Futz is the post-production word for making something sound like it is coming out of a small bad speaker. The filtering is the easy half, and most people stop there, which is exactly why their phone calls sound like a plugin rather than a phone.
Numbers you can dial straight in, plus the part nobody mentions.
Every one of these devices is defined mostly by what it throws away. Start with a high-pass and a low-pass at these points, then adjust.
| Source | High-pass | Low-pass | Character |
|---|---|---|---|
| Landline telephone | 300 Hz | 3.4 kHz | The classic. That band is the actual telephone standard, not an approximation |
| Mobile, normal call | 300 Hz | 3.4 kHz | Same band, plus codec artefacts and the odd dropout |
| Mobile, HD voice | 100 Hz | 7 kHz | Noticeably fuller. Use this if you want the call to read as modern |
| Small TV or clock radio | 200 to 400 Hz | 5 to 8 kHz | Add a resonant peak: the little cabinet rings somewhere around 800 Hz to 1.5 kHz |
| Laptop speaker | 400 Hz | 10 kHz | Thin rather than muffled. Keeps the top, loses everything underneath |
| PA or megaphone | 400 Hz | 4 kHz | A strong honk around 1 to 2 kHz and real distortion. Horns are peaky by design |
| Intercom or walkie | 400 Hz | 3 kHz | Heavy compression, a squelch of noise at the top and tail of each transmission |
Use steep filters here, 24 dB per octave or more. This is the opposite of the advice for sound through a wall, and for a good reason: a wall is a gentle physical roll-off, whereas a phone line is a designed electronic band limit with hard edges. Making a phone sound right means being ruthless.
Every one of these devices squashes dynamics aggressively, either deliberately or because it runs out of headroom. A phone call has almost no dynamic range. Set a high ratio and pull several dB, then turn it back up. Flat and loud is correct here.
Small speakers and cheap amplifiers distort. Not much, just enough to add grit on the loud syllables. Saturation rather than fuzz.
A modern mobile call does not just sound band-limited, it sounds slightly granular and metallic, and it occasionally drops out entirely. A very short dropout, a fraction of a second, once in a call, does more for believability than any amount of EQ. Bitcrushing gently in the top octave gets you close to the texture.
A landline has a faint hum and hiss. A walkie has a burst of squelch. A PA has room noise around it. Silence between words is the giveaway that it is a plugin.
You can get every number above right and the result will still sit oddly in the scene. The reason is simple once you see it: a real speaker is in a room, and your futzed voice is not. It is perfectly dry, sitting in the middle of the stereo field, with no reflections. The ear knows that is impossible.
So after the futz, put it back in the space. A short reverb matching the room the device is in, panned to wherever the device is on screen, and the level dropped so it reads as being over there rather than in your head. This one step does more than everything before it.
The technique is called worldizing. You play the futzed sound out of a real speaker in a real space and record it again with a microphone. Walter Murch used it on American Graffiti, and it still beats any plugin, because everything the room does is genuinely there. If you have a speaker and a recorder and an afternoon, it is worth doing once just to hear the difference.
If we see the device, the sound should move with it. A radio at the far side of the room when the camera is close should be quieter and duller than the same radio in a close-up. Static treatment across a scene where the camera moves is the second most common giveaway after the dry-signal problem.
None of this needs anything you do not already have. In Premiere: Parametric EQ for the band limits and the peak, Dynamics for the compression, Distortion for the grit, Studio Reverb for the room. In Final Cut: Channel EQ, Compressor, Distortion and Space Designer. In Resolve: the Fairlight EQ, the dynamics section on the channel strip, and the reverb effect.
Build it once on a bus and save it. You will use it on every project.
Getting the filtering right takes ten minutes. Getting the sound to sit somewhere real, and to keep sitting there as the shot moves, is the part that eats the afternoon.
THALM handles that half. You place the sound in a room on a building cutaway, place the listener somewhere else, and the distance, walls and doorway do the work. It also models how the sound is being played in that space, including a phone speaker, a PA horn and a gramophone, so the futz and the placement come from the same decision rather than from two plugins that do not know about each other.
We write these when there is something worth writing down. One email when a new one lands or a new Tunary instrument ships. No newsletter, no schedule.
Tunary