Even in film times, we sometimes had to « Up Rez » some negatives. Duplicating a 24x36 negative to a 4x5 inch one for contact printing sake. The classical wet print was usually an « enlargement », quality of which was mostly depending of the enlarger lens.
Nowadays we juggle with digital files, their sizes and resolution depending of the camera’s sensor. For an old time Nikon user, it would go from about 4,26 Mpx to 45,70 Mpx (from the D2 to the Z7, with many in between)!
On photographic forums there seems to be a race to the biggest Mpx size, mostly for the « cropping » capabilities (birders) or detail capacity (landscapers) with the general idea that it may be printed « large » in some sort of gallery or at least one’s living room.
However most today just stare at a screen (of different size, definition and color gamut, a whole new game), often zooming in (just to see if it’s sharp) to a pixel to pixel equivalency.
Of course, a greater Mpx size induces a greater file « weight » in Mo ! But as most say, « hard disk space is cheap » ! True enough, but those who dwell with marriages or corporate events, with hundreds of files to sort have a tendency to appreciate smaller files in weight and not to have to buy another « speedy » computer every year. For most of us the process of producing images goes from the camera to the printing or the screen viewing.
In my own case, after going to 24 Mpx with the D3x, I’ve settled to a smaller camera and a smaller file size, the Df and it’s 16 Mpx. Of course I feel the pang of N.A.S. (Nikon Acquisition Syndrome) just as everybody and go all « Wow, Aah, Gee », and other social noises when I see the result (on screen) of the newer « big » files of the Z7.
Still, when I project the way I would be using it, it would be mostly in mRaw or sRaw, with the occasional original size of 45,7 Mpx. With the Df, I travel a bit lighter with a My Passport HD and an iPad Pro (not to speak of the batteries). The latter two being also my usual professional devices as an Architect.
So, I do think that 16M px is quite enough for most of my pictures, but what if I’m asked for an A1 or A0 print of a group shot of the graduates of the year ? Or for some gallery show in A2 or A1 ?
That was my state of mind when I stumbled upon an ad for a new software from Topaz (I don’t know them), the « Gigabyte AI ». With a 30 day free trial, why wouldn’t I download it and try it.
So I did that, and thought that the good people from NG might want to see what it’s about and if it could be useful, « just in case » !
Let’s start with a basic Df file. A part of my desk at 1.00 A.M, no flash, a 28mm f/2.8 AI-s lens with a little stuffed owl as a target (dont look at the mess behind).
Of course, due to the forum rules and ease of use, every picture is reduced to a 800x532 pixel size. the crops are equally sized but at a One to One ratio.
This is the original 4928x3280 pixel picture reduced to 800x532. It was opened in Lightroom 6 and saved as a Tif file in sRGB.
Followed by a 100% crop.
Df base 800x532
100% Df
Then you have the screen of the software: It handles previews or the original file and of the intended output.
Topaz Screen Shot.
Three main menus:
Scaling - by ratios (ex; x4), by width size or by height size. These two sizes can be in Pixels, Centimeters or Inches. When using Centimeters or Inches, you can also define the final PPI.
Two smaller menus called « Suppress Noise » and « Remove Blur » with settings going from None to High (with in between steps).
Output - Allowing to choose a folder somewhere and adding a Prefix or a Suffix. (Even if in the same folder you happen to have the exact same name, it will add a (1), (2), etc.. to the file name, so no crashing a file with another ).
File Format - Settings for most of the usual file formats (but no Raw). In this case going from Tif to Jpeg. The « Image Quality » (like the usual Jpeg settings) and the possibility to change the color profile with a wide choice.
In the
first test, I tried to emulate a Z7 file size (8256 pixels wide). Almost a x2 ratio. (output weight 10,3 Mo in Jpeg high).
In a
second test I converted the 8256 pixel width to centimeters (almost 3 meters !) in 300 PPi. It’s about 32000 pixel wide. A bit more then a x6 ratio! (output weight 446,9 Mo in Jpeg high).
In a
third test I thought of a A1 size at 300 PPI (84 cm) for some gallery print. 9921 pixels and 300 PPI. (output weight 56,9 Mo in Jpeg high).
For each test, the first image will be a reduced picture to the 800x532 pixels size (to compare with the original Df file reduced accordingly), and a 800x532 crop of the picture at 100%.
First test:
8256(Z7)-800x532
100% Z7
Second test:
32000-800x532 (or Z7 at 300 PPI)
100% 32000 pixels (or Z7 at 300 PPI)
Third test:
A1-800x532
100% A1x300
In my Photoshop era, there were plugs and techniques to « Up Rez » files, but I must say that the evolution of Artificial Intelligence for pictures (Smartphones) have reached an interesting level.
Here, the ability to compute the « missing pixel », working with the neighborhood ones is quite fantastic !
Computer time varies between 20 to 45 minutes… So it’s more for a « one in a kind » (though the software allows for batches, set it, and call it a night !!!)…
Of course, a bad picture will always be a bad picture and the IQ of a sensor, the lens it's paired with, the correct settings, will always be paramount whatever the file size.
But I think I will buy this software, « just in case », and continue with my nice little Df files…
