Hey folks, new Murf user here and really impressed with the voice quality overall! I've been using it to generate narration for some internal data pipeline training videos.
However, I've hit a snag. I'm working with a "Consultant" voice for a longer explainer section, and I'm finding the natural 'breathing' sounds a bit too pronounced and distracting in the final audio. It's pulling attention away from the actual content about API sync patterns.
Is there a way to adjust the level of these breathing sounds, or the "naturalness" of the speech, within Murf itself? I've poked around the voice settings and the SSML options but haven't found a clear parameter for this. If it's an SSML tweak, a quick example would be super helpful.
If this isn't adjustable, does anyone have a good workaround? Maybe a specific post-processing step or a known setting in a tool like Audacity that can target those frequencies without making the voice sound robotic?
Thanks in advance for any pointers. Trying to get this training module ready to ship! 😅
ship it
ship it
The breathing artifacts are a known side effect of some of Murf's more naturalistic voice models, and they don't expose a direct control for that specific parameter. You've already looked in the right places.
A targeted workaround I've used is a combination of post-processing:
* First, apply a high-pass filter (around 80-100 Hz) to remove any rumble that might be amplifying the breath perception.
* Then, use a spectral editing tool, like the "Spectral Edit Multi-Tool" in Audacity or a similar feature in a DAW, to manually select and reduce the gain on just the visible breath sounds between phrases. This is more precise than a general EQ notch and preserves the voice's timbre.
It's a bit manual, but for a final training module, it gives you full control without sacrificing the voice quality you started with.
Your bill is too high.