Curiosity only: If we could print a RAW file, would it look better then a printed JPEG file?

[HEA-45]

I am curious about RAW and JPEG file formats. We all know that photographing and post-processing RAW files are considered the best technique. But, in order to obtain a print by an outside source, the RAW file must be converted to a JPEG format. Does this conversion degrade the quality of the electronic image and, consequently, the quality of the printed image?

Note: I love to print my photographs, and I use an outside source. They only accept JPEG format. I have no complaints about the quality of the printed image. I pose the question to satisfy my curiosity...not to solve any specific problem.

 

[Mark S Abeln]

HEA-45 wrote:

Does this conversion degrade the quality of the electronic image and, consequently, the quality of the printed image?

Yes. JPEG is a lossy file format, and typically has a much smaller bit depth than RAW images. We should expect lower final image quality.

However, RAW images are not directly printable for these reasons:

• There are many RAW formats, and each manufacturer has its own, which sometimes changes between camera models. Is it reasonable to ask each printer maker to support the arbitrary file formats of all the manufacturers, or is it more reasonable to have a standard file format used to print images?
• The Bayer array found in most digital cameras means that each pixel only has one color associated with it: red, green, or blue. Interpolation must be used to estimate full color for each pixel location. There are various interpolation algorithms in use, with differing trade-offs.
• The color space in RAW files do not correspond to any standard color space such as sRGB. RAW converters will convert the color space used by the camera to a standard which can be interpreted accurately by the printer.
• The flat tonal curve of RAW images often looks poor. Adding contrast to images usually makes them look better.
• The eye does not respond linearly to brightness, in that something twice as bright only looks somewhat brighter. For this reason, we use gamma correction, reordering the brightness numbers so that more bits of data are assigned to the darkest tones. To get good shadow detail, a non-gamma corrected image would have to have a much greater bit depth. JPEGs can get away with only using 8 bits of data (making files smaller) because of gamma correction, while non-gamma-corrected RAW images have a number of more bits per pixel, and so is somewhat larger in size.
• RAW images are rather soft. JPEG conversions will add sharpness to the image.
• There are other considerations also.

RAW images must be converted to a format suitable for printing and also converted so that the image will look good. The main question then becomes where should the RAW conversion take place? In the camera or on the computer? I suppose it could be done in the printer, but this adds considerable complexity to the printer software: also, we would not be able to preview the image rendering until after the image is printed, which seems rather late in the process.

 

[apaflo]

HEA-45 wrote:

I am curious about RAW and JPEG file formats. We all know that photographing and post-processing RAW files are considered the best technique. But, in order to obtain a print by an outside source, the RAW file must be converted to a JPEG format. Does this conversion degrade the quality of the electronic image and, consequently, the quality of the printed image?

Actually there is no need for a JPEG at all.

Demosaicing a Bayer Encoded raw data set converts the sensor data to a 16 bit RGB encoded image (and it is some manipulation of that which can be viewed on screen). The raw sensor data contains information for essentially an infinite number of slightly different images and does not define exactly one image. The process of converting to a 16 bit RGB image defines exactly one specific image. To save that image to disk it is converted to one of several possible image file formats. PNG, JPEG, PPM, TIFF and probably a few others are all possible.

Note that TIFF and PPM are examples of "lossless" formats that will retain all of the information in the 16 bit RGB encoded image, while others such as JPEG are lossy and will instead encode only information that is essential for viewing.

The amount of information encoded into a JPEG format can be varied by setting the "quality", and if set high enough there will be no visible degradation from what would be seen with any other format. (However, since not all information was saved, any editing will necessarily not be as good as what can be accomplished if a lossless format had been used.)

Note: I love to print my photographs, and I use an outside source. They only accept JPEG format. I have no complaints about the quality of the printed image. I pose the question to satisfy my curiosity...not to solve any specific problem.

JPEG is just fine for printing. It has the advantage of accomplishing what is needed with just about the smallest file size possible. That makes it great for email or other electronic transfers.

