Converting from lossy to lossless codec

[sebelk]

Hi,

I have a file with 3gp and inside of it a video stream in MPEG4 format and audio stream in amr_nb format.
Just to play, I've converted video to huffyuv. I've noted that file size grew and as it is expected the actual quality didn't change.
Perhaps it sounds somewhat silly, but I'd want to know why the bitrate, fps are increased... if after all, some data are lost, and you cannot come back to the "original"

Is it right my rationale? Does make sense in some case convert from lossy to lossless codec?

Thanks in advance!

[hech54]

Hi,

I have a file with 3gp and inside of it a video stream in MPEG4 format and audio stream in amr_nb format.
Just to play, I've converted video to huffyuv. I've noted that file size grew and as it is expected the actual quality didn't change.
Perhaps it sounds somewhat silly, but I'd want to know why the bitrate, fps are increased... if after all, some data are lost, and you cannot come back to the "original"

Is it right my rationale? Does make sense in some case convert from lossy to lossless codec?

Thanks in advance!

This isn't a "ZIP" file. Once the quality is gone, it is gone for good.

[smrpix]

First, if fps increased you did something wrong. It should remain the same.

There are two good reasons for using a higher bitrate lossless file, 1) to do further processing -- color correction, etc while minimizing further quality losses. 2) To make it easier for editors to handle creating a smoother experience. With i-frame only editing you trade increased filesize for reduced load on the CPU.

http://en.wikipedia.org/wiki/Video_compression_picture_types

[sebelk]

First, if fps increased you did something wrong. It should remain the same.

There are two good reasons for using a higher bitrate lossless file, 1) to do further processing -- color correction, etc while minimizing further quality losses. 2) To make it easier for editors to handle creating a smoother experience. With i-frame only editing you trade increased filesize for reduced load on the CPU.

http://en.wikipedia.org/wiki/Video_compression_picture_types

Thanks are you talking about of advantages of creating a lossless file from stratch, aren't you?
I ask if there is an advantage in converting from an existing lossy file to a lossless file, I see that only increase its size, but I wonder why/how does it, it extract data from compressed (lossy file) even when some bytes was descarted in order to make that file smaller?

Thanks in advance!

[hech54]

I ask if there is an advantage in converting from an existing lossy file to a lossless file

No. Like I said before.....this isn't a ZIP file. All you are doing is making a small-sized crappy looking file into a large-sized crappy looking file.

[poisondeathray]

Thanks are you talking about of advantages of creating a lossless file from stratch, aren't you?
I ask if there is an advantage in converting from an existing lossy file to a lossless file, I see that only increase its size, but I wonder why/how does it, it extract data from compressed (lossy file) even when some bytes was descarted in order to make that file smaller?

Lossless compression is with respect to uncompressed video (ie. the term lossless is in terms of the source video being decoded to uncompressed data) . So in that regard, lossless compressed files are much smaller than their uncompressed counterpart

The difference is it's decoded to uncompressed first, then re-compressed with a lossless codec. If you're familiar with audio, the analogy would be flac

In contrast, something like file archiving utilities like zip, winrar, 7zip, the video is not decoded to uncompressed video, so the original compression remains

[smrpix]

Thanks are you talking about of advantages of creating a lossless file from stratch, aren't you?

No. I'm referring to your situation and answering why one would bother converting lossy to lossless. (Though it works from scratch too.)

[docbillnet]

Perhaps it would help to use a visual analogy. Suppose you had news paper article you wanted to save. So you stuck it on a photo copy machine and made a copy. If you have a high quality copy machine, it will look close to the original, but never exactly the same. For example, maybe there was a photo that was dithered, and the letters became blurry and smudged. Now lets say you turn around and decide you need a perfect copy. If you make a perfect copy of your "copy" it will contain the exact same dithering, the exact same blurryness, and the exact same smudges. All of these "defects" in you copy are actually additional data. Crisp clear sharp lines actually take less data to store losses than blurry smudgy lines. The reason being is there are many ways a line can be smudged, while there is only one way for it to be crisp sharp and clear. So you will find your lossless copy of the copy actually needs to store significantly more information than a lossless copy of the original.

[jagabo]

I ask if there is an advantage in converting from an existing lossy file to a lossless file

Just for watching? No.

I see that only increase its size, but I wonder why/how does it, it extract data from compressed (lossy file) even when some bytes was descarted in order to make that file smaller?

The process of recompressing with a lossless codec involves first decompressing the lossy codec back to uncompressed frames. During that decompression something that hopefully looks like the lost video data is substituted in its place. The lossless codec has to include that new data.