Is there a reason the makers decided to choose this other basis function instead of using the fourier transform that the professor mentioned could be a possible method as well? Also is jpg similar to jpeg or is the similarity in the name not representative of the true difference/similarity.
TiaJain
Given that the DCT is applied separately to different color channels with different resolutions (as mentioned for the Y'CbCr channels), how does this affect the final image quality, especially when it comes to high-frequency details in the chroma channels?
buggy213
i think part of the reason is that DCT will only have real-valued coefficients after the transform, whereas the DFT might require working with complex numbers
spegeerino
I had heard that JPEG compression worked by "frequencies," and I guess those frequencies are the frequencies that get picked up by the discrete cosine transform (similarly to a DFT i imagine)
jananisriram
I find it interesting to see how different file formats have different types of interpolations and transforms, making them different pixelated structures.
Is there a reason the makers decided to choose this other basis function instead of using the fourier transform that the professor mentioned could be a possible method as well? Also is jpg similar to jpeg or is the similarity in the name not representative of the true difference/similarity.
Given that the DCT is applied separately to different color channels with different resolutions (as mentioned for the Y'CbCr channels), how does this affect the final image quality, especially when it comes to high-frequency details in the chroma channels?
i think part of the reason is that DCT will only have real-valued coefficients after the transform, whereas the DFT might require working with complex numbers
I had heard that JPEG compression worked by "frequencies," and I guess those frequencies are the frequencies that get picked up by the discrete cosine transform (similarly to a DFT i imagine)
I find it interesting to see how different file formats have different types of interpolations and transforms, making them different pixelated structures.