Image comparison: how to ... (1 Viewer)

Hello everybody,
I've been reading very eagerly lots of threads re. comparison of shots taken with different cameras (e.g 40D vs 50D, 1D mkIII vs 50D etc.) and I've been sometimes disappointed by the methods adopted to resize, or crop, or upres etc ..
My question is now: which is the safest and soundest method to effectively (and fairly) compare two images taken with different Mp cameras? I understand that using RAW and leave the image unprocessed or applying the same amount of PP is needed to test sharpness or other elements, but I'm more concerned here with relative size ...

I guess also that this could be a good starting point for beginners as well - why not making this thread sticky? :smoke:

Thanks for your replies,

Max

I wouldn’t get too worked up about this; as to me it’s a pointless exercise. A camera should be bought to obtain individual requirements. The cameras mentioned have all different specifications and price tags. Each person will want to achieve something totally different with their images so they should target the camera that suits their needs.

I can understand what you want to achieve in comparing, but you can’t. Professionally it’s the amount of enlargement factor the image will take before degrading. The larger amount of pixels doesn’t necessary mean better images. Colour quality is also important and overlooked by a lot of people in the quest for faster AF and the latest model, etc.

You can’t really tell how good an image is on screen; you can make an educated guess that it will be ok. The only way to judge would be to have all the images at the same % enlargement, then professionally proof them. This way you can judge colour, sharpness, detail and shadow.

It would be daft of me to buy a D3, if I was only showing images as low res on the web, I would be better off buying a D90 and use the difference to buy better lenses.

I think the best way to compare 2 images, each taken with a different camera of differing MP levels, would be to open both images and view at 100% zoom. This will certainly give you a good way to compare sharpness, especially if the subjects are very similar.

I do think you have a problem viewing the images at less than 100% zoom, because varying levels of zoom can appear differently on different monitors, so you'd never be sure you were really seeing something or it was just the way it was being displayed.

My thoughts on the subject are that with where we are in the game it is a bit pointless comparing image files at the pixel level - i.e. at 100% size. There is no easy way to compare 10MP files with 15MP or 21MP files at that level in a meaningful way. I certainly don't agree with upsizing photos for comparison. What does that prove about the camera/lens other than how good a job the software did at inventing non-existent data. In any case, are we shooting to capture pixels or shooting to create images, art, masterpieces?

Surely the poof of the pudding is in the end result, not the bits that make it up. So for my purposes I am interested in what sort of IQ I can get in my final image. For personal use I view my pictures at home on a 40" 1920 X 1080 LCD TV, from about 8-9' away. That screen requires about 2MP of data to fill it. I could not care less whether my camera has 8MP, 10MP or 15MP (other than the ridiculous file sizes and onerous processing times generated by the high MP cameras). I need to run some proper comparisons but I am seriously considering using my 50D in sRaw1 mode rather than pure raw. Either way, it will be the final image quality I am interested in, not how great each pixel looks.

I do my editing work on a 17" 1920 x 1200 calibrated laptop. If my images look good there from 12-18" away (the laptop screen fills a larger angle of view at that distance than my 40" TV at 8') then I know I've got a good picture.

Personally I don't print my work, but if I did then I would want to see how my photos look in print. I understand the desire to compare pixels but at the end of the day that is not really what photography is about - is it?

I do some wedding photography, as a second shooter, and the end clients there will not be judging what they get at 100% magnification. They will get prints prepared by the prime photographer. He will not waste his time studying my images at 100%. He just wants to know if I have supplied him with a useable image.

I also sell through a microstock agency. The QA may or may not look at my work at 100%, but the clients will probably not. They will be buying images, not pixels.

EDIT : By the way, I should add that for checking an individual image or comparing two from the same camera I see no harm at all in zooming in to 100%, but to compare different cameras with different pixel densities at 100% seems a bit meaningless. The higher the sensor resolution the worse any softness will appear from misfocus, falling outside the DOF, diffraction or movement(camera or subject). It seems sort of ironic that the better the sensor in terms of resolving power (and all other things being equal, like noise per pixel), the worse it will appear, theoretically, when viewed at 100% due to limitations elsewhere in the capture system.

tdodd said:

My thoughts on the subject are that with where we are in the game it is a bit pointless comparing image files at the pixel level - i.e. at 100% size.

In any case, are we shooting to capture pixels or shooting to create images, art, masterpieces?

Surely the poof of the pudding is in the end result, not the bits that make it up. So for my purposes I am interested in what sort of IQ I can get in my final image.

Click to expand...

I cannot but entirely agree with you and pe'rigin ... Although I'm quite interested nontheless in how to do it ... I use two cameras, with or without a TC, and when - during PP - I apply the same corrections, results are quite disappointing at times ... I just want to know how to compare similar shots, in order to grasp any relevant info on how to use my kit better ... perhaps my intentions are not that clear, or perhaps I'm hopeless :king:

I find when viewing on screen at sizes other than 100% I always seem to get a sharper image when viewing at a reduction factor of 1/integer - i.e. 1/2, 1/3, 1/4 or 50%, 33%, 25% if you prefer. It seems to me that the software has an easier time of it to neatly fit four pixels (2X2) into one, or nine pixels (3X3) into one or 16 pixels (4X4) into one than at some arbitrary reduction factor to simply fit the image to the screen. This is certainly my experience with both DPP and Lightroom. I suspect the same is true on final output, if displaying on a fixed pixel device like an LCD monitor.

For the images I display for home use on my TV I do endeavour to crop my 40D files to a 16:9 format of exactly 3840 X 2160 pixels, if I can yield an aesthetically pleasing composition that way. That then allows me to output my files at exactly 50% size for wonderfully crisp and vibrant images that are a joy to behold (he says, modestly ). Of course, I'm not hung up on doing that. If it works then great, if not then so be it.

So, what you might try, if comparing shots with and without a teleconverter.....

- if it's a 2X teleconverter then view the image with the converter at 50% and compare that to the one without at 100%. That will give you a clear indication of which image will give you superior IQ in your final product.

- if it's a 1.4X teleconverter then perhaps you could view the one with converter at 33% and the one without at 50%. This would close the gap on the difference in image size when comparing which will yield the better end product. Either that or simply fit both to the screen and hope for the best in the resizing algorithms that reconstruct the images. If it's 1.5X teleconverter then the 50%:33% resizing should yield identically sized images for your perusal.

EDIT : Another thought - Sometimes it is good to view images side by side. to make comparisons. At other times it is better to toggle back and forth between them and that way any differences can jump out more readily.