What resolution to choose (From RAW)

elfving73

Howdy friends!

For big prints, when "developing" from RAW, what resolution is preferable. I usually go with the 2048 X 3072 (6,3 MP). If I use the (+)-resolutions, will the image be "interpolarated" then? I figured, since the options is there from the beginning, perhaps it could be of any use?

Regards / Matty

pathfinder

Matty, the explanation I have heard is that it is better to uprez from within RAW, rather than later in Photoshop because the data in RAW is still linear while once the data has been passed into Photoshop it has a gamma adjustment applied to it and thus uprezzing in Photohsop will induce more compromises than uprezzing in RAW. Is that helpful or just a lot of confusing words?
I usually bring the images into Photoshop from RAW in their native size, unless I know I will be needing a bigger file because I plan a larger than usual print.
I import the files from ARC as 16 bit files in ProPhoto RGB, or sRGB if I know it is going to the web. There are a lot of theoretical arguments either way, but the fact is that I can't see much difference usually. For large prints I do uprezz whithin RAW if possible.

Andy

colourbox

I just checked my copy of Bruce Fraser's Real World Camera Raw CS2 and it says "the differences between upsampling in Camera Raw and upsampling in Photoshop using Bicubic Sharper are quite subtle." The sense I get from the text surrounding that quote is that it basically doesn't matter, and the popup is only in Camera Raw for convenience. The advantage of the data being raw affects quality for color and tone, but appears to be a non-factor for resampling.

Factors that might affect your decision: Camera Raw doesn't let you choose the resampling method, while Photoshop gives you 3 bicubic methods plus a few others. Also, because sharpening values are based on the resolution, your decision could be affected by whether you're going to sharpen in Camera Raw or after conversion.

If it's going straight to the web, I'll choose the smallest size in Camera Raw to avoid editing more converted data than necessary. If it's going to be a large print, I'd rather convert and work at the native resolution to save memory and then upsample later, before sharpening and printing. I don't see any point in loading, manipulating, and storing the made-up (interpolated) pixels any sooner than necessary. This is personal preference, definitely not gospel.

pathfinder

Colourbox's reply is straight from page 109 in Bruce Fraser's Real World Camera RAW with Photoshop CS2, and Fraser does indeed say it probably does not matter, particularly if you uprez in PSCS2 with Bicubic Smoother.

He does say that cameras with non-square pixels ( think Nikon D1x) may be different.

Thus I think the answer is it probably is not a significant concern, and can be done whenever it is most convenient in your workflow. I process large 16bit images on my computer without much delay, so I enlarge early. Some folks prefer to edit smaller images, and then uprez just before sharpening for printing.

I want to thank Colourbox for bringing this information to my attention. I read Fraser's book on Camera RAW for PS carefully, but actually have not been diligent in reading the new version for PSCS 2, as I didn't think there was that much different. I learned some thing too tonight

elfving73

Thank you, guys! You are the best!

I think I got an idea of what workflow I'll go with: Open image in it's native size, do whatever pp I need to do, then upsizing in photoshop. I have a rather old and sluggish computor so that'll suit me best. I tried to upsample first, but doin' pp in printsize, even cloning out just a pimp was teadious! (I gotta get myself a faster computor!) How to explain that to my spouse! They never just "get it", do they? Hahaha

Tell me, this Bruce Fraser's Real World Camera Raw CS2 - is it a good investment?

Matty

Angelo

elfving73 wrote:

Howdy friends!

For big prints, when "developing" from RAW, what resolution is preferable. I usually go with the 2048 X 3072 (6,3 MP). If I use the (+)-resolutions, will the image be "interpolarated" then? I figured, since the options is there from the beginning, perhaps it could be of any use?

Regards / Matty

Matty - there's something I'm not reading in the previous answers and wonder what information others can offer on this:

In RAW, your orignal image is a certain size. (your screen shot shows the drop down for "size" not "resolution")

For argument's sake, let's say the original size is 2048 x 3072 in RAW. That's the number of pixels wide by high. In RAW the "resolution" of that image is 300 dpi.

Any experienced printer will advise you can accomplish very good prints at 150 dpi output. My understanding of that information is that I can upsize by a full size and still get good prints. (ex: if 8x10 = 300 dpi ~ 16x20 = 150 dpi)

In the same manner, if you downsize your orignal by half and do NOT reset your resolution the resulting image will be 600 dpi which is way more resolution than needed and results in a huge file size.

I know this is true from practical experience and wonder if anyone can shed light on something I may be doing wrong.

pathfinder

elfving73 wrote:

Thank you, guys! You are the best!

I think I got an idea of what workflow I'll go with: Open image in it's native size, do whatever pp I need to do, then upsizing in photoshop. I have a rather old and sluggish computor so that'll suit me best. I tried to upsample first, but doin' pp in printsize, even cloning out just a pimp was teadious! (I gotta get myself a faster computor!) How to explain that to my spouse! They never just "get it", do they? Hahaha

Tell me, this Bruce Fraser's Real World Camera Raw CS2 - is it a good investment?

Matty

As I said, I read his version about RAW for PSCS and that book was what convinced me that RAW was the way to go and explained each of the controls available in RAW such that they were useful and understood.

As a result, I have not yet read the version for PSCS2, but I suspect it is equally valuable. Fraser has the ablity to explain things clearly and simply and usefully.

I just sat down and thumbed through the new version for PSCS2, and realize I need to read it also. Good stuff about using the Calibration tool to set up profiles for each individual digital camera. SO yes, I think it is a worth while purchase MSRP 39.99 USD.