 

[John Deerfield]

To put it simply, you can't print a Raw file because the Raw file is NOT an image. It's a set of instructions on how to make an image if you will. So now the question becomes whether it is better to print a 16-bit file (TIFF, PSD) or an 8-bit JPEG. In terms of finding a lab that will actually print from a 16-bit TIF, good luck! I know I have seen them but they are a very small niche. Epson makes 16-bit print drivers for Macs if you want to do your own printing. And then either way it really boils down to whether you can see any visible difference. I am sure in the fine art community there may be those that can discern a difference on the right image, but 99.99% of us will never tell a difference between an 8-bit print vs 16-bit print.

http://www.earthboundlight.com/phototips/16-bit-printing.html

http://www.olegnovikov.com/technical/epson4880/epson4880_8vs16bit.shtml

 

[Sandy]

Printing from RAW without using the JPEG/TIF should be better - though whether you would see the difference is another matter. You could try Qimage Ultimate free trial


http://www.ddisoftware.com/qimage-u/downloads.htm

which allows printing from the RAW file without going to JPEG, TIF, etc. nb the RAW file will need sharpening and possibly other adjustments.

If I want a top quality print I start from the RAW file, make appropriate adjustments and then print direct from the adjusted file. I suppose one way to compare is to start with the RAW file, make the necessary adjustments, print. Also convert that adjusted RAW file to JPEG and print the JPEG. Using Qimage the original RAW file, adjusted RAW file and JPEG will all be available.

If you have set the camera to take RAW+JPEG there is a possible comparison with the camera's JPEG but that is sure to have been processed rather differently from the way you do the RAW file so you wouldn't really be able to tell how much was processing and how much was RAW/JPEG.

Try it and let us know.

 

[hedwards]

apaflo wrote:

HEA-45 wrote:

Note: I love to print my photographs, and I use an outside source. They only accept JPEG format. I have no complaints about the quality of the printed image. I pose the question to satisfy my curiosity...not to solve any specific problem.

JPEG is just fine for printing. It has the advantage of accomplishing what is needed with just about the smallest file size possible. That makes it great for email or other electronic transfers.

Right, as long as one isn't doing extreme crops or blow ups there should be little or no difference between TIFFs, PPMs and JPGs other than filesize.

However, if you make use of continuous tones in an image and you insist on leaving individual pixels visible without resamping then you may occasionally have the compression algorithm leave you with a series of dots all over the place.

In the real world though, you're unlikely to notice that these days as the images are large enough that the samples are small enough not to be noticed on normal sized prints.

On a side note, JPGs are really, really old at this point, I wish that JPG2000 or one of the newer formats would catch on as you get better quality with the same filesize.

 

[apaflo]

hedwards wrote:

apaflo wrote:

HEA-45 wrote:

Note: I love to print my photographs, and I use an outside source. They only accept JPEG format. I have no complaints about the quality of the printed image. I pose the question to satisfy my curiosity...not to solve any specific problem.

JPEG is just fine for printing. It has the advantage of accomplishing what is needed with just about the smallest file size possible. That makes it great for email or other electronic transfers.

Right, as long as one isn't doing extreme crops or blow ups there should be little or no difference between TIFFs, PPMs and JPGs other than filesize.

Lets state that more accurately. As long as the JPEG file is not edited there won't be any differences.

Extreme crops or "blow ups", when done prior to saving the image in JPEG format are saved just as accurately as the image at original size. But that is also true of other edits. The point is that a JPEG file should be a final product, not an intermediate file that will be edited further.

 

[hedwards]

Not true. I've seen it myself in the past where you're displaying a size where the image permits you to see the sample plots being used. I'm not sure if I've still got those photos around, but it's definitely something that happens from time to time if you've got smooth tones and are needing to stretch the data a bit.


As far as I know, the only way to avoid that is to use a format that doesn't average adjacent values and block sort them. It's completely inevitable if the block size is large enough to be seen and JPG uses blocks that are large enough to be seen.


It's a moot point as when one is trying to push the limits like that, one should have the technical knowledge to know that's going to happen.

As far as I know, this doesn't apply to prints which are typically of a much higher resolution and subject to interpolation by the printer, but at a typical DPI of 72, your eyes don't have to be that good to see the sampling going on.

 

[GreenMountainGirl]

Right, as long as one isn't doing extreme crops or blow ups there should be little or no difference between TIFFs, PPMs and JPGs other than filesize.

Eventually I would like to purchase a printer that will use larger paper. Any words of advice on how to handle the extra enlarging that will occur?

