lossless compression of still images - recommendations?

Joel Rees joel.rees at gmail.com
Tue Feb 21 00:19:49 UTC 2017


On Tue, Feb 21, 2017 at 7:43 AM, Aryan Ameri <public at aryanameri.com> wrote:
> On 20/02/2017 18:01, Ralf Mardorf wrote:
>> On Mon, 20 Feb 2017 11:45:48 +1100, Aryan Ameri wrote:
>>> What you are looking for is using some film compression algorithm such
>>> as h.264. Those algorithms excel at doing frame by frame comparisons
>>> and then only storing the pixels that have changed. If storage as a
>>> movie format meets your needs, it would be easy to turn the series of
>>> jpeg pictures into a h.264 encoded movie, and the resulting file size
>>> would be tiny. A few million 5MP pictures (frames) would easily fit in
>>> a 1 GiB file.
>>
>> The quality of the single pictures of even professional film cameras is
>> less good, than those of a consumer reflex camera. A film aren't
>> stills, so it's missing the details and definition a still provides.
>>
>
> Depends on your lens. With mirrorless cameras, that is not true. If you
> use a camera such as a Panasonic Lumix DC-GH5, you'll see that it treats
> video just as a series of continues photos shot at 24-60 fps.
>
> Besides, what does this have to do with my statement? I didn't say that
> the OP should use a video camera. I said the OP would see significant
> file size reductions if they turned their series of pictures into a film
> using a codec such as h.264 (IFF archival in movie form is sufficient
> for their purposes).
>
> --
> AA

And thats lossy, although, as the conversation continues, we find that
the OP is not really interested in perfectly lossless encoding. He
just doesn't want to lose too much more than what the JPEG compression
already loses for him.

Which may not be what he should be focusing on, so to speak.

-- 
Joel Rees

I'm imagining I'm a novelist:
http://joel-rees-economics.blogspot.com/2017/01/soc500-00-00-toc.html
More of my delusions:
http://reiisi.blogspot.jp/p/novels-i-am-writing.html



More information about the ubuntu-users mailing list