Audio Quality

Why does my audio sound robotic or unnatural?

AI-generated voices have come a long way, but some combinations of text and voice can sound less natural than others. Here are some things to try:

• Switch voices. Different voices handle different types of dialogue better. Experiment with a few options.
• Add emotion instructions. When editing a script segment, add notes like "spoken softly" or "with excitement" to guide the voice.
• Use voice design. Instead of choosing a preset, describe the voice you want — age, tone, personality — and let the AI create one that fits your character.
• Simplify the text. Long, complex sentences with unusual punctuation can trip up voice generation. Break them into shorter, more natural phrases.

Why is there background noise in my sound effects?

AI-generated sound effects can sometimes include subtle artifacts or unintended background sounds. This is a normal part of the generation process.

To get cleaner results:

• Be more specific in your prompt. Instead of "explosion," try "single large explosion with debris, no echo, clean recording."
• Describe what you don't want. Adding "no background noise" or "isolated sound" can help.
• Regenerate. Each generation produces a slightly different result. If the first attempt has artifacts, try again.

How do I make dialogue sound more natural?

Great dialogue audio comes from a combination of good writing and the right settings.

• Add emotion and delivery cues to your script segments — "whispered," "shouting," "sarcastically," etc.
• Vary the pacing. Mix short and long sentences. Let characters interrupt each other or trail off.
• Use pauses. Adding "..." or explicit pause notes gives dialogue room to breathe.
• Apply voice effects. A phone filter on a phone conversation or slight reverb on a large room scene adds realism.

Can I improve music quality?

Yes. The quality of generated music depends heavily on how specific your prompt is.

• Specify genre and mood. "Melancholic acoustic folk ballad" is much better than "sad music."
• Mention instruments. "Solo piano with light strings" gives the AI something concrete to work with.
• Include tempo hints. "Slow tempo, around 70 BPM" helps set the right pace.
• Keep vocal sections short. For best quality, vocal music segments should be under 60 seconds. Instrumental pieces can be longer.
• Iterate. Regenerate and compare results. Small prompt changes can produce very different music.

Why does the same prompt give different results?

AI generation is inherently variable — the same prompt will produce different outputs each time. This is by design and is actually a feature, not a bug.

If you get a result you love, it's saved in your project. If a result isn't quite right, just regenerate that segment. You can regenerate individual segments without affecting the rest of your project.