spectrumtransfer: making two recordings sound like one

You know the situation. Two takes from the same video shoot. One sounds warm and clear. The other sounds thin, a bit dull, maybe slightly harsh, because the mic was different, or the person sat somewhere else. Both are usable. Together, they don’t feel like the same production.

I kept fixing that by hand, again and again. So I built a tool for it.

The idea

spectrumtransfer compares the tonal balance of a good reference with the recording that needs help, and calculates the difference:

EQ correction = reference spectrum − target spectrum

That correction is applied as a static EQ. It’s not magic and it’s not adaptive. It simply makes the weaker take sound like it came from the same setup as the good one.

The whine problem

The second thing it does turned out to be just as useful. A lot of recordings have a thin, high electrical whine from screens, lights, power supplies or camera electronics. The voice is fine, but that tone is always there.

The whine mode removes exactly that, with a gentle high cut and narrow notch filters, and leaves the rest of the voice alone.

Made to be used, not configured

I think tools like this only get used if they’re easy. So there’s one launcher on macOS that walks you through the steps: auto level, tone transfer, de-esser, a bit of clarity, peak control and whine reduction. There’s also a fast mode with sensible defaults. It works on single files or whole folders, WAV or MP4, and writes down the settings it used next to every result.

It’s a small, very practical thing. But it saves me time on every video project, and honestly, that’s my favourite kind of tool.