However, if you make use of continuous tones in an image and you insist on leaving individual pixels visible without resamping then you may occasionally have the compression algorithm leave you with a series of dots all over the place.

What is resampling? How do you know you are leaving individual pixels visible? (Maybe this is a Photoshop CS process? I use LR4 and PSE 11.)

I tried doing an enlargement of just one portion of an image the other day, the picture appeared to have a pebble-type surface! Must have stretched beyond normal limits...

 

[apaflo]

hedwards wrote:

Not true. I've seen it myself in the past where you're displaying a size where the image permits you to see the sample plots being used. I'm not sure if I've still got those photos around, but it's definitely something that happens from time to time if you've got smooth tones and are needing to stretch the data a bit.

As far as I know, the only way to avoid that is to use a format that doesn't average adjacent values and block sort them. It's completely inevitable if the block size is large enough to be seen and JPG uses blocks that are large enough to be seen.

It's a moot point as when one is trying to push the limits like that, one should have the technical knowledge to know that's going to happen.

As far as I know, this doesn't apply to prints which are typically of a much higher resolution and subject to interpolation by the printer, but at a typical DPI of 72, your eyes don't have to be that good to see the sampling going on.

What you are describing is the result of re-sampling the JPEG for display, which is an "edit", and as stated should be avoided.

For example, if you have and image that is 341x228 pixels, but display it with software that uses a 1024x684 window on you monitor screen. If it is not resampled before display the artifacts you mention will not be visible. Likewise if it were originally saved as a 1024x684 JPEG instead of the smaller one, then it would not be resampled before display and also would not have the visible artifacts.

That was exactly the point: do not edit the JPEG image!

 

[AnandaSim]

HEA-45 wrote:

I am curious about RAW and JPEG file formats. We all know that photographing and post-processing RAW files are considered the best technique.

It isn't if the operator is not skilled or the workflow is not colour calibrated or the screen is not the best.

But, in order to obtain a print by an outside source, the RAW file must be converted to a JPEG format.

No, not really. I can pretty much guess high quality labs will do 16 bit TIFF. At a store near me, they have also two classes of printing - normal use, lower priced photos vs higher prices, colour calibrated photos - if you give them a high quality photo file and choose low priced photos, it's NOT "the best"

Does this conversion degrade the quality of the electronic image and, consequently, the quality of the printed image?

Please do not think of RAW as a displayable, printable format. RAW is not useable until you open the file in a photo program and the photo program applies the rendering recipe - gamma contrast curve, white balance, noise reduction (or not) etc... What you see on the screen when you open the RAW file HAS ALREADY been processed this way by a recipe either stored in the photo program or read from suggestions stored in the raw file.

Once you have the photo on the screen, you can then save the file in TIFF, PSD, JPEG, PND, BMP etc.... If you save the photo as 16 bit TIFF, it is pretty much the photo you see on the screen. If you save it as 8 bit JPEG with maximum quality - the file is much smaller but still not bad.

Note: I love to print my photographs, and I use an outside source. They only accept JPEG format. I have no complaints about the quality of the printed image. I pose the question to satisfy my curiosity...not to solve any specific problem.

Find a lab that can do 16 bit tiff, print a high priced print for one print and see for yourself.

 

[apaflo]

GreenMountainGirl wrote:

Right, as long as one isn't doing extreme crops or blow ups there should be little or no difference between TIFFs, PPMs and JPGs other than filesize.

Eventually I would like to purchase a printer that will use larger paper. Any words of advice on how to handle the extra enlarging that will occur?

It won't be a problem if you adjust the pixel dimensions of the image correctly before it is saved to a print file.

And example would be if you get really excited and end up with an Epson printer that can produce 17 inch wide prints! It prints at 360 pixels per inch. So you shoot a great picture of your child, and want to send Grandma a 16x20, plus you want an 8x10 for your desk at work. Lets assume the camera is a Nikon D5100 that produces 4928x3264 images. (You can adjust the various numbers to whatever it might actually be, and then redo the same arithmetic as below to get a "real" example.)

