Bryan -

Wiser heads will probably correct or refine this, but ...

The 'raw' file is supposed to represent data very close to what is provided by the camera's sensor array.

Before saving images in non-raw formats (JPEG, etc.) the camera performs some conversions of the image field as well as - usually - applying some compression. These conversions are irreversible and represent a compromise intended to give pleasing images of common subjects.

This is crudely analogous to the processing applied to color negatives by most photo processors as they are printed - you have probably seen the quite surprising results this sometimes gives: colors gone crazy!

UFRaw gives you the chance to make different assumptions and adjustments, and keeps the original image file untouched in case you want another chance. It also provides as default a compensation similar to what the camera would have done if left to its own algorithms. That's what you get if you don't intervene.

If you aren't making any adjustments in UFRaw you may be just as well-off to work with [say] JPEG images out of your camera. The intermediate stage of adjusting your image in UFRaw is one of the main reasons for using RAW format in the first place.

At least that's my [mis?]understanding.

- Mills

On Wed, 30 Sep 2009, Bryan wrote:

For starters, is there any good reason why UFRaw should be used for anything other than opening a RAW file before transporting it to Gimp?

What I really wanted to ask was, Why, when open up an image in UFRaw and it looks great, even when zoomed in as much as allowed. However, when you sent it to Gimp via UFRaw, when you zoom into even a small amount the image is very pixelated. When or how did it become downsized or some sort of processing that might cause this to happen? Is it possible to bring in a RAW file and retain it's sharpness?

Thanks for anyone's help in regards to this.

On 2009-09-30, John Mills wrote:

If you aren't making any adjustments in UFRaw you may be just as well-off to work with [say] JPEG images out of your camera. The intermediate stage of adjusting your image in UFRaw is one of the main reasons for using RAW format in the first place.

This is as far from being true as one can get.

A hint: a camera uses as-much-pessimized-as-possible computer, and does RAW-->JPEG conversion in a fraction of a second. RAW converter most probably uses hundreds times more powerful hardware, and takes up to minute(s) to perform convertion.

Guess how the results compare...

Hope this helps, Ilya

On 2009-09-30, John Mills wrote:

If you aren't making any adjustments in UFRaw you may be just as well-off

to work with [say] JPEG images out of your camera. The intermediate stage

of adjusting your image in UFRaw is one of the main reasons for using RAW

format in the first place.

This is as far from being true as one can get.

A hint: a camera uses as-much-pessimized-as-possible computer, and does RAW-->JPEG conversion in a fraction of a second. RAW converter most probably uses hundreds times more powerful hardware, and takes up to minute(s) to perform convertion.

Guess how the results compare...

Hope this helps, Ilya

I'll embellish the OP's original question for further discussion.

Obviously, if you adjust in UFRaw, your 'range' of adjustments is much wider than what would be available if you were starting straight to GIMP with a camera generated JPG.

In the spirit of the OP's question, if you make no adjustments in UFRAW, is there any more latitude for adjustment in the resultant JPG file (in Gimp or other editing application) than what you might get straight from the camera?

Personally, I use UFRaw to fine tune the exposure (I like those blinking whatchamacallits) although, most of the time, having shot in RAW and adjusted exposure in the camera according to the histogram view, I generally already have the best exposure balance so that, if there is any blinking in UFRaw, it will show up for both overblown highlights and underexposed shadows. I might do a little noise control on some shots, and that's about it. The rest, in my view, is better handled in GIMP.

It feels as though I have a lot of latitude in GIMP. It's not unusual for me to rescue details in deep shadows by adding duplicate layers in screen mode (masking out all but the shadowed areas).

I finally have that process down, and it works quite well.

Since I don't shoot in JPEG at all, I cannot say if the process would be greatly compromised using in camera jpegs, but someone else may be able to share some thoughts along those lines.

Caruso

On 2009-09-30, Carusoswi wrote:

In the spirit of the OP's question, if you make no adjustments in UFRAW, is there any more latitude for adjustment in the resultant JPG file (in Gimp or other editing application) than what you might get straight from the camera?

This is not a very have-a-clear-answer topic.

I would guess that with Canon, the answer is straightforward: the RAW-converted output would be SIGNIFICANTLY better than in-camera one in ALL respects. Dynamic range, handling of clipping, handling of noise, sharpness, etc.

With cameras which use more advanced versions of the Apical Iridex hardware or firmware (starting with Sony, but Nikon is reported to be in process of catching up), the situation is not as clear. I did not see any report of RAW processor which can match Apical-style "Dynamic Range Optimizations".

