MOBI to EPUB conversion

[zoliky]

Hello, I tried converting a MOBI file which is about 50MB in size to EPUB using Caliber. The EPUB gets reduced to about 30MB, so I suppose there's some compression involved in the process.

I'm wondering is there a way to do a 1:1 conversion from MOBI to EPUB with no compression (or minimal compression)? Any suggestion appreciated. Thanks.


Zoltan

[mbovenka]

Originally Posted by zoliky

Hello, I tried converting a MOBI file which is about 50MB in size to EPUB using Caliber. The EPUB gets reduced to about 30MB, so I suppose there's some compression involved in the process.

I'm wondering is there a way to do a 1:1 conversion from MOBI to EPUB with no compression (or minimal compression)? Any suggestion appreciated.

The only compression that could occur is a resizing of the images in in the MOBI to a smaller size to fit your output device. Set the output profile to 'tablet' to disable this and have your device do the resizing on the fly when needed.

[kovidgoyal]

Not to mention that since MOBI and EPUB use different compression techniques, comparing their file sizes does not mean much.

[dwig]

Both classic MOBI and ePub are compressed formats. You can't have an ePub without it being compressed. An ePub file is a special case ZIP archive. Classic MOBI files are encoded much like compiled programs. As Kovid mentioned, they are different compression methods that produce differing results. Also, the file structure is different and that will have an impact on file size.

mbovenka's point comes into play if the book contains images. The output profile you choose for conversion can resize and recompress the images leading to differing sizes.

[JSWolf]

Originally Posted by dwig

Both classic MOBI and ePub are compressed formats. You can't have an ePub without it being compressed. An ePub file is a special case ZIP archive. Classic MOBI files are encoded much like compiled programs. As Kovid mentioned, they are different compression methods that produce differing results. Also, the file structure is different and that will have an impact on file size.

You can have an ePub that's not compressed. You just store the files in the ePub in store mode. But it will be a lot larger than it needs to be. I don't see any good reason to just store all the files in an ePub. Mobi is also compressed. Just not as good as ePub and if you use the higher leve of Mobi compression, you end up using a lot more processor and battery in your Kindle to extract the files.

[sjfan]

The OP is probably conflating compression with lossiness. ZIP* is lossless, à la FLAC or PNG. You don’t lose quality or information with it like you would with a lossy format (e.g. MP3 or JPG).

So as long as you’ve got the output format set correctly (as mbovenka says), then you shouldn’t worry about it.

*Technically, DEFLATE (LZ77/Huffman) compression, which is what ZIP uses in the vast majority of real-world cases

[dwig]

Originally Posted by sjfan

The OP is probably conflating compression with lossiness. ZIP* is lossless, à la FLAC or PNG. ...

Correct for ZIP archives, but not so with PNG.

PNG (Portable Network Graphic) is a bitmap format and not a compression algorithm. PNG images can be compressed or uncompressed, and if compressed can use either lossless or lossy compression. It is quite false to assume that PNGs are compressed or that if compressed the compression is lossless.

[sjfan]

Originally Posted by dwig

Correct for ZIP archives, but not so with PNG.

PNG (Portable Network Graphic) is a bitmap format and not a compression algorithm. PNG images can be compressed or uncompressed, and if compressed can use either lossless or lossy compression. It is quite false to assume that PNGs are compressed or that if compressed the compression is lossless.

This is mostly incorrect, though you're right that uncompressed PNGs are also allowed. But the standard allows only one compression algorithm, which is the (lossless) deflate algorithm. See the PNG specification:

PNG compression method 0 (the only compression method presently defined for PNG) specifies deflate/inflate compression with a 32K sliding window. Deflate compression is an LZ77 derivative used in zip, gzip, pkzip and related programs…

Some image encoders may do a lossy pre-pass to decrease file size (using things like Blurilizer and quantization filters) before they encode to PNG. But they are effectively editing the image in a lossy manner prior to encoding it as a PNG; any conforming PNG encoder/decoder will decode to a byte-for-byte identical copy of the image that was encoded.

You could obviously do lossy pre-processing on any format, and you wouldn't say that FLAC is lossy just because someone chose to do a pre-pass on the data to lower the effective bitrate before FLAC-encoding it.