CODEWITHSUNDEEP
imagecompressionjpeg
John
John

Reputation: 1596

JPEG Compression Implementation Questions

I'm trying to implement JPEG Compression (or as close to it as I can), but there are some points that I need clarity on with the actual implementation. I will explain what I currently know and where I see the issues, if anyone could clear them up that would be fantastic.

The first step is to split the image into 8x8 blocks. But I am not sure on the best way to do this, for example, what dimension array would be best to use to store all of these segments considering that it will have to have chroma down sampled and then DCT applied. Would it be 3D array (two dimensions to store 2D elements of the image and then one dimension for the color channels) and then iterate through in groups of 8 or a 4D array (with an extra dimension for storing each 8x8 group) or another method entirely.

I can then potentially see issues with the chrominance down sampling because then the size of the array would have to change size once the number of chrominance values have been reduced and these would then have to be put into the DCT which cant really take all of the different sized arrays for chrominance and luminescence at the same time.

Also is the idea of the DCT that it takes all three color channels of the 8x8 group and then converts the three values to one value thus saving space or does it take each color channel one at a time (if so I don't really understand the point how converting to Fourier space makes the compression more efficient)? Also I have noticed that the values I get for the DCT are well out of bounds of 0-255 and are instead much higher. As far as I know these values for each 8x8 block would be divided by the IJG standard quantization matrix follow by different entropy encoding.

I realize this question covers a lot of areas and is quite messy, but I can provide any additional information if required, any help would be greatly appreciated.

Upvotes: 2

Views: 1106

Answers (1)

Alex I
Alex I

Reputation: 20327

"Digital Video Compression" by Peter Symes has a chapter on JPEG, and is a good introduction to compression in general.

The reference implementation for JPEG might be a good start.

Upvotes: 2

Related Questions