So: there might be one respect (tonal mapping, sometimes called "dynamic range") in which RAW-processed-JPEG might be not as good as in-camera one...

It feels as though I have a lot of latitude in GIMP.

8-bit is good enough for "minimally postprocessed" images, since noise would provide sufficient dithering, both in highlights and in darks. However, significant noise reduction and/or substantial tonal mapping has a risk to make banding visible. Which makes GIMP not very suitable for such styles of photography. (Not so with the subjects I favor most, so I did not see that.)

Hope this helps, Ilya

Ilya -

Thanks for your earlier note -- I am quite happy to stand corrected and your post suggests a basic experiment I can easily do: compare in-camera processed and post-processed RAW images for the same scene and settings.

I'll have a limited sample to work with: my only camera delivering RAW images is a Pentax K100D, quite dated now by newer technology. On the other hand, I can compare UFRaw into GIMP and Photoshop Elements with the Pentax plug-in into PSE and (if I have a suitable intermediate format) into GIMP. At the very least I'll learn something.

If I see anything surprising or interesting I may share it and hopefully get useful feedback. Anyway, there's no substitute for knowing what one's own equipment does.

On Thu, 1 Oct 2009, Ilya Zakharevich wrote:

On 2009-09-30, Carusoswi wrote:

In the spirit of the OP's question, if you make no adjustments in UFRAW, is there any more latitude for adjustment in the resultant JPG file (in Gimp or other editing application) than what you might get straight from the camera?

This is not a very have-a-clear-answer topic.

I would guess that with Canon, the answer is straightforward: the RAW-converted output would be SIGNIFICANTLY better than in-camera one in ALL respects. Dynamic range, handling of clipping, handling of noise, sharpness, etc.

With cameras which use more advanced versions of the Apical Iridex hardware or firmware (starting with Sony, but Nikon is reported to be in process of catching up), the situation is not as clear. I did not see any report of RAW processor which can match Apical-style "Dynamic Range Optimizations".

So: there might be one respect (tonal mapping, sometimes called "dynamic range") in which RAW-processed-JPEG might be not as good as in-camera one...

I'm not sure I follow that, unless the sensor's bit-depth and that of the camera's RAW format are different.

I'm not at the stage of getting full scale from my images: still working for consistent, decent quality prints from straight-forward subjects.

- Mills

There was nothing wrong with your question: It was perfectly clear.

What was strange is that everyone on the list who replied, did not answer your question but answered the question that thought that they had read. Perhaps, they simply do not want to answer it, or the explanation missed its mark and had to be reformatted to that of the layman :) Myself included.

I had a quick look on the UFraw website. It briefly explains how it saves the file from the programme to the disc, but does not explain in which format it passes the file to Gimp. Sorry, but I don't know the answer.

You could try asking on a UFraw mailing list if there is one, or the UFraw forum:
http://sourceforge.net/projects/ufraw/forums/forum/434060

Simon.

Bryan wrote:

On 2009-09-30, Carusoswi wrote:

In the spirit of the OP's question, if you make no adjustments in UFRAW,

is

there any more latitude for adjustment in the resultant JPG file (in Gimp

or

other editing application) than what you might get straight from the

camera?

This is not a very have-a-clear-answer topic.

I would guess that with Canon, the answer is straightforward: the RAW-converted output would be SIGNIFICANTLY better than in-camera one in ALL respects. Dynamic range, handling of clipping, handling of noise, sharpness, etc.

With cameras which use more advanced versions of the Apical Iridex hardware or firmware (starting with Sony, but Nikon is reported to be in process of catching up), the situation is not as clear. I did not see any report of RAW processor which can match Apical-style "Dynamic Range Optimizations".

So: there might be one respect (tonal mapping, sometimes called "dynamic range") in which RAW-processed-JPEG might be not as good as in-camera one...

It feels as though I have a lot of latitude in GIMP.

8-bit is good enough for "minimally postprocessed" images, since noise would provide sufficient dithering, both in highlights and in darks. However, significant noise reduction and/or substantial tonal mapping has a risk to make banding visible. Which makes GIMP not very suitable for such styles of photography. (Not so with the subjects I favor most, so I did not see that.)

