I’m a long-time audio professional new to SL. I go back to the days of razor blades so SL is pretty mind-blowing for me. Wow…
That said, I’ve been tasked to improve a 1980’s cassette recording of a (vocals, two guitars, bass, and drums) new wave band. It’s a bad recording. Not only is it a cassette, it’s clearly not first-generation. I can’t even tell if it’s Dolby encoded.
So, the reason I’m using SL is that I was hoping to deconstruct the songs, sweeten the stems, and then put it back together. I played with it last night for the first time and, not surprisingly, SL did a so-so job of de-mixing. I don’t blame the software at all, but…
Can anyone offer suggestions on how to get better stems from a crappy recording? Should I “pre-sweeten” it before trying the extraction? I would welcome any suggestions. Many thanks,
I dont have an answer. Just a few tips.
Try stem separation twice. It can help. Also trying different settings. And once with the checkbox whose name I can’t remember checked and once unchecked.
Then Separate each by tonal/noice.
That may give you the separation you need.
And of course you can separate using the spectrum manually as well.
When you have achieved the technical separation you can merge the layers back again where you no longer need separation (it doesn’t make any sonic difference just makes for fewer lanes to deal with. Then you can process/filter and then add whatever sweetening you had in mind.
I do doubt them at the outcome will be great though. This type of processing really works best on hi quality material and less so with stuff that is bad quality.