Resize for each size needed?

[OsmosisStudios]

lynweber: If youre shooting RAW, there's an option in your RAW editor for DPI. It shouldnt matter, but I have mine set to 300 as default.

the DPI/PPI difference stems from the graphic design world vs the photography world. In photography, a 12mp image is 4000x3000 pixels because that's how many pixels are on the sensor. In graphic design, if I want a high-quality business card (3.5x2"), then I have to set my DPI to 300.

In photography, you cant adjust how many pixels the camera records. It records 12m, so you cant get any more out of it. In this case, DPI means nothing as it is essentially just a in-software issue.

In graphic design, when you set your DPI, you're actually telling the software how many pixels you want to have for every inch of print. If Im making a business card for instance, again at 3.5x2", a 100DPI image will be 350x200 pixels, 200DPI will be 700x400 and a 300DPI image will be 1050x600.

When it gets to the printer, the printer sees the file, sees what sizes its supposed to be printed at (ie 3.5x2) and prints it at that size with all that lovely resolution. The reason we choose 300DPI is because it's the best compromise between size and quality: lower DPI is noticeable, higher DPI isnt)

In short, as I said, DPI for prints is essentially meaningless. Just set it to 300DPI in your RAW converter and leave it alone.

[BCampbell]

Originally Posted by lynweber

I think that must be the problem it is never 8x10 its always somethng off. IS that the major difference between the full frame cameras and mine?

I have a client that wants to have a 16x20 but it prints funny.

In photoshop i set my DPI to 300 is that all i need to do and then when i upload to mpix it crops a bunch off my photo say at a 8x10 size.

As others have said, don't touch the DPI setting in photoshop.

What's happening, and I'm not sure this is being really clearly explained because there's a LOT to explain here, is that you're talking different aspect ratios. Your camera records in a 3:2 aspect ratio. If you look for the lowest common denominator, you'll see that 8x10" is a 5:4 aspect ratio. It gets even more confusing because many of the popular print sizes are different ratios: 4x6" is 3:2, 8x10" is 5:4, 5x7" and 11x14" are their own ratios (though 11x14" is very close to 5:4), etc.

This means that, like comparing a wide screen TV to a normal, old school set, your image is often going to be wider. In GIMP this is solved by choosing the rectangle select tool and setting the aspect ratio to whatever the desired size is. In Photoshop I'm not exactly sure but it's going to be just as simple using whatever tool you use to crop. Just set it to 8:10, 7:5, 1:1, whatever you need, and crop away.

[lynweber]

So you are saying i should never use the Image size box in photoshop just use the Crop tool or rectangle tool and use the sizes there to crop the image. Should I do this for each size then and save it as its own file? Just trying to get a good workflow.

Originally Posted by BCampbell

As others have said, don't touch the DPI setting in photoshop.

What's happening, and I'm not sure this is being really clearly explained because there's a LOT to explain here, is that you're talking different aspect ratios. Your camera records in a 3:2 aspect ratio. If you look for the lowest common denominator, you'll see that 8x10" is a 5:4 aspect ratio. It gets even more confusing because many of the popular print sizes are different ratios: 4x6" is 3:2, 8x10" is 5:4, 5x7" and 11x14" are their own ratios (though 11x14" is very close to 5:4), etc.

This means that, like comparing a wide screen TV to a normal, old school set, your image is often going to be wider. In GIMP this is solved by choosing the rectangle select tool and setting the aspect ratio to whatever the desired size is. In Photoshop I'm not exactly sure but it's going to be just as simple using whatever tool you use to crop. Just set it to 8:10, 7:5, 1:1, whatever you need, and crop away.

[BCampbell]

Originally Posted by lynweber

So you are saying i should never use the Image size box in photoshop just use the Crop tool or rectangle tool and use the sizes there to crop the image. Should I do this for each size then and save it as its own file? Just trying to get a good workflow.