Hope this helps, Ilya

Ok, I want to make sure that I've asked my question clear enough before I decide that you guys have blown me away with your technological knowledge. I'm shooting in RAW and so I'm opening up a RAW file with UFRaw because without opening the file first with UFRaw, I can't get it into into Gimp for Post Processing. I hope I'm right so far. Well, after opening the RAW file in UFRaw and whether I perforn any adjustments or not in UFRaw, if I hit OK to send it to Gimp isn't it still a RAW file when it's in GIMP or has UFRaw converted it to a jpg automatically and that is why the image looks crappy in GIMP, particulairly when zoomed in on? Isn't there is only a relatively small amount of things you can do to an image in UFRaw? Which is why you'd want to get that RAW file to GIMP to be able to really do some post processing because there is only so much you can do to a jpg?

On Thu, 2009-10-01 at 07:11 -0700,

Ok, I want to make sure that I've asked my question clear enough before I decide that you guys have blown me away with your technological knowledge. I'm shooting in RAW and so I'm opening up a RAW file with UFRaw because without opening the file first with UFRaw, I can't get it into into Gimp for Post Processing. I hope I'm right so far. Well, after opening the RAW file in UFRaw and whether I perforn any adjustments or not in UFRaw, if I hit OK to send it to Gimp isn't it still a RAW file when it's in GIMP or has UFRaw converted it to a jpg automatically and that is why the image looks crappy in GIMP, particulairly when zoomed in on?

Well you would need to "Sharpen" the RAW image, JPG's are normally sharpened in camera but not so with RAW. In fact even most JPG images could benefit from some sharpening.

Cheers George

Having a look at UFRaw: In my case, it's maximum zoom is 50%. In GIMP, you may zoom 200%, 400%, 800% and, of course, such magnifications will be pixelated. Plus: File formats are, generally, defined as standardized methods of storing data/pictures on hard disk/CD/USB-Stick etc. GIMP's internal representation must follow other rules. E.g. GIMP uses an alpha channel, JPEG doesn't have an alpha channel. In future, GIMP will hold images internally by 4x32bit floating point numbers (RGBA). IMHO, none of the external picture formats uses floating point numbers. Naturally, UFRaw will send the picture to GIMP in GIMP's internal picture format.

Joerg

Am 30.09.2009 23:27, schrieb Bryan:

For starters, is there any good reason why UFRaw should be used for anything other than opening a RAW file before transporting it to Gimp?

What I really wanted to ask was, Why, when open up an image in UFRaw and it looks great, even when zoomed in as much as allowed. However, when you sent it to Gimp via UFRaw, when you zoom into even a small amount the image is very pixelated. When or how did it become downsized or some sort of processing that might cause this to happen? Is it possible to bring in a RAW file and retain it's sharpness?

Thanks for anyone's help in regards to this.

I use UFRaw all the time to convert my Fuji RAF (RAW) files to TIFF and/or JPEG. A RAW file is very much different than a TIFF or JPEG file. It is a collection of the individual light sensors in the cameras CCD or CMOS sensor.

You can read about camera sensors on wikipedia. They are way more complicated than I could explain.

RAW files generally have 10-14 bits of information from each sensor site. And there may be more than just three sensors sites that are combined into one pixel in a TIFF or JPEG image. If you look at UFRaw you will see on the very first tab a selector that has 5 different ways to interpolate the RAW file. You will also see color temperature and white balance.

If you just open your raw file with UFRaw and then send it to GIMP you will get just the conversion that UFRaw is adjusted to do at the time you send it to GIMP. When you send it to GIMP it is in GIMP's own format and isn't a JPEG or TIFF or any other kind of file until you export it.

If that looks bad to you then you can and should do some adjustments in UFRaw because that's what it's for. And you can do quite a lot in the conversion process. You won't be able to do the same kinds of adjustments in GIMP as in UFRaw. UFRaw uses all the bits from each sensor site (10-14 bits) and GIMP works (now) only in 8 bits so you lose a lot of information when the RAW image goes to GIMP. That's why you need to be careful about how you convert the RAW file.

UFRaw is based on DCRaw you should read more about DCRaw. You can answer all your questions yourself with a little reading. The use of UFRaw is somewhat complicated. It's too complicated to expect all your answers from this news group and all the information you need is on the links below.

http://ufraw.sourceforge.net/Guide.html

http://www.cybercom.net/%7Edcoffin/dcraw/

Bryan wrote:

On 2009-09-30, Carusoswi wrote:

In the spirit of the OP's question, if you make no adjustments in UFRAW,

is

there any more latitude for adjustment in the resultant JPG file (in Gimp

or

other editing application) than what you might get straight from the

camera?

This is not a very have-a-clear-answer topic.

