I uploaded a picture and it listed some sounds that might accompany the picture like "Water flowing in a stream nearby." It took a couple of minutes to generat

I uploaded a picture and it listed some sounds that might accompany the picture like "Water flowing in a stream nearby." It took a couple of minutes to generate 6 seconds of poor sounding audio. Seems like using canned sounds would be better. That'd be kinda cool. Use AI to identify what sounds might be applicable to a scene, let the user select all the sounds which are applicable and add more, and then play the selected sounds with mixer board controls for volume and frequency. Save the result as an mp4.