First, the image does not have the same aspect ratio as a 16x20 print, so it needs to be cropped. The largest it can end up is 4080x3264. But that pixel dimension would, with a 360 pixel per inch printer, produce an image that is 11.33" x 9.07". That is smaller than the 16x20 you want and larger than the 8x10. So at that point two different "print files" should be produced. One that is (360*20 x 360*16) and one that is (360*10 x 360*8). The first is 7200x5760 pixels and the second is 3600x2880 pixels. Each image, at that size, should be sharpened to taste and then can be saved to a JPEG file. (Any time before this point if the image is saved to an intermediate file, that file should be either a TIFF format or the native format for you editor; but never as a JPEG format.)

Then the 7200x5760 pixel JPEG file is printed as a 16x20, and the 3660x2880 JPEG is printed as an 8x10.

However, if you make use of continuous tones in an image and you insist on leaving individual pixels visible without resamping then you may occasionally have the compression algorithm leave you with a series of dots all over the place.

What is resampling? How do you know you are leaving individual pixels visible? (Maybe this is a Photoshop CS process? I use LR4 and PSE 11.)

"Resampling" is the process used to change the pixel dimensions of an image. For example if you have a 2000x2000 pixel image and need to display it in an 800x600 pixel window on a computer... to get it down to 2000 pixels wide the display software will essentially average every 2.5 pixels and generate 1 pixel for a new image, which will be what is actually displayed. That is "down sampling". If the same image is to be printed it would necessarily have to be "up sampled" to produce more pixels.

I tried doing an enlargement of just one portion of an image the other day, the picture appeared to have a pebble-type surface! Must have stretched beyond normal limits...

That is the effect that we discussed in the previous couple of articles. One of the ways that JPEG reduces file size to save blocks of pixels as if they are all the same. So it might divide then entire image into 8x8 pixel blocks and save each as just one RGB value using three 8 bit bytes instead of 64 separate 8 bit bytes. Big space saver, but...

If you crop out a section of that image, and enlarge it... the 8x8 pixel block is enlarged, and becomes visible. Do that just a little and it is just "poor quality", do it a lot and it is downright ugly. A "pebble-type surface" sounds like you were just into the downright ugly range!

 

[ARShutterbug]

If you print a RAW file onto paper, it will fill many sheets of paper. This is because a RAW file is just lines of code, a representation of the recorded data. It doesn't become a viewable photograph until you develop it. Of course you can print from a TIFF file, but the increase in perceptible quality over JPEG is not worth the increase in the file size.

 

[GreenMountainGirl]

apaflo wrote:

It won't be a problem if you adjust the pixel dimensions of the image correctly before it is saved to a print file.

I could use some help understanding this process.

I looked through my programs (LR4 and PSE 11), and found where to change the image size in PSE 11. It includes a place to check for resampling the image. With an original, untouched raw image on the screen, it gives pixel dimensions as 46.0M, then it gives dimensions in pixels 3264W and 4928H. The next part is document size, with width as the parameter I can change. When I change it, the other numbers change. For instance, I changed to 16"W, resulting in 24.157"H and a resolution of 204 pixels/inch.

And example would be if you get really excited and end up with an Epson printer that can produce 17 inch wide prints! It prints at 360 pixels per inch. So you shoot a great picture of your child, and want to send Grandma a 16x20, plus you want an 8x10 for your desk at work. Lets assume the camera is a Nikon D5100 that produces 4928x3264 images. (You can adjust the various numbers to whatever it might actually be, and then redo the same arithmetic as below to get a "real" example.)

I do not know how to find out the pixels per inch of my printer or the pixel dimensions of my camera images. My camera is a Nikon D7000. I have the camera set to take only RAW, using a 14-bit depth.

First, the image does not have the same aspect ratio as a 16x20 print, so it needs to be cropped. The largest it can end up is 4080x3264. But that pixel dimension would, with a 360 pixel per inch printer, produce an image that is 11.33" x 9.07". That is smaller than the 16x20 you want and larger than the 8x10. So at that point two different "print files" should be produced. One that is (360*20 x 360*16) and one that is (360*10 x 360*8). The first is 7200x5760 pixels and the second is 3600x2880 pixels. Each image, at that size, should be sharpened to taste and then can be saved to a JPEG file. (Any time before this point if the image is saved to an intermediate file, that file should be either a TIFF format or the native format for you editor; but never as a JPEG format.)

Then the 7200x5760 pixel JPEG file is printed as a 16x20, and the 3660x2880 JPEG is printed as an 8x10.

