I shoot micro 4/3s (MFT), so my image sensor is a 20.1-megapixel sensor that produces raw files that average in size around 17 to 18 megabytes.
To scale the images, here is the process I follow. Each image can take up to 5 minutes, depending upon size. That is because it uses AI to create the missing pixels to come up with a proper scaling that interpolates new pixels that work with the image. This is a slow process and CPU intensive. Gigapixel AI uses artificial intelligence to look at the image compared to millions of similar images and creates new pixels with this algorithm. However, significant changes in size of images can be particularly problematic.įor scaling larger, I have found the best way to increase the size of an image, as of the beginning of 2020, is to use a product from Topaz called Gigapixel AI. Even for a 36-megapixel image, you only end up with printable dimensions of 24.5″ x 16.4″.Īs you get close into the image, you can see the pixelation.Īlthough Photoshop has improved much of its algorithms for image size changes, these work reasonably well for smaller changes in size. If you use 300 dpi to give you a print, the maximum size is 14.3″ x 9.3″ (this is not overly large). For example, a 12-megapixel image (common cell phone resolution) on a 3:2 sensor translates into pixels that are roughly 4290 x 2800 on the sensor. If you simply use pixels, there is no weird conversion. This means that megapixels divided by dpi will give you the maximum dimensions natively produced by your sensor.
When you print, typically, you are looking at print resolution, in dots-per-inch (dpi) rather than megapixels, that is appropriate for the medium and the size of print you are going to make.įor most print media you might hold in your hands, you need a minimum of 300 dpi. In order to actually print it, you need to figure out how big you want your print and then do some very basic math to figure out what works best for your image.
Megapixels just provides you with the total number of pixels.
0 kommentar(er)
