Best methods to get high quality enlargements

[qveda]

While I save up for a full frame DSLR, and/or consider getting a 4x5 field camera for occasional use, I am currently using a 10mpixl point-shoot. I'd like to find out the best methods for big enlargements - up to 20x30.

I've heard that in Photoshop (CS4), you can maintain high quality by increasing image size in small increments. 110% at a time, rather than doing it in one step. Which type of algorithm should be used? I've heard that you should select the one that is 'best for smaller size'.

I'm sure software plugins , such as Fractal, are perhaps best for this. But is it realistic to get to 20x30 from 10mpixls just using Photoshop ?

 

[Ryan1524]

10MP can be plenty for 20x30, depending on the printer and the sensor.

 

[103734]

high qveda you linked me to the wrong thread (you linked to this) in the other post but I was able to find this thread.

So anyways about your question, I have shots that I did with a Canon Rebel XT (its only 8mp) that I have blown up to 20x30, but thats as high as I would go with only 8mp, above that and the quality starts to drop.

And are you planning on printing the pictures yourself? I don't see why you have to blow it up in photoshop first, what size is the resolution of the picture?

 

[compuwar]

While I save up for a full frame DSLR, and/or consider getting a 4x5 field camera for occasional use, I am currently using a 10mpixl point-shoot. I'd like to find out the best methods for big enlargements - up to 20x30.

I've heard that in Photoshop (CS4), you can maintain high quality by increasing image size in small increments. 110% at a time, rather than doing it in one step. Which type of algorithm should be used? I've heard that you should select the one that is 'best for smaller size'.

I'm sure software plugins , such as Fractal, are perhaps best for this. But is it realistic to get to 20x30 from 10mpixls just using Photoshop ?

Probably not the best in terms of image quality if the 10MP are from a small sensor like on a P&S- bu a lot depends on the image, focus and exposure. A 6MP DSLR with an APS-C sensor will go to 20x30 well enough for normal viewing distances. The algorithm depends a lot on the picture. The best bet though is to use a service that has a good RIP and let them upsample the picture, as the RIP will be matched to their printer.

 

[toxic]

for 20x30...i don't think you need to bother, as long as you don't stick your nose into the print. if you're having it done by a lab, they'll upres it for you if they feel like it needs to be.

 

[dmz]

Your ten-megapixel sensor outputs an image that is roughly 4,000 x 2,666 pixels. Your output is 30" x 20". That means your final image, without resampling ("upres"-ing), will contain approximately 133 pixels-per-inch.i.e 4000 pixels divided by 30 inches.

Ideally, you want about 180-240 ppi (pixels-per-inch) for a good, sharp photographic or inkjet print. That's the ideal, but acceptable prints can be made from as little as 75-100 ppi, as long as the viewing distance is beyond arm's length. Most commercial photo print papers can't resolve anything beyond 200 ppi, and while inkjets and dye-sub printers can use more resolution, you don't need more than this for a print which will be seen from arm's-length or greater. So, you're print won't be "perfect", it will be a little "soft", but only if you're looking at it at less than arm's length.

Upsampling (upres-ing) destroys information. Repeated upsampling destroys more information, so, if you do need to upsample (and you DON'T), do it ONCE. Photoshop offers some latitude with the upsampling algorithm it uses, you can choose a sharper or smoother result as well as the standard bicubic resampling algorithm.

And, of course, upsampling a JPEG image is not going to give a good result no matter what algorithm you use. Shoot RAW with a BIG sensor (meaning lots of pixels, not physically large!) if you want BIG prints - in your case, beyond 20x30. Good Luck and

Happy Shooting!

dmz