I would guess that with Canon, the answer is straightforward: the RAW-converted output would be SIGNIFICANTLY better than in-camera one in ALL respects. Dynamic range, handling of clipping, handling of noise, sharpness, etc.

With cameras which use more advanced versions of the Apical Iridex hardware or firmware (starting with Sony, but Nikon is reported to be in process of catching up), the situation is not as clear. I did not see any report of RAW processor which can match Apical-style "Dynamic Range Optimizations".

So: there might be one respect (tonal mapping, sometimes called "dynamic range") in which RAW-processed-JPEG might be not as good as in-camera one...

It feels as though I have a lot of latitude in GIMP.

8-bit is good enough for "minimally postprocessed" images, since noise would provide sufficient dithering, both in highlights and in darks. However, significant noise reduction and/or substantial tonal mapping has a risk to make banding visible. Which makes GIMP not very suitable for such styles of photography. (Not so with the subjects I favor most, so I did not see that.)

Hope this helps, Ilya

Ok, I want to make sure that I've asked my question clear enough before I decide that you guys have blown me away with your technological knowledge. I'm shooting in RAW and so I'm opening up a RAW file with UFRaw because without opening the file first with UFRaw, I can't get it into into Gimp for Post Processing. I hope I'm right so far. Well, after opening the RAW file in UFRaw and whether I perforn any adjustments or not in UFRaw, if I hit OK to send it to Gimp isn't it still a RAW file when it's in GIMP or has UFRaw converted it to a jpg automatically and that is why the image looks crappy in GIMP, particulairly when zoomed in on? Isn't there is only a relatively small amount of things you can do to an image in UFRaw? Which is why you'd want to get that RAW file to GIMP to be able to really do some post processing because there is only so much you can do to a jpg?

On 2009-10-01, John Mills wrote:

With cameras which use more advanced versions of the Apical Iridex hardware or firmware (starting with Sony, but Nikon is reported to be in process of catching up), the situation is not as clear. I did not see any report of RAW processor which can match Apical-style "Dynamic Range Optimizations".

So: there might be one respect (tonal mapping, sometimes called "dynamic range") in which RAW-processed-JPEG might be not as good as in-camera one...

I'm not sure I follow that, unless the sensor's bit-depth and that of the camera's RAW format are different.

Sensor bit-depth is an absolutely bogus metric (unless one uses it as "an indicator of amount of R&D", which may correlate with other, important issues; such as read noise and correlation of noise of nearby pixels).

If RAW files were compressed to 8-bit gamma=2, they won't loose "practically any" information; 9-bit would be a significant overkill (assuming full-well about 70K electrons, as typical large-sensor dSLRs have). gamma=2.2 is very similar. (The special significance of quantization after gamma=2 is that Poisson noise becomes "constant-width", thus dithers in dark parts as well as in highlights.)

If you do not know what DRO is, look on dpreview, and/or look for examples on Apical site.

Ilya

On 2009-10-01, Bryan wrote:

Well, after opening the RAW file in UFRaw and whether I perforn any adjustments or not in UFRaw, if I hit OK to send it to Gimp isn't it still a RAW file when it's in GIMP or has UFRaw converted it to a jpg automatically and that is why the image looks crappy in GIMP, particulairly when zoomed in on?

I have no idea how your communication channel is configured. I would use sRGB 8-bit TIFF (do not know whether deflation makes sense [called ZIP in GIMP]) with no downscale, or at most to-75% downscale (to-65% should be OK with most pocket cameras).

Isn't there is only a relatively small amount of things you can do to an image in UFRaw?

Probably true (never used UFRaw ;-). On the other hand, AFAIK, GIMP has practically no tools to do photo-related work either (one can't even apply a curve to L channel without going through a hundred hoops).

Hope this helps,
Ilya

There was nothing wrong with your question: It was perfectly clear.

What was strange is that everyone on the list who replied, did not answer your question but answered the question that thought that they had read. Perhaps, they simply do not want to answer it, or the explanation missed its mark and had to be reformatted to that of the layman :) Myself included.

Actually, we're all making what I think are sincere efforts to contribute comments that might help answer the OP's question. I never gave any thought to what format UFRaw sends to Gimp, but I bet it's not some Gimp native format, as I doubt Gimp has one. Gimp saves xcf files which, if my understanding is correct, are equivalent to edit decision files in video applications, or .indd files in Adobe's InDesign page layout application. These are, for the most part, reference files that keep track of the changes you are trying to make to an image (in the case of photo editing apps). There is no image format until you save the file in Gimp. That's why, if you try to print a non-flattened image, Gimp complains and then offers to export it for you.

