nanoStudio Audio Lab — denoise / EQ / dynamics

1 Source & analysis

Audio or video file

Any container ffmpeg reads — the audio track is extracted server-side.

Clip length seconds analysed

Derives the chain from the clip's own measurements — noise floor, spectrum, loudness, crest — and explains every choice.

Load a source, then press Analyse source — nothing here is guessed, it is all measured from the clip.

2 Mixer — active preset: custom

the rack shows the filters you reach for; a preset reveals anything extra it uses

Raw args — every native option of each filter

Reaches options with no fader (afftdn noise_type, agate knee, alimiter asc …), validated against the option table read out of the ffmpeg binary. Typing here overrides that strip's fader until you press reset.

ffmpeg chain

No filters enabled — bypass.

payload (native filter args)

{}

3 Process & compare— load a source first

export gains (dB)

(bypass)

measurements

	dry	wet	Δ

Press Process to measure dry vs wet.

Crest (peak − RMS) is the one that moves when the compressor or limiter works — RMS alone cannot see them. None of these can hear artefacts: for that, listen.

energy in the bands this chain targets

band	dry	wet	Δ
—	—	—

Watch this rather than overall RMS for hum, buzz and rumble: a tone 8 dB under the voice is well under a dB of total energy, so a perfect notch barely shows up above.

Hum reads as horizontal lines, a gate as vertical gaps, a denoiser as a lifted floor.

live analyser — the audio you are hearing

— · space flips dry/wet without losing your place

Download the processed clip at matched gains, to compare offline:

Process a clip first.