Originally Posted by darinp2
I am hoping you will clearly address why you think the left side of the image of the bride looks worse than doing the original at 1:1 and zooming to the same size would look.
The reason is the NN sampling I did of the original image I posted of the car and the windowed zoomed image were chalk and cheese. I could have posted screen captures of the observed results, but my intension was to let those who really wanted to see the results for themselves, go through the process so there was no bias.
In fact the projector scaled image of the original on my set up was inferior to the zoomed windowed image.
The above is a living example why.
Please try it out I'll e-mail anyone with the original image set....it's a crystal clear revelation. Words nor discussion required.
Originally Posted by darinp2
If a person had a 1080p projector that displayed pixels as perfect squares with no seams (I know we don't have that) would you still argue that zooming was better than using an anamorphic lens in all cases?
This is an interesting question. There is a limit to the data that can be addressed to a pixel...yes? If not we would have developed higher fill ratios and could have addressed infinite detail to the pixel/s as long as the capture device was indeed capable of capturing it, but alas there is a limit, so we develop panels with more pixels and address more data to these additional pixels.
1st the let's define "the Image sought" to me it's as found on the encode available to us currently via the BluRay disk. I wish to see the image with minimal manuplation, no scaling induced additional components(artifacts).....no additional glass other than the primary lens to act as a filter...etc. True as possible to the original is what I seek with my projected image.
Others may have different objectives.....Does scaling and the use of an a-lens alter the original even by a small amount, of course it does....scaling will have an affect, does an A-Lens affect MTF? it has to, it's not 100% transparent. I'm sure some may even like the result, just as some would like to adjust the tone of a musical track.....to suit their personal objectives.
But why vertically scale a perfect pixel mapped 1920 x 810 image that is perfectly zoomable without panel structure becoming evident to an A-Lens size image when viewed from the optimal angle for 1920, with no reduction in the available precious ANSI, no scaling softening, no MTF reduction, no inducing pincushion and depth of field reduction to gain 270 vertical pixels where the 810 height image detail is spread over an additional 270 pixels to achieve an image that is 10 to 15% brighter?!
Gaining the additional brightness is the primary reason to use an A-Lens but at a cost of purity. Even if we had one big pixel addressed with the 1920 x810 detail, one still couldn't sit closer as the detail adressed would be insufficient to resolve perfectly from the less than ideal viewing angle.......perhaps one day we will have capture devices and projection devices that will have near infinite detail capture/display capability...so we can do so......but 4k will make things better....for now...well soon anyway I hope.
There is no right or wrong, It's the individuals journey....I take the least image manupilated path........I too would love the ability to sit at a closer ratio than I currently do....but I want addtional source data mapped to additional pixels, not existing source data stretched over 25% more pixels. Looks like 4k for me.
Do you use an A-Lens Darin?...............somehow I see you as a pedantic perfectionist......who though debates the pros and cons....from a neutral stance, perhaps even playing the devils advocate at times......but remains a purist.