Raw Image Capabilities of CS2

wrote:

I currently have Photoshop 7 and am considering upgrading to CS2. Do all CS2 functions (filters, masks, color correstions etc.) work with raw images, or just a subset?

That’s sort of an invalid question. :o)

You need to import your RAW file into Photoshop before you can do anything with it. Once you have it in there you can do what you like with it. As long as your camera is supported by the RAW importer utility you can consider that all features work.

—
<a href="http://www.derekfountain.org/">Derek Fountain</a> on the web at http://www.derekfountain.org/

Sorry for not being precise. I assume that when the raw image is imported into CS2, it is 16-bit. My question is, therefore, whether all CS2 functions work with 16-bit images, or just a subset. Thanks.

writes …

I currently have Photoshop 7 and am considering upgrading to CS2. Do all CS2 functions (filters, masks, color correstions etc.) work with raw images, or just a subset?

If you mean digital-camera RAW files, you have to convert them from RAW to tiff or jpeg first.

Sorry for not being precise. I assume that when the raw image is imported into CS2, it is 16-bit. My question is, therefore, whether all CS2 functions work with 16-bit images, or just a subset. Thanks.

Some of the filters are still 8 bit only … this is likely because Adobe bought filter sets from 3rd party users long ago and hasn’t bothered to re-write them for 16 bit. For sure pretty much all the things you’d want to do with 16 bit files are supported but not *all* the filters. There is extended 16 bit support compared to version 7, for example.

Bill

wrote:

Sorry for not being precise. I assume that when the raw image is imported into CS2, it is 16-bit.

Not necessarily. You decide that and you can also choose 8 bits. If you do all the color correction in the RAW conversion, there is no compelling reason to keep working in 16 bits.

My question is, therefore, whether all
CS2 functions work with 16-bit images, or just a subset.

All major functions work in 16 bit, but most filters are still 8 bits only. These are the ‘creative’ filters mainly, filters like Unsharp Mask (and the new Smart Sharpen) do work in 16 bits.

—
Johan W. Elzenga johan<<at>>johanfoto.nl Editor / Photographer http://www.johanfoto.nl/

On 2 Jan 2006 08:39:51 -0800, wrote:

Sorry for not being precise. I assume that when the raw image is imported into CS2, it is 16-bit. My question is, therefore, whether all CS2 functions work with 16-bit images, or just a subset. Thanks.

Once you have adjusted your exposure and color and are ready to save the file as a non-RAW image you have the choice of saving it in 8 or 16 bit format. As you well know, Some filters and functions do not work on 16 bit files. Also, I do not think you can save a 16 bit as anything other than a .PSD file.

I could be wrong here as I am still pretty new to CS. Want to upgrade to CS2 real soon!

Once you have adjusted your exposure and color and are ready to save the file as a non-RAW image you have the choice of saving it in 8 or 16 bit format. As you well know, Some filters and functions do not work on 16 bit files. Also, I do not think you can save a 16 bit as anything other than a .PSD file.

TIFF and PNG are still possible when using 16 bits per channel.

Waldo

Waldo wrote:

Once you have adjusted your exposure and color and are ready to save the file as a non-RAW image you have the choice of saving it in 8 or 16 bit format. As you well know, Some filters and functions do not work on 16 bit files. Also, I do not think you can save a 16 bit as anything other than a .PSD file.

TIFF and PNG are still possible when using 16 bits per channel.
Waldo

So is JPEG2000.

Clyde

"Johan W. Elzenga" wrote in message

wrote:

Sorry for not being precise. I assume that when the raw
image is imported into CS2, it is 16-bit.

Not necessarily. You decide that and you can also choose 8 bits.

Agreed.

If you do all the color correction in the RAW conversion, there is no compelling reason to keep working in 16 bits.

Here it becomes a matter of personal choice. The question is what comprises "all the color correction". Some (myself included) like to avoid potential (bad) surprises with regards to posterization, which would mean a 16-b/ch workflow right up to final profile conversion and sharpening. Others change mode to 8-b/ch earlier in the process, to each their own.

My question is, therefore, whether all
CS2 functions work with 16-bit images, or just a subset.

All major functions work in 16 bit, but most filters are still 8 bits only. These are the ‘creative’ filters mainly,

Indeed, because their purpose is less restricted to accurate color.

filters like Unsharp Mask (and the new Smart Sharpen) do work in 16 bits.

Sharpening, which (to me anyway) is the last step in processing, signals the need to postpone fall-back to 8-b/ch mode (or re-initiate 16-b/ch mode) to avoid the propagation of round-off errors with each processing step. After all, who wants to sharpen posterization in otherwise smooth gradients?

