This week's blog is on the subject of digital signal processing. To start with I had to download a file "speechtone.wav" to edit in CoolEditPro (the software used to edit sound). When I opened CoolEditPro and loaded in the downloaded sound, the wave initially had this appearance:
I then listened to the sound for the first time. There was a highly annoying background tone, which had a sort of piercing sound. The speech in it was far from precise and the overall sound resembled a radio frequency which was going out of range. I thought one way to change this would be to edit the notch filter, adding in tones of 440HZ. Below are the standard frequencies applied to sounds in the software.
Firstly, I experimented with only applying a 440Hz sound to see how that would affect the sound, to see if it could make the sound more clearer and try and remove the piercing tone if possible. Below denotes how I edited the settings of the notch filter.
The wave then changed in appearance to look like this:
This method added far more clarity to the speech contained in the sound file, but the piercing tone remained, so I then opted to try something different to improve the quality of sound. Instead of solely using a Notch Filter at 440Hz, I thought trying to add a notch filter at that frequency to the standard settings would be beneficial. I edited the notch filter as follows:
This then made the wave take this shape:
By this point, the speech is getting far clearer and the muffled tones that resembled at out of range radio have all but disappeared, but the piercing tone on top of the speech still remains, so my next move was to try and remove that. Having analysed both of the methods, I came to the conclusion that the first one was better, so I edited the notch filter settings accordingly.
No comments:
Post a Comment