How to reduce the bit depth of the bmp image from 8-bit to n-bit, where n<8 andn>=1. Choose n=3.

-bit depth reduction -what function to be used

[Eleanor Betton]

I have an application where a peice of kit will only read 4 bit images. Is there a workaround to save to a 4 bit bitmap? Currently I have to converst through photoshop which is time consuming.

[Image Analyst]

I don't know of any standard way to write 4 bit images unless you just wrote out raw data to a binary file with fwrite(). I don't even see a way in Photoshop to do it. You might start a new question if you can get more information about what format this "peice of kit" expects to see, like what format/extension, what kind of header, etc.

[Mayan Rumbak]

I need something similar...

I have a 4-bit depth BMP that I would like to read - manipulate - and save back as BMP with 4-bit depth.

but I don't know how to do it...

Does anyone have an idea?

[Image Analyst]

Yes:

rgbImage = imread(filename); % Read in.
% Code to manipulate - whatever that means.
% Make 4 bit
if max(rgbImage(:)) > 15 % If it's more than 4 bits now, rescale
    rgbImage = rescale(rgbImage, [0, 15]);
end
% Save back out.
imwrite(outputFileName, uint8(rgbImage));

Answer 1

[Image Analyst]

Just divide the image by a power of 2. For example to convert an 8 bit image to another, lower bits image:

image7bit = image8bit / 2;
image6bit = image8bit / 4;
image5bit = image8bit / 8;
image4bit = image8bit / 16;
image3bit = image8bit / 32;
image2bit = image8bit / 64;
image1bit = image8bit / 128;

[Image Analyst]

sunnie, this does not change the number of bits to n. It changes the number of colors to n. The resulting image is an indexed image - an array where the values go from 1 to n. The image is an 8 bit image but the values go from 1 to 4 in your case. 4 can be represented with 3 bits (4 in binary is 100). Then there's a color map with it that specifies what each RGB color will be for the values 1, 2, 3, and 4. If you specified 6 you'd get 6 colors, not 64 which is what 6 bits would be. Why do you think you need 4 bits? Why did you not take my advise? You do know, right, that even if the values are less than 8 bits, the values are stored in a uint8 variable which means it will be 8 bits? You can only get 4 bits if you store it in a compressed manner, like with imwrite and tell it to use PNG or some other compressed format. And you would not be guaranteed of getting exactly 4 because you don't really have control over how imwrite() does its compressing. You could do it yourself but I don't want to get into that because I don't think it's necessary for you.