I also erred in my previous post about UFRaw sending jpegs to Gimp. Actually, I'm not certain what sort of file it sends to Gimp, but, I'm guessing it's starts out as (or with components capable of being saved as) an 8-bit Tiff.

I believe that, if you use the stand alone version of UFRaw, it will give you a choice of file types when you 'save as.'

. . . and that brings us back to what I read as the OP's central question. In terms of image quality, are we better off making most of our adjustments in Gimp after having used UFRaw to convert the raw image, or would we be wiser to make whatever adjustment available to us in UFRaw before sending the image to Gimp.

I would bet that, in my case, it's a moot point, as I will get acceptable results either way that are close enough, given the uses to which I will put the images, that my viewers would not discern between either approach. Theoretically, however, I'm guessing that there may be a difference, and I'd be curious to hear from someone with the technical depth to offer some enlightenment.

This has turned out to be quite an interesting thread. I'm glad I took the time to read and respond. Thanks, OP.

Caruso

Ok, I want to make sure that I've asked my question clear enough before I decide that you guys have blown me away with your technological knowledge. I'm shooting in RAW and so I'm opening up a RAW file with UFRaw because without opening the file first with UFRaw, I can't get it into into Gimp for Post Processing. I hope I'm right so far. Well, after opening the RAW file in UFRaw and whether I perforn any adjustments or not in UFRaw, if I hit OK to send it to Gimp isn't it still a RAW file when it's in GIMP or has UFRaw converted it to a jpg automatically and that is why the image looks crappy in GIMP, particulairly when zoomed in on?

I have been following this thread with interest so I decided to do some tests. The results may be relevant or not as the case may be but I think that they are interesting. My little camera gives a RAW image = 8.6 MB and a jpeg image = 2.6 MB. The developed RAW image from UFRaw saved at 98% jpeg = 17.2 MB and then that image loaded into GIMP and saved as xcf = 17.0 MB.

Norman

On 2009-10-02, Norman Silverstone wrote:

I have been following this thread with interest so I decided to do some tests. The results may be relevant or not as the case may be but I think that they are interesting. My little camera gives a RAW image = 8.6 MB and a jpeg image = 2.6 MB. The developed RAW image from UFRaw saved at 98% jpeg = 17.2 MB

What do you think are benefits of using jpegs with quality above 95%?

Better use compressed 8-bit sRGB TIFF instead (all minilabs I know would reject TIFF with *any* compression, though...).

Ilya

I have been following this thread with interest so I decided to do some tests. The results may be relevant or not as the case may be but I think that they are interesting. My little camera gives a RAW image = 8.6 MB and a jpeg image = 2.6 MB. The developed RAW image from UFRaw saved at 98% jpeg = 17.2 MB

What do you think are benefits of using jpegs with quality above 95%?

I have absolutely no idea, it is just that I came across a reference somewhere which said "I use 98% jpeg compression when archiving images.

Better use compressed 8-bit sRGB TIFF instead (all minilabs I know would reject TIFF with *any* compression, though...).

Why is that?

Norman

On Sat, 03 Oct 2009 08:59:11 +0100, Norman Silverstone wrote:

I have absolutely no idea, it is just that I came across a reference somewhere which said "I use 98% jpeg compression when archiving images.

If you're archiving images that aren't compressed with lossy compression, use a format that stores them losslessly. If the images are already compressed with lossy compression (=JPEG), don't recompress them.

On 2009-10-03, Norman Silverstone wrote:

What do you think are benefits of using jpegs with quality above 95%?

I have absolutely no idea, it is just that I came across a reference somewhere which said "I use 98% jpeg compression when archiving images.

Better use compressed 8-bit sRGB TIFF instead (all minilabs I know would reject TIFF with *any* compression, though...).

Why is that?

Why is WHAT??? You need to Trim the posting more precisely...

If you mean "why TIFF?": Since TIFF provides no-artefacts storage at the same size as what you got from your JPEG (which WILL create artefacts).

Why 8-bit? I already discussed that; and since your input is 8-bit, most probably you won't get any advantage from 16-bit. (Well, currently GIMP won't even produce 16-bit...)

Why compressed? To save size...

Puzzled, Ilya

Better use compressed 8-bit sRGB TIFF instead (all minilabs I know would reject TIFF with *any* compression, though...).

Why is that?

Why is WHAT??? You need to Trim the posting more precisely...

I wonder if the questioner meant why do minilabs reject TIFF with *any* compression?

Norman