There are two honest ways to make your own sound effects: record them, by performing the sound with props in front of a microphone, or generate them, by describing the sound in enough physical detail that a model can build it. Both give you a cue nobody else has used.
Which one to reach for is usually decided by the sound itself rather than by your preference.
Record it or generate it
| The sound you need | Record | Generate |
|---|---|---|
| Footsteps, cloth, prop handling | Yes | No |
| Something with a specific room around it | Yes | No |
| A creature, engine, or weapon you do not own | No | Yes |
| A cue you need in five minutes | No | Yes |
| Twenty variations of the same hit | No | Yes |
| Anything unsafe or destructive to stage | No | Yes |
What foley recording is
Foley is the craft of recording everyday sounds to match action on screen: footsteps on a surface, cloth movement, a door, a cup set down on a table. You record it yourself with a microphone, performing the sound in time with the picture or capturing it clean to place later. It gives you complete control over texture and performance, and it can capture the specific object or space you actually have in front of you. The cost is setup: a quiet room, a microphone, the right props, and the time to record and clean each take.
What generating from a text description is
The other route is to describe the sound you want and generate it. On Morphic you write the cue in plain language (the object making the sound, its texture, the space around it, roughly how long it runs) and get an original effect made to fit, using models such as ElevenLabs and Seed Audio. There is no recording setup and no prop hunt, and you can produce a sound you could not easily record, such as something large, dangerous, or imaginary. You can generate several takes and pick the one that fits, then download it for your editor. Effects you generate are yours to use commercially on paid plans, so check the current terms.
When to reach for each
Neither route is always right. A quick way to choose:
- Record foley when you have the object and a quiet space, the performance timing matters, or you want the very specific texture of a particular prop. It rewards setup with total control.
- Generate on Morphic when you need the sound quickly, cannot record it (too big, too loud, or not physically available), or want to audition several options fast. It trades the hands-on control of foley for speed and reach.
- Use both on a real project. Record the handful of sounds that need a specific performance, and generate the rest, including the ones that would be impractical to capture.
Browse the Morphic sound library to hear the range of cues you can describe, and see how to add sound effects in Premiere Pro for placing a finished sound in an edit.
How to describe a cue so the result fits
Whether you are briefing yourself for a foley take or writing a prompt to generate one, the same four details make a sound specific:
- The source: what actually makes the sound. "A wooden chair" is clearer than "a creak."
- The texture: how it feels. Heavy or light, sharp or soft, wet or dry.
- The space: where it happens. A small room, a stone hallway, and open air all sound different for the same action.
- The length: roughly how long it lasts, including whether it has a short tail or cuts off cleanly.
Put together, "a heavy wooden chair dragging on a tile floor in a small room, short and dry" gives you a sound cut for the moment instead of one you have to force. Vague cues get vague results; specific cues get usable ones.
What description actually buys you
The same intended sound, asked for three ways. Nothing changed but the words.Better
0:04Source and action named. Closer, but the texture and the ending are still open.
If you would rather start from a finished cue and change one word, the footstep foley, cloth foley and magic and fantasy cues pages each open in Studio with a working prompt already in the box.
FAQs
Only if you record foley. That route needs a microphone, a quiet space, and props. If you generate the sound from a text description on Morphic instead, you need none of that; you describe the cue and download the result.
Yes. Generating from a text description produces an original effect made to your brief rather than a stock file pulled from a shared library, so the sound is specific to your moment.
Generate three and audition them against the moment rather than on their own. Sound effects vary more between takes than images do, and the take that stands out in isolation is often not the one that disappears correctly into a scene.
Sounds tied to a specific prop or a precise performance timing often reward recording, since you control every nuance. Sounds that are impractical to capture, such as large, loud, or imaginary ones, are usually faster to generate.
Name the source, the texture, the space, and the rough length. "A short metallic latch click in a small room" gives the model far more to work with than "click," so the result fits the moment.
Yes. The same Morphic workspace that generates sound effects also covers music and voiceover, so a project's audio can come from one place instead of several tools.