I understand the math, just not sure about how to change the pixel dimensions. Unless the program does it for me when I change the dimensions through the resampling process?

"Resampling" is the process used to change the pixel dimensions of an image.

If you crop out a section of that image, and enlarge it... the 8x8 pixel block is enlarged, and becomes visible. Do that just a little and it is just "poor quality", do it a lot and it is downright ugly. A "pebble-type surface" sounds like you were just into the downright ugly range!

If I use this process on a photo that I am enlarging, will it alleviate the "pebbling" effect that I found?

I really appreciate the explanation you gave, and I understand the theory, just need some help with the execution! Thanks so much.

Susan

 

[John Deerfield]

I could use some help understanding this process.

I looked through my programs (LR4 and PSE 11), and found where to change the image size in PSE 11. It includes a place to check for resampling the image. With an original, untouched raw image on the screen, it gives pixel dimensions as 46.0M, then it gives dimensions in pixels 3264W and 4928H. The next part is document size, with width as the parameter I can change. When I change it, the other numbers change. For instance, I changed to 16"W, resulting in 24.157"H and a resolution of 204 pixels/inch.

The 3264 x 4928 is your resolution. When you change the width or height, the PPI changes to reflect the new height & width. You only have so many pixels (in this case, 3624 x 4928). If you put more of those pixels in a one inch space, you will get a smaller document (NOT a smaller file size because you haven't changed the number of pixels). If you put fewer pixels in a one inch space, you will have a larger document. For most print work, the target PPI/DPI is 300. So, a 3264 x 4928 with a resolution of 300 ppi in Photoshop gives you a "native" document size of 10.88 x 16.427. If we change the PPI to 72 (NOT resampling), those some number of pixels (3624 x 4928) now create a document measuring 45.333 x 68.444. The problem is that for a print you need far more PPI.

And example would be if you get really excited and end up with an Epson printer that can produce 17 inch wide prints! It prints at 360 pixels per inch. So you shoot a great picture of your child, and want to send Grandma a 16x20, plus you want an 8x10 for your desk at work. Lets assume the camera is a Nikon D5100 that produces 4928x3264 images. (You can adjust the various numbers to whatever it might actually be, and then redo the same arithmetic as below to get a "real" example.)

I do not know how to find out the pixels per inch of my printer or the pixel dimensions of my camera images. My camera is a Nikon D7000. I have the camera set to take only RAW, using a 14-bit depth.

3624 x 4928 is your pixel dimensions. As to whether it is tagged as a 300 ppi or 72 ppi image doesn't matter (see above!). the 300 or 72 ppi is merely an instruction set telling how to display or print the pixels.

First, the image does not have the same aspect ratio as a 16x20 print, so it needs to be cropped. The largest it can end up is 4080x3264. But that pixel dimension would, with a 360 pixel per inch printer, produce an image that is 11.33" x 9.07". That is smaller than the 16x20 you want and larger than the 8x10. So at that point two different "print files" should be produced. One that is (360*20 x 360*16) and one that is (360*10 x 360*8). The first is 7200x5760 pixels and the second is 3600x2880 pixels. Each image, at that size, should be sharpened to taste and then can be saved to a JPEG file. (Any time before this point if the image is saved to an intermediate file, that file should be either a TIFF format or the native format for you editor; but never as a JPEG format.)

Then the 7200x5760 pixel JPEG file is printed as a 16x20, and the 3660x2880 JPEG is printed as an 8x10.

I understand the math, just not sure about how to change the pixel dimensions. Unless the program does it for me when I change the dimensions through the resampling process?

For the most part, you don't want to change the pixel dimensions. If you crop an image, then Photoshop will create either fewer (no problem) of more (this is when Photoshop will resample and essentially make up pixels) pixels. As example, you take your image sized 3264 x 4928, if your "crop" it to a 16x24" print, you now have 4800 x 7200 pixels... that is more than your original 3264 x 4928, Photoshop resampled (made up) those extra pixels.

"Resampling" is the process used to change the pixel dimensions of an image.

If you crop out a section of that image, and enlarge it... the 8x8 pixel block is enlarged, and becomes visible. Do that just a little and it is just "poor quality", do it a lot and it is downright ugly. A "pebble-type surface" sounds like you were just into the downright ugly range!

