Sunday, 10 July 2016

Surprising Results or Something Else?

If you have an open mind about photographic equipment and enough curiosity there is the potential to be surprised by all manner of tools and methods.  Lets see if we can surprise you just as little, from a technical point of view.

The pic at the top is a test pic, but not the whole pic, just a small 1000 by 1000 pixel section, the full pic below shows the complete frame, we'll get to that soon enough. 

Before moving forward let me just explain a couple of things about my testing methods and philosophies, just so, you know, we don’t feed the trolls with misinformation and cause anyone undue confusion.  See I have some "testing" rules, strict rules and I adhere to them almost religiously.

1) When I test a lens (or camera) I never ever pay any attention to out of camera jpegs, they just don’t figure in my world, unless of course I am shooting on my iPhone.  I know you will read all sorts of guff about jpegs being great and easy and all you ever need, spare me…they are rubbish if you are really interested in knowing the worth of a lens or sensor.  I know most punters don’t see the difference “on line”, well let's be frank here, when was the last time you saw a raw file actually displayed "on line" in a raw convertor application.  Never, all you have seen "on line" are compressed jpeg versions of the extracted RAW files, trust me it is not the same thing.

2)   All my RAW test files are converted and processed in just one RAW convertor, Iridient Developer.  There are significant differences between the detail and renderings that come out of different RAW programs, Iridient (a Mac only application) in my experience is the one that extracts the most information out of the any raw file.

3) Lenses are only tested at the lowest native ISO setting on the camera, trust me on this, you can only extract the maximum possible detail that the lens can project onto the sensor if the noise levels are kept at their absolute minimum.

4) I don’t turn on the distortion correction option in the RAW convertor, most folk probably don’t realise this but the corrections can seriously erode the edge and corner definition of an image and in most cases the correction is not needed.  I would rather apply it on a "case by case" basis.  Turning it off also quickly alerts me to any major distortion problems the lens may have in its native state.

5) I always custom fix the Chromatic Aberration, often the pre-potted settings work fine but the fact is they are rarely perfect, there are actually sample variations from lens to lens even of the same model and in any case I can often dial up a better setting set myself.

6) I leave the vignetting correction off, that way I get a better idea of any vignetting issues, but I also try the file with it on to see if the degree of correction required is going to mess with the tonal rendering.

7) I take lens test shots at pretty much the same location under the same light (bright sunlight), that way I am comparing apples with apples so to speak.

8) I never take test shots of brick wall or close up charts, (unless I am testing a macro lens or for macro capability) all shots are 3D subjects in the near, mid and infinity distance ranges.  God knows why people run tests of brick walls, maybe there is some odd sect of photographers out there whose lifelong work revolves around recording the masonry arts, beats me!

Ok, so all clear then, good, now I ask that you please remember these above aspects when you look at future posts on lenses and camera tests.

So back to our test pic…..

Now the original frame (below) the section was lifted from is a file a tad over 33 megapixels, bet that has the keen amongst you guessing, what camera shoots 33 megapixels?   Ah actually none, but I will come back to that.

So allowing for the odd resizing and damaging compressing images for a blog site does, in essence what we are looking at in the 1000 by 1000 px crop at about a 100% view.  Let's put that into a little more perspective, if you printed the image at 300 psi (pixels per inch) you would end up with a print about 22 by 17 inches, pretty big but of course you would not generally be looking at such a print with your nose just a few inches from the papers surface.  Anyhow this file would easily print tack sharp to 36 by 27 inches and much bigger again if your viewing distance relaxed a bit.

After looking at the full frame above something might be tinkling around in the back of your photographers brain… is not a 3:2 ratio image, ……actually it looks kinda like…..well, 4:3.  Something fishy going on here eh?

Ok we’ll come back to that, but lets go back to the 1000 px section, now even allowing for the rubbish the net inflicts on images I bet you can see that is pretty detailed, I mean you can actually make out the barbs on the barbed wire just above the chain link fence, (at least I can on my screen before the web does its business), you can almost see the cross frames in the windows on the long white building in the background, which I might add is a good kilometre from the camera (in fact in the original in Photoshop you can see them at a 200% view.  Heck even the blades of the dead grass inside the compound are pretty easy to see.