Bart

"Bart van der Wolf" wrote in message
[re correcting in 16 bits]

Here it becomes a matter of personal choice. The question is what comprises "all the color correction". Some (myself included) like to avoid potential (bad) surprises with regards to posterization, which would mean a 16-b/ch workflow right up to final profile conversion and sharpening. Others change mode to 8-b/ch earlier in the process, to each their own.

….

Sharpening, which (to me anyway) is the last step in processing, signals the need to postpone fall-back to 8-b/ch mode (or re-initiate 16-b/ch mode) to avoid the propagation of round-off errors with each processing step. After all, who wants to sharpen posterization in otherwise smooth gradients?

Bart, is this in linear gamma space? It sounds as if you speak from experience, and have an example of an image that shows posterization, or sharpening artifacts, that are present after an 8 bit correction, but absent in a similar 16 bit correction. If this is the case (and you are not working in linear gamma space), I am still questing for such an example image.

—

Mike Russell
www.curvemeister.com

Mike Russell wrote:
[snip

Bart, is this in linear gamma space? It sounds as if you speak from experience, and have an example of an image that shows posterization, or sharpening artifacts, that are present after an 8 bit correction, but absent in a similar 16 bit correction. If this is the case (and you are not working in linear gamma space), I am still questing for such an example image.

I have recently been doing some testing to see how safely I can work in 8-bit rather than 16-bit after a raw conversion. In summary, if I am careful with the settings, (in ACR 3.x), I can normally JUST get away with 8-bit. But only just.

Normally, I can get a good histogram and no posterisation in Photoshop. I rarely have to do significant global corrections in Photoshop if I take enough care in ACR. (If I did, I might query my ACR settings, and try again). If I do LOCAL changes, however, they are likely to be more dramatic than would a global one, and looking at the local histogram where I’ve done such edits I sometimes see the histogram breaking up into spikes. On recent pictures, I haven’t actually seen bad results from this when printed at A3, but if I did much more with those local areas in Photoshop, (or perhaps if they were larger areas), I think I would. (I really need 9-bits for comfort!)

Although I have previously been working at 16-bits, I am trying to work at 8-bits instead. (For a few reasons, such as size, speed, filters, etc). I think if I got posterisation in local areas, as a result of local edits, I might go back to ACR and do another raw conversion suitable for that local area so that I could get a good 8-bit result in Photoshop, then combine the two conversions.

This is still a bit tentative. I intend to run at 8-bit for a while, and keep monitoring things closely just in case.

—
Barry Pearson
http://www.barry.pearson.name/photography/
http://www.birdsandanimals.info/

Barry Pearson wrote:

Mike Russell wrote:
[snip

Bart, is this in linear gamma space? It sounds as if you speak from experience, and have an example of an image that shows posterization, or sharpening artifacts, that are present after an 8 bit correction, but absent in a similar 16 bit correction. If this is the case (and you are not working in linear gamma space), I am still questing for such an example image.

I have recently been doing some testing to see how safely I can work in 8-bit rather than 16-bit after a raw conversion. In summary, if I am careful with the settings, (in ACR 3.x), I can normally JUST get away with 8-bit. But only just.

Normally, I can get a good histogram and no posterisation in Photoshop. I rarely have to do significant global corrections in Photoshop if I take enough care in ACR. (If I did, I might query my ACR settings, and try again). If I do LOCAL changes, however, they are likely to be more dramatic than would a global one, and looking at the local histogram where I’ve done such edits I sometimes see the histogram breaking up into spikes. On recent pictures, I haven’t actually seen bad results from this when printed at A3, but if I did much more with those local areas in Photoshop, (or perhaps if they were larger areas), I think I would. (I really need 9-bits for comfort!)

Although I have previously been working at 16-bits, I am trying to work at 8-bits instead. (For a few reasons, such as size, speed, filters, etc). I think if I got posterisation in local areas, as a result of local edits, I might go back to ACR and do another raw conversion suitable for that local area so that I could get a good 8-bit result in Photoshop, then combine the two conversions.

This is still a bit tentative. I intend to run at 8-bit for a while, and keep monitoring things closely just in case.

—
Barry Pearson
http://www.barry.pearson.name/photography/
http://www.birdsandanimals.info/

I have never seen an actual photograph that was harmed by doing anything in 8 bit. I have seen plenty of people with numbers to show that 8 bit is bad. I have personally seen a few extreme gradient exercises where 8 bit will show a difference. I still haven’t seen any real-world photographs where 8 bit makes a bit of difference. Particularly on the printed result – which is all that really matters.