If I use this process on a photo that I am enlarging, will it alleviate the "pebbling" effect that I found?

Yes and no. You can only blow up an image so much. For the most part, with a D7000 and a technically good shot, you should be able to get almost size print. But if you do extensive cropping, you won't be able to get large prints. As example, if you take a full length shot of someone you could get virtually any size print. But if you now crop the image so that it is only a head shot, you won't be able to enlarge that area of the image the same way you could enlarge the whole image: you have deleted a large number of pixels when you did the crop.

I really appreciate the explanation you gave, and I understand the theory, just need some help with the execution! Thanks so much.

Susan

 

[Roberto de La Tour]

HEA-45 wrote:

I am curious about RAW and JPEG file formats. We all know that photographing and post-processing RAW files are considered the best technique. But, in order to obtain a print by an outside source, the RAW file must be converted to a JPEG format. Does this conversion degrade the quality of the electronic image and, consequently, the quality of the printed image?

Note: I love to print my photographs, and I use an outside source. They only accept JPEG format. I have no complaints about the quality of the printed image. I pose the question to satisfy my curiosity...not to solve any specific problem.

IN practice (what my eyes can see): perfectly OK to print from JPEG. No visible loss. Shoot in RAW, post process from this raw if needed, look at it well, convert to JPEG maximum res/quality, look again, and print. It will be just fine even using a very good printer

 

[GreenMountainGirl]

John Deerfield wrote:

For most print work, the target PPI/DPI is 300. So, a 3264 x 4928 with a resolution of 300 ppi in Photoshop gives you a "native" document size of 10.88 x 16.427. If we change the PPI to 72 (NOT resampling), those some number of pixels (3624 x 4928) now create a document measuring 45.333 x 68.444. The problem is that for a print you need far more PPI.

In PSE 11, when I look at the resizing panel, there is a place to check for sampling. Will using this tool, will it do additional sampling so an enlarged photo can be printed without "pebbling"?

3624 x 4928 is your pixel dimensions. As to whether it is tagged as a 300 ppi or 72 ppi image doesn't matter (see above!). the 300 or 72 ppi is merely an instruction set telling how to display or print the pixels.

I think I do understand this part, seems like it means I can only rearrange the dimensions so the total ppi come out the same, not change the ppi, if I want a decent print.

As example, you take your image sized 3264 x 4928, if your "crop" it to a 16x24" print, you now have 4800 x 7200 pixels... that is more than your original 3264 x 4928, Photoshop resampled (made up) those extra pixels.

Are you referring to the CS photoshop, or do the LR4 and PSE 11 do this automatically as well? I suspect the PSE 11 does not, because it has that panel to fill in and ask for resampling.

If I use this process on a photo that I am enlarging, will it alleviate the "pebbling" effect that I found?

Yes and no. You can only blow up an image so much. For the most part, with a D7000 and a technically good shot, you should be able to get almost size print. But if you do extensive cropping, you won't be able to get large prints. As example, if you take a full length shot of someone you could get virtually any size print. But if you now crop the image so that it is only a head shot, you won't be able to enlarge that area of the image the same way you could enlarge the whole image: you have deleted a large number of pixels when you did the crop.

The photo where I got the pebbling was of one of my kitties - as you said, a full body shot cropped to just a head shot. I guess this is why people suggest filling up the screen with what you like best! Of course, with moving subject this is not always possible.

Thanks for your explanations. Hopefully you will see this and help clear up these few more points.

Susan

 

[John Deerfield]

GreenMountainGirl wrote:

John Deerfield wrote:

For most print work, the target PPI/DPI is 300. So, a 3264 x 4928 with a resolution of 300 ppi in Photoshop gives you a "native" document size of 10.88 x 16.427. If we change the PPI to 72 (NOT resampling), those some number of pixels (3624 x 4928) now create a document measuring 45.333 x 68.444. The problem is that for a print you need far more PPI.

In PSE 11, when I look at the resizing panel, there is a place to check for sampling. Will using this tool, will it do additional sampling so an enlarged photo can be printed without "pebbling"?

To a certain degree, yes. To another no! Resampling the image is simply making up pixels. If you start making up too many pixels quality is going to suffer. As a general rule, a 10% increase in pixels is safe. At 20% you start pushing the envelope.

