My arms have been better this week so I completed an edit of my film. I was very pessimistic about how my film would turn out after shooting as I was not happy with the footage and felt lots of it could and should have been done better. It was useful, therefore, to have a bit of time after shooting to distance myself from the project. When I came back to the project, I was able to do so as an editor rather than a director, and instead of wanting to reshoot every scene my rejected instead was to make the film work as best I could with the footage I had. This meant being very creative and also ruthless; some lines of dialogue had to be lost as I was not happy with any of the takes. Once I had my edit together I was surprised by how much I liked it.
Unfortunately, the audio was very poor quality, largely because I shot my film in a noisy classroom and not a proper sound studio. During the shoot I had two mics recording at all times, one going straight into the camera and the other processed by Nvidia RTX, an AI noise removal. I thought the noise removal would fix the problem was it a noisy studio but unfortunately it often collect and distorted the actors voices, making it unusable. The audio recorded by the camera what’s far clearer, but also very noisy. To fix this I first experimented with various AI noise remove this like runway ML and Adobe podcast. Can use, however, why no good as they applied a blanket effect that did not take into account the level of noise or type of noise, which left then audio sounding distorted. I then tried the premiere pro noise reduction feature. This was better as I had more control over the amount of noise to remove, but after some research, I learned that Adobe audition has a very good noise remover. After some experimentation I was amazed by how powerful the Adobe audition know its removal was. It works by the use of making a manual selection of a piece of audio containing only noise and then select how intense the noise cancellation should be. It also allows he’s had to make a specific adjustments to different waveforms say different levels of noise can be targeted. This made most of my audio usable but there are still some lines which are heavily distorted by the process.