Clyde

"Clyde" wrote in message
SNIP

I have never seen an actual photograph that was harmed by doing anything in 8 bit.

"Doing anything" makes it easy to present a sample crop: <http://www.xs4all.nl/~bvdwolf/main/downloads/SatDish.jpg> It is just the result of a gamma adjustment from linear to average PC monitor gamma in 8-bit/channel versus 16-b/ch mode with Photoshop. Look at the left side of the satellite dish, and to the shadows. The original linear gamma file was from a 12-b/ch P&S digicam.

Things will be harder to pin-point in image areas with lots of detail or noise, but artifacts will be there.

Bart

"Mike Russell" wrote in message

"Bart van der Wolf" wrote in message
[re correcting in 16 bits]

SNIP

Sharpening, which (to me anyway) is the last step in processing, signals the need to postpone fall-back to 8-b/ch mode (or re- initiate 16-b/ch mode) to avoid the propagation of round-off errors with each processing step. After all, who wants to sharpen posterization in otherwise smooth gradients?

Bart, is this in linear gamma space?

No, not necessarily.

It sounds as if you speak from experience, and have an example of an image that shows posterization, or sharpening artifacts, that are present after an 8 bit correction, but absent in a similar 16 bit correction. If this is the case (and you are not working in linear gamma space), I am still questing for such an example image.

Sure, and it’s not even any particular or underexposed example: < http://www.xs4all.nl/~bvdwolf/main/downloads/8vs16-bpch%20pr ocessing.png> It’s just from a Pixmantec RSP raw conversion, a Levels color balancing adjustment (sky was a bit too pink for my taste), some profile conversions (e.g. from Wide to final sRGB for Web publishing), and RGB to Lab (for potential(!) Luminosity adjustment) and back, with CS2 small radius Smart Sharpening at output size, shown at 200% zoom.

Personally, my workflow would attempt to avoid ANY 8-b/ch operation, until the final output image stage.

Bart

"Bart van der Wolf" wrote in message

[re request for an example image]

Sure, and it’s not even any particular or underexposed example: < http://www.xs4all.nl/~bvdwolf/main/downloads/8vs16-bpch%20pr ocessing.png> It’s just from a Pixmantec RSP raw conversion, a Levels color balancing adjustment (sky was a bit too pink for my taste), some profile conversions (e.g. from Wide to final sRGB for Web publishing), and RGB to Lab (for potential(!) Luminosity adjustment) and back, with CS2 small radius Smart Sharpening at output size, shown at 200% zoom.

Thanks for the example. I can see the artifacting in the sky. Am I correct that this is a photograph of a painting?

Personally, my workflow would attempt to avoid ANY 8-b/ch operation, until the final output image stage.

Yes, I’m certainly aware of that, and I thank you for your past examples as well.
—
Mike Russell
www.curvemeister.com

"Mike Russell" wrote in message
SNIP

Thanks for the example. I can see the artifacting in the sky. Am I correct that this is a photograph of a painting?

No it’s from part of a panoramic stitching job:
Here’s the relevant full frame, and the crop was around the block at the top:
<http://www.xs4all.nl/~bvdwolf/main/downloads/Batavia.jpg>

The sample crops were treated exactly the same, only difference was a mode change to 8-b/ch after Raw conversion for one of them.

Bart

Bart van der Wolf wrote:

"Clyde" wrote in message
SNIP

I have never seen an actual photograph that was harmed by doing anything in 8 bit.

"Doing anything" makes it easy to present a sample crop: <http://www.xs4all.nl/~bvdwolf/main/downloads/SatDish.jpg> It is just the result of a gamma adjustment from linear to average PC monitor gamma in 8-bit/channel versus 16-b/ch mode with Photoshop. Look at the left side of the satellite dish, and to the shadows. The original linear gamma file was from a 12-b/ch P&S digicam.

Things will be harder to pin-point in image areas with lots of detail or noise, but artifacts will be there.

Bart

Actually you illustrate my point very well. There are and always will be artifacts and imperfections at 200% on the screen. I don’t create pictures to be viewed at 200% on the screen.

I do see a difference between the two pictures in your example. I don’t see a difference that is relevant photographically. I can’t tell which one is "better". From my point of view, it doesn’t matter. They both look good enough to tell the story of the picture.