So this is a very sharp test pic.  OK you want more, everyone wants more.  The section is actually from the outer edge of the frame well away from the centre where lenses usually resolve at their best, and I can vouch that right on the frames perimeter it's actually no worse.  So this must be an expensive and rather good lens?

Well actually it is not an expensive lens, but yes its rather good and maybe a bit better than it might first appear.  You see this is a kit lens, just a lowly kit lens and it is being used just beyond the middle of its focal range where its performance actually starts to tail off a little. Intriguing no?

More intrigue, the lens is being shot wide open! Yes one could not consider f5.3 a particularly wide aperture but nonetheless that's as wide as this puppy goes at 28mm.  Oops that rather lets the cat out of the bag, the 28mm focal length being beyond the mid range of a kit lens, well yes it is if the cameras sensor format is M4/3.  

That’s right folks this is a kit lens on a Micro 4/3 camera, in this case a Panasonic series 2 14-42 mm jobbie.  As kit lenses go it really is rather excellent despite the basic specs and utilitarian intents.

The camera in question, an Olympus EM 5 Mk2.  Did I just hear someone go ah ha, I'ved heard about that camera, I know what you did Braddles, you shot it using that new fangled hi res mode where the camera takes 8 frames and blends them together for 40 plus megs and then you cropped it…didn’t ya, didn’t ya, fess up?

Ah, um, so sorry but you are completely wrong, this is actually a single frame, hand held and shot at just 16 megapixels, the file is 33 megapixels because the image is so detailed it can be uprezzed easily in the raw conversion stage.

So fact is that this lens is really very very good, the second fact is that M4/3 is capable in conjunction with the right convertor of producing some technically amazing results.

But there is something else happening here and I am not going to let the cat out of the bag too much, as it is actually a bit of a trade (my trade) secret, but allow me to tease you a little for the time being.

Look at the full frame and ponder this for a moment. This shot is taken in bright and fairly contrasty sunlight, agreed.  Ok now look at the highlights, the white building on the left, the white on the platform, the almost specular highlights on the light at the top of the pole.  There is no tonal clip off, everything is rendered and detailed.  Look at the tonality of the sky, absolutely natural.  

Now take a look at the shadows, you can see details in the deep shadows under the carriages. Take a peek in the trees in the background, you can see into the deep shadows inside the trees and make out the details.

This "out of the RAW convertor image" has a very long tonal range, very film like, the colour is muted but it could be easily boosted in saturation.  However, where the saturation should be high, like on those orange/red safety flags, it is just right! This is a file could be now edited with impunity and still hold together perfectly, add contrast, boost saturation, stretch the worries.

Is it a HDR image, stacked image….nope, like I said it is a single frame.

Take another look at the crop, can you see any noise at all, nothing.  Think about that for a minute, this file has been optimally sharpened to bring out every little fine detail and then uprezzed to 33 megapixel, by any normal accounts there should be some obvious noise showing, both chrominance and luminosity…but nothing, nada, not a spot.

So now if you have had some experience in the dark digital arts you would be wondering, how is this possible? How is it that a lowly, cheap kit lens on a 16 megapixel M4/3 camera can end up with image quality as good or better than say a Nikon D810 or Sony A7Rmk 2  (Bear in mid folks that if this image was a frame taken by those cameras cropped to the 4/3 aspect ratio it would have been almost the same number of pixels as our test pic, so its a fair comparison.

Well like I said, I am not giving the game away, but I can tell you this, I have been doing this very thing for years and it is part of a full system, and I have run workshops on it.  It is aimed at serious photographers who want an edge and it has a name, “True Light Capture” 

Stay tuned cause pretty soon I will be announcing the next round of TLC workshops and I will tell you a little about it.

No comments:

Post a Comment