Harsh vocals

I had a badly recorded vocal – too close to the mike, too many peaks at various frequencies (450 Hz, 13 kHz, 17 kHz), big low end, too much presence at 2.4 kHz. I tried everything – compressors, EQ – but ended up with a do-nothing pitch shift.

Static EQ

This did not work. A few narrow shelves and notch filters did not either. The peaks shift, and a fixed EQ just changed the vocals too much.

Multiband compressor

I had some success with the Orinj multiband, multi threshold compressor. I will write more about it separately.

I want to deal with the peak at 13 kHz, I choose a band around it. In that band, I prepare a soft-knee compressor: no compression up to -30 dB, very slight 1.1:1.0 compression between -30 dB and -20 dB, some larger 1.5:1.0 compression above, and so on.

This has some problems. There are too many peaks to deal with, each needs a separate compressor, and these compressors are not independent. Changing one requires changing others. I find multiband compressors too difficult to tune and too destructive over vocals. When the glare is controlled, the compression becomes too noticeable.

Dynamic EQ

I went with TDR Nova first and Soothe second. The results were similar. Both are good. Soothe self-tunes and allows more precision.

I could not tune either of these well. The EQ-ed bands seemed too wide. Too much air disappeared from the vocal. Maybe I could correct, but I moved on.

Pitch shift

Surprise, surprise – a pitch shift was the winner.

I ran the vocals through the Orinj pitch shift, but without changing the pitch. The Orinj pitch shift processes without a pitch change. I am not sure about other pitch shift effects. This one runs the FFT and then the inverse FFT even if you don't change the pitch. Since the FFT is naturally imprecise, it smooths out frequency peaks. The glare is largely gone. Some air is gone too, but that can be fixed with extra EQ.

This was step 1. I now have a session with a pitch shift that does "nothing".

Big vocal delay

Step 2 was a delayed stereo vocal to augment the main vocal. I chose a different recording take. Importantly, I chose one that would have never made it in normal circumstances, because it was too mellow, uninteresting, and lacking energy.

Mellow can be boring but controlled and smooth. It did not reduce or hide problems in the main vocal but drove attention away from them. It took care of everything the pitch shift could not.

authors: mic