I have to finish The Canyon Conundrum first though

Sam

Angelo wrote:

Matty - there's something I'm not reading in the previous answers and wonder what information others can offer on this:

In RAW, your orignal image is a certain size. (your screen shot shows the drop down for "size" not "resolution")

For argument's sake, let's say the original size is 2048 x 3072 in RAW. That's the number of pixels wide by high. In RAW the "resolution" of that image is 300 dpi.

Any experienced printer will advise you can accomplish very good prints at 150 dpi output. My understanding of that information is that I can upsize by a full size and still get good prints. (ex: if 8x10 = 300 dpi ~ 16x20 = 150 dpi)

In the same manner, if you downsize your orignal by half and do NOT reset your resolution the resulting image will be 600 dpi which is way more resolution than needed and results in a huge file size.

I know this is true from practical experience and wonder if anyone can shed light on something I may be doing wrong.

Hi Angelo,

Here are my thoughts on this. Looks like both our original files are 3072 ppi X 2048 ppi. (Note: ppi, not dpi) At 300 ppi our photos will be 10.24" X 6.827". This is an aspect ratio of 3:2. When I crop to get a 5:4 ratio, and size for an 8X10 I end up with a resolution of 256 ppi.

If I size for 16X20 I end up with a resoluton of 128 ppi.

My reading seems to sugest a resoluton of between 180 ppi, and 300 ppi for a good quality print. I try to shoot for a minumum of 240, and if after croping I am a little short I might up size in CS2. My reading (can't remember where right now) also sugests that there could be some small quality advantage to up sizing in PS rather than in the RAW conversion.

I have had image files as low as 200 ppi printed outside that came out very well. 200 ppi seems to work well on my Canon i9900.

I sent out 2 image files (prox 10X6.6 @ 300ppi) to EL-CO color labs, and had them printed at 20X30. They came out great! I don't know what they use to upsize, but it seems to work very well. My home experiments show a softer print if I try to double the image file size, but depending on the image could be acceptable.

Sam

Angelo

Sam wrote:

Hi Angelo,

Here are my thoughts on this. Looks like both our original files are 3072 ppi X 2048 ppi. (Note: ppi, not dpi) At 300 ppi our photos will be 10.24" X 6.827". This is an aspect ratio of 3:2. When I crop to get a 5:4 ratio, and size for an 8X10 I end up with a resolution of 256 ppi.

If I size for 16X20 I end up with a resoluton of 128 ppi.

My reading seems to sugest a resoluton of between 180 ppi, and 300 ppi for a good quality print. I try to shoot for a minumum of 240, and if after croping I am a little short I might up size in CS2. My reading (can't remember where right now) also sugests that there could be some small quality advantage to up sizing in PS rather than in the RAW conversion.

I have had image files as low as 200 ppi printed outside that came out very well. 200 ppi seems to work well on my Canon i9900.

I sent out 2 image files (prox 10X6.6 @ 300ppi) to EL-CO color labs, and had them printed at 20X30. They came out great! I don't know what they use to upsize, but it seems to work very well. My home experiments show a softer print if I try to double the image file size, but depending on the image could be acceptable.

Sam

I have a headache I need my slide rule!!!
I guess the simple difference is that ppi is a unit of file measurement and dpi is a mechanical measurement used to determine the quality of output in the printing process. (I think )

When printing I always transfer my images into a graphics program (CorelDraw) size as needed, set the resolution, convert to CMYK and away we go. Of course my office printer may have something to do with this... I use an Epson Photo 1270 which has a 1200 x 1200 dpi output.

Sam

Angelo wrote:

I have a headache I need my slide rule!!!
I guess the simple difference is that ppi is a unit of file measurement and dpi is a mechanical measurement used to determine the quality of output in the printing process. (I think )

When printing I always transfer my images into a graphics program (CorelDraw) size as needed, set the resolution, convert to CMYK and away we go. Of course my office printer may have something to do with this... I use an Epson Photo 1270 which has a 1200 x 1200 dpi output.

Trying to learn about ppi, dpi, and printing can put you on Jack Daniel's permanently. While it doesn't stop the voices as well as the tin foil, you don't really care.

ppi are set side by side very neatly, and can be seen and counted. dpi on the other hand are not side by side, but put down all over the place including on top of each other. As an example if you use the same ink droplet size and other technology as printer A, but add additional ink cartridges / colors in printer B your claimed dpi will increase because you have more cartridges capable of laying down more droplets, IE more dpi.

ppi is the number of pixels per inch the image file has. Dpi is the number of droplets an output device is capable of putting on paper. Each output device (they are different) can only interpret a finite number of ppi to use in determining the number, location, and color of the ink droplets to put on the paper.

As for CMYK, I think I have heard of it. That's it.

This is my understanding of this. If it's incorrect please let me know, I am probably operating on more false assumptions than any one person should, so any corrections would be appreciated. :

Sam

Angelo

Sam wrote:

Trying to learn about ppi, dpi, and printing can put you on Jack Daniel's permanently. While it doesn't stop the voices as well as the tin foil, you don't really care.

ppi are set side by side very neatly, and can be seen and counted. dpi on the other hand are not side by side, but put down all over the place including on top of each other. As an example if you use the same ink droplet size and other technology as printer A, but add additional ink cartridges / colors in printer B your claimed dpi will increase because you have more cartridges capable of laying down more droplets, IE more dpi.

ppi is the number of pixels per inch the image file has. Dpi is the number of droplets an output device is capable of putting on paper. Each output device (they are different) can only interpret a finite number of ppi to use in determining the number, location, and color of the ink droplets to put on the paper.

As for CMYK, I think I have heard of it. That's it.

This is my understanding of this. If it's incorrect please let me know, I am probably operating on more false assumptions than any one person should, so any corrections would be appreciated. :

Sam

"...~~~I'd like to teach the world to print, in perfect harmony...~~~"

We agree. What you said is what I said.