How to process vocal stems in Spectralayers 11

SpectraLayers
1

When I unmix a song I find that when I solo the vocals, they are muffled usually when the voice is lined up with a kick or snare. I figured out how to take random sounds out of the vocal stem by using vocal denoise but can’t figure out how to clean up muffled vocals. I am new to spectralayers/audio restoration so trying to learn as much as possible.

1 Like
2

Was wondering about the definition of ‘lined up’ and ‘clean up’ here. Do you mean that before unmixing the vocal sounded brighter than the result after unmixing and vocal denoising when compared with the sounds of the kick and snare? and you now want to enhance the high frequency definition of the vocal stem?

3

After unmixing, when I solo the vocals yes, some of the vocals are muffled and I would like to enhance or the highs on the muffled parts of the vocal

4

Of course, it’s extremely difficult to advise without hearing the audio. And your description is not very clear.

I would suggest that the unmixing may not be directly causing a loss of high frequencies in the vocal. There’s probably a lack of high frequency definition in the vocal source itself.

Unmix Song could be having a hard time with successfully separating things in badly recorded audio or when many instruments are operating within the same frequency range. The result is a compromise. If the source audio contains a badly recorded vocal element then the vocal stem will reflect this same badly recorded vocal element.

You’d probably need to apply EQ and / or compression to enhance the vocal stem, probably in another application, as part of an automated mix. For basic volume automation in SpectraLayers you could use the volume envelope curve available for each layer (see p.35 of the Operation manual).

5

You can also try the Debleed feature to get rid of drums or what not from the vocal stem.
Just make the vocal layer active, select the Debleed module, check Drums in that module’s window and go.

I think this will not muffle the vocals as much. At least you can play with the parameters to optimize the result.

1 Like