demo but the results where crap, it has added unnecessary compression.

1 reply

0 recast

1 reaction

Desh Saxena pfp

Yeah makes sense, I mean the little quick mastering I do for my demos on logic's mastering tools are just compressed anyways, but tbf I'm not doing much mixing anyways cos of my timeframe

1 reply

0 recast

0 reaction