I bet if you print this picture off at any size where it looks like a photograph, you will not be able to tell the difference without magnification. Even if you think you can tell the difference, it doesn’t matter to story of the photograph. I bet that anyone who has not seen your "flaws" or told about them will be able to notice any difference at all.

My point is that Pixel-Pushers and Measurebators can find difference between 16 bit processed images and 8 bit processed one. In the final print, I have never seen any difference, let alone a significant one.

Clyde

Bart van der Wolf wrote:

"Johan W. Elzenga" wrote in message

[snip]

If you do all the color correction in the RAW conversion, there is no compelling reason to keep working in 16 bits.

Here it becomes a matter of personal choice. The question is what comprises "all the color correction". Some (myself included) like to avoid potential (bad) surprises with regards to posterization, which would mean a 16-b/ch workflow right up to final profile conversion and sharpening. Others change mode to 8-b/ch earlier in the process, to each their own.

[snip]

I offer an alternative workflow. Work in 8-bit where it makes sense. "Convert" to 16-bit where necessary. Yes, I know that you can’t just do this! Once you have lost the data with 8-bit, you can’t retrieve it with 16-bit. But …

I now, at least until I prove to myself that it is the wrong method, raw-convert to 8-bit. I take care to label my layers with some details about the raw conversion settings I used as input to those layers. If it became very important to have a 16-bit version, I probably have enough information to repeat the raw conversions, then build up the Photoshop image in 16-bit form.

Here is a thought: suppose I converted a Photoshop file, with layers intact, from 8-bit to 16-bit. Any IMAGE layers won’t convert properly – they need new raw conversions into 16-bit instead of into 8-bit.

However, it seems to me that ALL the layer masks and adjustment layers are still valid. Given that a lot of the effort for a photograph may go into these, it seems to me that I can convert from 8-bit to 16-bit for a fraction of the original effort, mainly by replacing 2 or 3 image layers, not all the layer masks and adjustment layers.

So I feel that I can work in 8-bit mode most of the time, and if the need arises convert to 16-bit mode for a relatively small extra overhead.

—
Barry Pearson
http://www.barry.pearson.name/photography/
http://www.birdsandanimals.info/

"Bart van der Wolf" wrote in message

<http://www.xs4all.nl/~bvdwolf/main/downloads/Batavia.jpg>

It’s the Golden Hind, isn’t it? I should have recognized it right off. 🙂

Thanks very much for the example. I’m familiar with 8 vs. 16 bit problems when using ProPhoto or Wide Gamut RGB. Those color spaces essentially require 16 bit processing, since so much of their gamut is devoted to colors that cannot be viewed or printed, and it is easy to create artifacts. —
Mike Russell
www.curvemeister.com

"Mike Russell" wrote in message

"Bart van der Wolf" wrote in message

<http://www.xs4all.nl/~bvdwolf/main/downloads/Batavia.jpg>

It’s the Golden Hind, isn’t it? I should have recognized it right off. 🙂

No, the reason you didn’t recognize it is because it’s a different ship 😉
<http://www.bataviawerf.nl/en/batavia.html>. It’s the same ship I used for a sharpening example (for someone questioning Bayer CFA resolution) at:
< http://www.xs4all.nl/~bvdwolf/main/downloads/Batavia_Crop.jp g>

Thanks very much for the example. I’m familiar with 8 vs. 16 bit problems when using ProPhoto or Wide Gamut RGB. Those color spaces essentially require 16 bit processing, since so much of their gamut is devoted to colors that cannot be viewed or printed, and it is easy to create artifacts.

Indeed, however there are other colorspaces (e.g. "Best RGB" or "Beta RGB") that can efficiently encode the image’s entire gamut, should Adobe RGB be too limiting, while running less of a risk compared to ProPhoto RGB.
< http://www.brucelindbloom.com/index.html?WorkingSpaceInfo.ht ml>

Bart

"Barry Pearson" wrote in message
SNIP

I offer an alternative workflow. Work in 8-bit where it makes sense. "Convert" to 16-bit where necessary. Yes, I know that you can’t just do this! Once you have lost the data with 8-bit, you can’t retrieve it with 16-bit. But …

The trick would be in avoiding cumulative rounding errors from propagating with multiple adjustments. Obviously, accuracy would get lost once reduced to 8-b/ch, but depending on the use it may be tolerable. The use of dithering in gamut conversions may also help to hide some artifacts.

Adjustment layers are very useful because the final calculation is postponed for the final flattening of the layer composite, flattening is therefore best done in 16-b/ch. Masks may be a little less accurate, but if they’re part of a final flattening then things may be tolerable.

Bart