How to save variables in matlab in most optimal way

Hello, i have being trying to implement multiple image compression algorithms in matlab (such as RLE) but when encoding an image and trying to save the encoded matrix as a .mat file, it always becomes larger when it should be smaller thoeratically

so i want to know what other ways can you save variables such that it gets the original size it should have

thanks in advace

[dpb]

As stream (aka "binary") files. See fwrite and friends...

[Adel Hafri]

which precision should i use ?

[dpb]

What's the precison of your raw data? Use the same.

There may be another place where MATLAB is increasing the storage size that you're not expecting -- by default it turns virtually everything numeric into double precision, even image data that may only be 8-bit depth.

We would have to know the details of what you're doing to a much greater depth than you've divulged so far to know anything specific.

[Adel Hafri]

okay so i have a 750x750 jpeg pictures with values ranging from 0 to 255 and im supposed to apply losless image compression algorithms to reduce the size of those pictures, lossless image compression algorithms such as huffman work by reducing the length of frequent occuring symbols, for example if 150 was my most occuring then my huffman algorithm gives it the code 0 for example and so i ll be saving 7 bits times the frequncy of that data which means compression, the problem is matlab automatically makes that 1 bit length 0 into a 00000000 so essentialy, my algorithm is pointless since matlab will make all the data 8 bit length again, so i want a way to save data exactly the size i want, whether 1 bit,2bits.3....etc instead of it forcing all data to be 8bits

here is an example of how the algorithm changes the symbols

the picture i used isnt the best example of compression but you can get the idea

Answer 1

[Walter Roberson]

fwrite() with the 'bit1' precision.

For compression calculations purposes you should never be using the size of mat files. mat files have overhead for every variable, and mat files have their own compression built in.

[Adel Hafri]

can you please go more into detail how to use fwrite, i barely found any guids about it and it confused me alot

[Walter Roberson]

bits = {[1] [0 0] [1] [0 1 1] }
Bitstream = [bits{:}];

fid = fopen('test.bin','w');
fwrite(fid, Bitstream, 'bit1');
fclose(fid);

The sequence of bits in the file would then be 1001001 followed by an extra 0 to reach one byte.

[Walter Roberson]

Note that unless you have a fixed Huffman dictionary, then you also need to store the dictionary of bit streams and corresponding source symbols. What is stored in the file(s) needs to be enough information to recreate the original input. Just storing the encoding bits without the dictionary is not enough.

[Adel Hafri]

yes i do store my huffman dedctionary and use it in my decoding part such as here

my question is if my encoded data becomes 1001001 , how would my decoder know that it was orignally [1] [0 0] [1] [0 1 1] ?

[Walter Roberson]

You step through bit by bit.

You start with the collection of all known symbols as the current collection. You set the current offset to 0.

You read the next input bit, and you increment the current offset. You look through the current collection at the current offset and discard any entries that do not match the current bit. If the current collection is down to a single entry then you emit the corresponding symbol and go back to the start (resetting the current collection to the entire set and so on.) If you get the end of input but the collection has multiple entries then you have malformed input so stop without emitting a final output.

In practice it is more common to represent the collection as a binary tree, in which case you only need to walk the tree instead of forming sub-collections.

[Adel Hafri]

yeah man, you are a god send, your method worked perfectly thank you so much

i have just one issue now and it's that when im using bit1 for my bitstream it's saving all values as 0, while if i use 'bit2' it works perfectly, any idea why is that happening ?

it's happening even in the example u gave me

bits = {[1] [0 0] [1] [0 1 1] }
Bitstream = [bits{:}];
fid = fopen('test.bin','w');
fwrite(fid, Bitstream, 'bit1');
fclose(fid);

[Adel Hafri]

nevermind, i fixed it, i had to use ubit1 instead of bit1 for unsigned binary

again thank you so much man, you are a life saver