By the completion of this project, we intend to have a functional filter library that we can input a song into and get a remastered version of the song on the output. To do this we intend to build the following elements:

​

1. Further noise reduction:​

- We will fine tune and correct the initial noise filters to do a better job of reducing white noise and background noises, as well as correct the rubber banding problem that is currently plaguing the gaussian filter

​

   2. Transformation utilizing reference track:

- We will implement the reference track correction algorithm we have laid out in the progress report to sharpen the sound to make it sound crisper/sharper and not so dull using a sample from the song we want to fix and correcting it to create a transfer function. This is laid out in method 2 of the sound sharpening section on the progress report.

​

   3. Equalizer program:

- We will implement a general equalizer to filter out high, mid, and low frequencies using high, low, and band pass filters along with gain stages.

​

If time permits, we will also research and work on the following elements:

​

   1. Develop user interface to tie all filter elements into a toolbox for ease of use

- If time permits, we plan to look into generating a user interface for the functions we created so that people can easily use and play around. It can also be a cool and useful demo for people to see how DSP tools work, as well as how they interact with various songs.

​

   2. Investigate ways to algorithmically generate a reference track correction transform

- If time permits, we plan to look into seeing how we can algorithmically generate the reference track transfer function to create perhaps a more generalized mapping for various recording devices, such as gramophones, reel to reel tape recorders, etc. As laid out by  Koo Junghyun et al, machine learning would probably be the most effective method for this element.