EDIT: SEE MY THIRD POST FOR SUMMARIZED CONCEPT
I was up until the early morning looking at ways to potentially compress a file.
I have a basic concept of the huffman coding algorithm, which looks for repetition of bytes / patterns in a file, coding the most repetitive patterns with the lowest amount of bits and the least repetitive patterns with the highest amount of bits, in order that the most repetitive patterns will ideally consist of a larger majority of a file and will be represented in few bits.
What I came up with is slightly different.. It is based on repetition to a degree, but is slightly different to what the huffman algorithm looks for in a sense.
The idea is that there is a 50% chance that a file can be reduced up to 25% of the original file size.
It works on frames, which consists of headers and data, and it is potentially possible to run a tweaked huffman coding algorithm on the data portions of the frames to attempt to reduce the file size even more.
On top of that, as there is a 50% chance that a file can be reduced up to 25%, it is possible to run the compression algorithm on the produced file to reduce the size even more.
This idea should generally work better on 'random garbage files' than what the huffman coding algorithm can do.. i.e. random data files which will not compress well with something like WinRar should still have a 50% chance to reduce the file size up to 25%.
I am not good a programming, but I believe the concept is solid.
I am wondering if there is someone who would perhaps be able to help me to produce the program or have a look at my concept for constructive criticism.
If anyone is interested in being able to help out with this idea, I can post up my concept.
I was up until the early morning looking at ways to potentially compress a file.
I have a basic concept of the huffman coding algorithm, which looks for repetition of bytes / patterns in a file, coding the most repetitive patterns with the lowest amount of bits and the least repetitive patterns with the highest amount of bits, in order that the most repetitive patterns will ideally consist of a larger majority of a file and will be represented in few bits.
What I came up with is slightly different.. It is based on repetition to a degree, but is slightly different to what the huffman algorithm looks for in a sense.
The idea is that there is a 50% chance that a file can be reduced up to 25% of the original file size.
It works on frames, which consists of headers and data, and it is potentially possible to run a tweaked huffman coding algorithm on the data portions of the frames to attempt to reduce the file size even more.
On top of that, as there is a 50% chance that a file can be reduced up to 25%, it is possible to run the compression algorithm on the produced file to reduce the size even more.
This idea should generally work better on 'random garbage files' than what the huffman coding algorithm can do.. i.e. random data files which will not compress well with something like WinRar should still have a 50% chance to reduce the file size up to 25%.
I am not good a programming, but I believe the concept is solid.
I am wondering if there is someone who would perhaps be able to help me to produce the program or have a look at my concept for constructive criticism.
If anyone is interested in being able to help out with this idea, I can post up my concept.