planning: Ichigo Speech Enhancements

[dan-homebrew]

Goal

• Baseline enhancements, i.e. noise reduction
• Processing step: pre-model before Ichigo

[tikikun]

should we use current speech enhancement tech to tackle this instead of developing our own?

[PodsAreAllYouNeed]

We should very much use out of the box speech enhancement tech. The problem is general enough that the ichigo use-case is not differentiated enough to warrant domain-specific fine-tuning.

[tikikun]

@PodsAreAllYouNeed we can icebox this if this is not priority

[tikikun]

also cc @dan-homebrew