You'll only need to change the overall image or canvas size when you want to make a smaller physical file, or smaller image dimensions. For example, if someone wants the file for one of your photos so they can print a 5x7" image at 300ppi, you'll want to change the image size to 2100x1500 pixels. This way, if they try to print larger sizes, the image quality will degrade; it's a way to control how your images are used.

For the highest possible image quality, do not touch the image or canvas size. Use your cropping tool (again, in GIMP I use the rectangle select, it's probably slightly different in Photoshop) and constrict it to the aspect ratio you're printing to.

Then, if you want to know the largest acceptable print size you can make, ask your print service or consult your printer's manual to see what resolution they recommend. Usually this is 300ppi up to about 11x14, and starts dropping from there because of issues dealing with the angular resolution of the eye combined with the fact that we view larger images from further away. An example:

Let's say you have a Canon t1i which creates 4752x3168 pixel images. You want to print to 8x10", so keeping the height to 3168 (the 8" dimension) we need to crop the width (the 10" dimension) to 3168/8*10=3960px. Setting our crop tool to 10:8 does this for us. If we want to print at 300ppi, we then just divide the image dimensions to see what our maximum image size is: 3960x3168 / 300 = 13.20x10.56". Plenty good for 8x10. If 150ppi is acceptable, you could print to 3960x3168 / 150 = 26.40x21.12". If you wanted to give the file to someone and ensure they can't print to 8x10" at a resolution higher than, say, 200ppi, you'd then resize your image to 2000x1600px.

And then run an unsharp mask but that's a different topic.

Hope this all makes sense.

[lynweber]

Its beginning to, unsharp mask I'm wondering about that to. Do you have to sharpen every image or do you recommend it? Care to tell me about that?

Originally Posted by BCampbell

You'll only need to change the overall image or canvas size when you want to make a smaller physical file, or smaller image dimensions. For example, if someone wants the file for one of your photos so they can print a 5x7" image at 300ppi, you'll want to change the image size to 2100x1500 pixels. This way, if they try to print larger sizes, the image quality will degrade; it's a way to control how your images are used.

For the highest possible image quality, do not touch the image or canvas size. Use your cropping tool (again, in GIMP I use the rectangle select, it's probably slightly different in Photoshop) and constrict it to the aspect ratio you're printing to.

Then, if you want to know the largest acceptable print size you can make, ask your print service or consult your printer's manual to see what resolution they recommend. Usually this is 300ppi up to about 11x14, and starts dropping from there because of issues dealing with the angular resolution of the eye combined with the fact that we view larger images from further away. An example:

Let's say you have a Canon t1i which creates 4752x3168 pixel images. You want to print to 8x10", so keeping the height to 3168 (the 8" dimension) we need to crop the width (the 10" dimension) to 3168/8*10=3960px. Setting our crop tool to 10:8 does this for us. If we want to print at 300ppi, we then just divide the image dimensions to see what our maximum image size is: 3960x3168 / 300 = 13.20x10.56". Plenty good for 8x10. If 150ppi is acceptable, you could print to 3960x3168 / 150 = 26.40x21.12". If you wanted to give the file to someone and ensure they can't print to 8x10" at a resolution higher than, say, 200ppi, you'd then resize your image to 2000x1600px.

And then run an unsharp mask but that's a different topic.

Hope this all makes sense.

[BCampbell]

Originally Posted by lynweber

Its beginning to, unsharp mask I'm wondering about that to. Do you have to sharpen every image or do you recommend it? Care to tell me about that?

It's a different topic, I mention it because you DO want to run it whenever you shrink your canvas.

In general, it depends on how you shoot, what you shoot, and what you want your finished product to look like. For myself, I'm shooting RAW most of the time so no sharpness adjusts are done in-camera, and usually do a very light pass with unsharp mask at the end of my processing. Different levels of sharpening provide different effects. Sometimes I don't run it on people, or if the photo was shot at a high ISO.

If you're shooting JPG, even if you don't have your in-camera settings set to not sharpen, most cameras still apply some level of sharpening in the process.

And yes, unsharp mask is a sharpening tool even though it doesn't sound like it...