Posts
One reason for wanting to start with as many pixels as possible is that every time we manipulate the image, we lose a little quality - resize up or down, get rid of barrel or pin cushion distortion, blend images for depth of field, or stitching, or correct perspective distortion and each of those steps degrades the image a little. This all on top of the original raw file conversion and any manipulation the camera does to raw images - raw isn't always as virginal as one might think as the manufacturer does some noise reduction or some sharpening before you get to play with it.
All these changes do add up. If I sharpen an image, I'm playing with the data and losing just a tad of resolution to gain in sharpness. Run my image through Akvis Enhancer or any other programme that adds local contrast or helps shadows and highlights and even more loss.
Dodging and Burning further damage the image - which is why this is the last thing I do. Even all those fancy blending layers and adjustments distort the data even if they don't destroy it.
And all this doesn't even consider cropping.
Personally, I'll be happy to be have a camera with 36 or more megapixels.
Subscribe to:
Post Comments (Atom)
6 comments:
and I'll be happy to have the money to get such a camera :)))
I guess this is all technically correct but the issue is how little is little. If it's little but negligible that's one thing, and if the aggreagagte of all the changes is noticable - in print - that's another.
At the end of the day I still believe that what drives the resolution you need is how big you want to print (cropping aside).
By the way, sharpening doesn't add resolution, it adds acutance - edge contrast.
I agree to a point - can't have the pixel count outrunning the ability to control noise! But from the looks of things I think we can expect much higher counts in the near future from fx sized sensors.
Maybe I am missing something, but I do not follow the logic of this blog at all. What is the relevance of pixel count to the operations listed? They all change all the pixels within a selected range. If I select a feature then burn it, I alter all the pixels within that selection. If the feature occupies one quarter of the image, then whatever the pixel count I alter a quarter of the pixels in the image. What has the overall pixel count got to do with it? How does a higher pixel count protect the quality of the representation of the selected feature?
I guess I have an underlying problem with undefined quality. For example, consider blending images to achieve depth of field. Why do it? Because the original image does not provide the DOF sufficient to represent the image to the public in the way the photographer wanted (laws of physics!). So the original image is not fit for purpose. You cannot have anything that is simultaneously of quality and not fit for purpose - like a high quality oven that does not cook the Sunday roast! Pixel count changes neither the laws of physics nor a common sense definition of quality.
This reads to me like someone trying to justify purchase of a very expensive camera. Buy it because you want to print six foot by four foot prints. Buy it because it is there and you want it.
If there is a step in the argument I have missed, please publish it. Otherwise, this is misleading and adds to the impression many beginners have that you cannot make a great image without a prohibitively large outlay on kit!
I expected better from the author of great books like 'Take your photography to the next level' and 'Why images work'.
Seems that dear Anonymous (which I wonder why s/he kept its own identity anon.) forgot some facts about having larger count of pixels.
1. If you plan to print and lot of times you need to crop to adjust the framing (nature isn't perfect I guess), then it is better to have a bigger count of pixels to remain in the range of a good resolution and printable size.
2. Higher count of pixels (with addition of a full-frame) provides a lessening in the noise level.
3. Higher count of pixels (i.e. higher resolution) results in .... higher resolution! (Laws of Physics). i.e. greater details, i.e. nicer print.
this is just some.
Most beginners don't stitch, don't blend and don't print large.
One thing to consider is that the Epson 3000 is capable of printing 360 pixels per inch where previous printers really topped out at 300, no matter what you thought you were sending to the printer. At 360 pixels per inch, a 16X20 (the traditionally largest size photographers would print in darkroom days) this would equate to 20X360=7200 pixels, which coincidentally is just about the long dimension of the new D800/E.
Some are delighted when a 16X20 comes out at all, others need it tack sharp and max. resolution. I would never suggest a $3000 camera for any beginner, no matter how much money they have - APS-C cameras are a bit smaller and the lenses sig. smaller and with greater range and serve the needs of many hobby photographers, no matter how serious they are.
Post a Comment