Why JPEG?

[saynomore]

I've found I can make as many adjustments and to the same degree
with JPEG High as I can with Raw.

I take it you don't shoot much RAW or know the meaning of "shoot to the right", or appreciate being able to change the native WB and color settings of your picture.

Don't get me wrong, I think jpeg is a good compromise, but a compromise anyway. RAW is uncompromised, full-potential data. Even more than uncompressed, 16-bit TIFF. Jpeg can get very high quality, but I would only treat it as a final image format, ready to send to print or web.

 

[misha]

TIFF offers non-lossy compression and Olympus support TIFF files.

Compression? Those TIFF files in Olympus were huge, much bigger than RAW - so what's the point? And of course they locked up those weak processors for half a minute or so
--
Misha

 

[Mark B.]

Guess you've never set a wrong WB - such as flourescent while shooting outside. I'd much rather deal with that as a raw file, which can be fixed with one click.

Mark

 

[Kjeld Olesen]

While apart from the missing EXIF capabilities of PNG this format would be preferable over JPEG due to its support for 48 bits per pixel, I think JPEG2000 would provide the best compromise between compression level and data loss.

--
Kjeld Olesen
http://www.acapixus.dk

 

[janaco]

But, as I understood, the JPEG2000-Algorithm is told what SIZE the Image should get, where you tell the JPEG-Algorithm what QUALITY the Image should be! But how can the camera judge, what size would be appropriate to match the detailes the actual Shot has?
Or did I miss something with JPEG2000?
--
Canon EOS 33/300D/30D

 

[Kjeld Olesen]

But, as I understood, the JPEG2000-Algorithm is told what SIZE the
Image should get, where you tell the JPEG-Algorithm what QUALITY
the Image should be! But how can the camera judge, what size would
be appropriate to match the detailes the actual Shot has?
Or did I miss something with JPEG2000?

Well, what you describe is certainly not the only difference between the two formats. The JPEG2000 uses an entirely different form of compression. As you can see here

http://www.dpreview.com/news/0108/01080401lurajpeg2000.asp

a 284 KB image JPEG2000 has about the same quality as a 1,135 KB JPEG image. This would be the benefit of JPEG2000 over JPEG - that you could either have more images at the same quality or the same ammount at better quality on the same CF-card capacity.

The backside is that JPEG2000 is far more processor demanding, but since this would be hardware rather that software based I do not think it should pose any problem.

--
Kjeld Olesen
http://www.acapixus.dk

 

[avidday]

Apart from the computational complexity of the wavelet compression scheme,

some of the core elements of JPEG2000 are encumbered by patents, which almost guarantees none of the big camera manufacturers will use it.

 

[Kjeld Olesen]

BTW: wouldnt it be possible to store the RAWs as 16-Bit-TIFF-Gray?
So you save Interpolation-Information within the EXIF? So the Image
uncompressed would "only" be 8MPix*2 = 16MBytes? Maybe (lossless)
compressed we would get 10MByte...

That is actually not a bad idea I think. That would make it a LOT easier for software developers to read the RAW data, and make their own bayer interpolation algorithms - and especially, to apply whatever manipolation to the each color chanel BEFORE the bayer interpolation is performed.

I do not think that it would compress much however, since neigbouring pixels now would represent diferrent color chanels.

Unfortunately, I do not think that we will ever see this in action.

--
Kjeld Olesen
http://www.acapixus.dk

 

[Kjeld Olesen]

... ,
some of the core elements of JPEG2000 are encumbered by patents,
which almost guarantees none of the big camera manufacturers will
use it.

So is JPEG but PNG is not.

--
Kjeld Olesen
http://www.acapixus.dk

 

[saynomore]

You're at the mercy of only one level of compression, which is not much for regular pics. It's the same difference between mp3's and lossless Windows media or FLAC for example. The lossless codecs usually give you 2:1 compression or less. For PNG, you'd still get a file considerably bigger than a RAW file.

... ,
some of the core elements of JPEG2000 are encumbered by patents,
which almost guarantees none of the big camera manufacturers will
use it.

So is JPEG but PNG is not.

--
Kjeld Olesen
http://www.acapixus.dk

 

[Kjeld Olesen]

You're at the mercy of only one level of compression, which is not
much for regular pics. It's the same difference between mp3's and
lossless Windows media or FLAC for example. The lossless codecs
usually give you 2:1 compression or less. For PNG, you'd still get
a file considerably bigger than a RAW file.

As I recall PNG has the posibility og no compression, loss less, and lossy compression - at 24 or 48 bits.

--
Kjeld Olesen
http://www.acapixus.dk

 

[B0b]

JPEG is lossy (unless you use 100% quality).

actually any Quality format of JPEG is Lossy!
--
http://www.pbase.com/bob_hall/
http://www.graphic-artwork.co.uk

 

[PaulyOly]

I know you mean you lose information, but I honestly can't see it even in large prints.

--

http://www.pbase.com/paulyoly/root

 

[FidelCastro]

I don't have any problems adjusting white balance in JPEG

 

[PaulyOly]

in prints? I can't see it even with a compression of 10 as opposed to the highest quality of 12 in elements.

--

http://www.pbase.com/paulyoly/root

 

[MitchAlsup]

I've found I can make as many adjustments and to the same degree
with JPEG High as I can with Raw.

You may think this, but it is not true.

With RAW you get a full 12-bit data-set. With JPG you get 8-bit gamma corrected data-set. Where does this come into play? Whne you want to extract the subtle detail deep into the shadows. JPG only uses 8 counts to represent all of the information from 8 stops below saturation to 11.2 stops below saturation. For this same space, RAW uses 64 counts. If you ever end up underexposing images, RAW can enable a lot better image to be created from the underexposed starting point.

With complete 12-bit numbers for R, G, and B channels, the White balancing allows more sublte tonality shifts without any posterization effects that sometimes creep into JPG white balance adjustments.

Perhaps you are simply not critical enough of your post processing work to 'see' the subtle differences a bigger data-space can deliver, and thereby don't know what you are missing.
--
Mitch

 

[Ominous]

The tonal range is going to be different if you do your WB and get it correct in RAW vs playing with curves in JPEG
--
http://www.whalenphotography.net

 

[gilesjuk]

If you read the post, if you're a RAW shooter than you aren't making any use of the internal processing parameters of the camera?

Picture styles, parameters etc aren't applied to RAW.

 

[janaco]

I did some tests in ordering Printouts of my 300D pictures in Low-JPEG sized 50x75cm - thats normally far to big for 6MPix. But the results where so good that only when using a magnifying glass I could identify the JPEG artifacts.

Now after that experiences and with the 8MPix 30D I decided to "shoot" mostly in Low JPEG (sharpness 2) and "take fotos" in RAW only when I know that I want to postprocess the Image (i.e. because of high contrasts etc.)
--
Canon EOS 33/300D/30D

 

[TimR32225]

I know that in theory raw is better than jpg, but I have yet to use it very much because I can see little, if any difference. I shoot jpg most of the time, and sometimes shoot a few raw files and archive them for later.

For me, it's not worth using 3 times the memory and disk space with little or no payback. Maybe I'll see it some day, but I don't see it yet.
--
pbase gallery at: http://www.pbase.com/tim32225