3624 x 4928 is your pixel dimensions. As to whether it is tagged as a 300 ppi or 72 ppi image doesn't matter (see above!). the 300 or 72 ppi is merely an instruction set telling how to display or print the pixels.

I think I do understand this part, seems like it means I can only rearrange the dimensions so the total ppi come out the same, not change the ppi, if I want a decent print.

Changing the PPI is the same as changing the dimensions of the image (with the re-sample box unchecked). Again, PPI is only an instruction set.

As example, you take your image sized 3264 x 4928, if your "crop" it to a 16x24" print, you now have 4800 x 7200 pixels... that is more than your original 3264 x 4928, Photoshop resampled (made up) those extra pixels.

Are you referring to the CS photoshop, or do the LR4 and PSE 11 do this automatically as well? I suspect the PSE 11 does not, because it has that panel to fill in and ask for resampling.

I do not think LR will resample an image. PSE certainly will. In PSE, if you use the marquee tool to perform the crop, this is the same concept that LR would use. You are cropping away the pixels you don't want. When you do this, you are left with fewer pixels: you got rid of some.

If I use this process on a photo that I am enlarging, will it alleviate the "pebbling" effect that I found?

Yes and no. You can only blow up an image so much. For the most part, with a D7000 and a technically good shot, you should be able to get almost size print. But if you do extensive cropping, you won't be able to get large prints. As example, if you take a full length shot of someone you could get virtually any size print. But if you now crop the image so that it is only a head shot, you won't be able to enlarge that area of the image the same way you could enlarge the whole image: you have deleted a large number of pixels when you did the crop.

The photo where I got the pebbling was of one of my kitties - as you said, a full body shot cropped to just a head shot. I guess this is why people suggest filling up the screen with what you like best! Of course, with moving subject this is not always possible.

I would use the marquee tool to crop the image (from the original) again. Only now, go into the image size/resize panel of PSE. You will now see how many pixels you are left with. If you enter 300 ppi (do not resample), you will get an idea of how large a print you can now make from the cropped area. Again, you can only "enlarge" (resample) that area so much (between 10-20%) before quality starts to suffer. There are applications that will do a better jog resizing images than PSE. Perfect Resize is on example. However, there are still going to be limits! I can't take a small web image and resample it to a 16x20: there are simply not enough "original" pixels.

Thanks for your explanations. Hopefully you will see this and help clear up these few more points.

Susan

 

[John Deerfield]

Roberto de La Tour wrote:

HEA-45 wrote:

I am curious about RAW and JPEG file formats. We all know that photographing and post-processing RAW files are considered the best technique. But, in order to obtain a print by an outside source, the RAW file must be converted to a JPEG format. Does this conversion degrade the quality of the electronic image and, consequently, the quality of the printed image?

Note: I love to print my photographs, and I use an outside source. They only accept JPEG format. I have no complaints about the quality of the printed image. I pose the question to satisfy my curiosity...not to solve any specific problem.

IN practice (what my eyes can see): perfectly OK to print from JPEG. No visible loss. Shoot in RAW, post process from this raw if needed, look at it well, convert to JPEG maximum res/quality, look again, and print. It will be just fine even using a very good printer

Just in case this is being lost in this thread... you cannot print a Raw file. Or, at least as one other poster has mentioned, you can't print a Raw file as an image! The Raw file does not exist as an image. It is kind of like a recipe: it has all the ingredients and you can tweak the ingredients to come up with something more to your taste. But sooner or later, you have to cook the file. Unless you are using a printer/service that uses a 16-bit file, then you are printing an 8-bit file. In other words, the same file you would get printing a JPEG. So no, you won't see a difference because there isn't one.

 

[Mike_PEAT]

Roberto de La Tour wrote:

IN practice (what my eyes can see): perfectly OK to print from JPEG. No visible loss. Shoot in RAW, post process from this raw if needed, look at it well, convert to JPEG maximum res/quality, look again, and print. It will be just fine even using a very good printer

If you print the 12-14 bit RAW file on a top of the line printer, compared to printing an 8 bit JPEG, and you WILL see a difference if the RAW file contains colours the JPEG doesn't.

When you save an image as a JPEG millions of colours are thrown away!