One person through phone loud speaker and one person directly into the mic

Hi! I’m trying to edit a Zoom H1n file of an interview between two persons talking on the phone. The phone was set to loud speaker and the Zoom H1n recorder was placed near the phone (and the person conducting the interview). I tried to use an EQ and a Compressor, but while I’m tweaking one voice (the male voice on the phone), the other voice (female) starts sounding worse, and viceversa. What would be the best method to make the two voices sound about the same? (same level, same quality) The interview is 40 minutes long, so separating the two sources (the two voices) would be a very tedious workflow…

You could try to use spectralayer pro for separation, if you don’t have license, you could try free UVR (ultimate vocal remover), but that may require additional model download from the software

1 Like

@alin89c Hello. The only other solution I can think of is using AI powered software to auto detect the voices. An example is Cleaver Speaker Diarisation Software (I have never used it). Other than that it will require you to manually split the audio.

You could split it and do offline processing clip by clip with all the clips on the same track or split the clips to different tracks, which is an option you want to avoid but I think is the best choice (unless it is a budget limitation).

A final option is to automate your processing which I think would be more labor intensive than splitting the tracks.

1 Like

Exactly, the Unmix Mutiple Voices function in SpectraLayers 10 is what you need.

1 Like

I didn’t know there are multiple SpetraLayers products out there. Looks like it’s time to start saving up to get the pro SpectraLayers.

What do you mean exactly by automatic the processing?
Do you mean trying to find an easier way than manually splitting the two voices on separate tracks?

Thanks for taking the time to post this YouTube video link!

@alin89c hello, by automate your processing I mean you have it change overtime. An example is to have an eq preset for voice one and a preset for voice two. You will use the read and write function to automate (move back and forth) between both presets. You can do the same for dynamic processing, etc.

You can see how it’s easier just to manually split the audio or use AI software to split it. Hope it helps.

1 Like