Heinz, very nice job. I will say that if you used an M9, maybe even an M8, I believe it would be sharper in your crop area. That being said, I highly doubt that the M9 would have handled the bright lights from the street lamps as well as film did. About 2/3 over from the left in the center of the image the street lamps do not blow the highlights at all....which I see the M9 doing very often which is almost impossible to recover.
I would be very happy with the image you posted above
As the image was overexposed by 1 stop the highlights are actually blown. Should have done some bracketing, but I am really an idiot when it comes to night shots.
Ah, one other thing: I am waiting for a correction lens for my viewfinder and I could not focus at all. Distance was simply set to infinity, distance to the house in the crop is about 300 meters. But I doubt that that has any influence.
Heinz....look at the photo below taken with the M9. Look directly left of the fountain in the background. That is where I *think* film would handle it better
Here is another one. Look at the blown highlights at the base of the church. I will have to do a side by side test
BTW...do you guys recognize the church above? It is shown at the beginning of Forrest Gump where the feather falls from the steeple. The bench Tom Hanks sits at in the movie is the square that I was actually standing in when taking this shot.
Ryan, I see what you mean. But, as I said, my mistake. I just checked again and actually my image was 1.5 stops overexposed. All greenish light streetlamps and the white reflections in the river are RGB each at 100%. Totally blown. As this was shot from my balcony I might reproduce it and do some bracketing if anyone is interested in this.
I have some medium format shots from Ethiopia over in the Zeiss thread that Carsten started. Also extreme contrasts, but properly exposed for this challenging light situation. Nothing at all is blow there. So, in general I totally agree. I have some images shot with the 5DII of that situation in Ethiopia and the highlights are totally blow. I just don't agree on my example above
Ryan, nice sets with the 50 Cron. I really like its thick sort of "glossy" rendering.
Charles, great shots and excellent PP!
Heinz, thanks for posting the scan crop. I'd say that's about the equivalent of what you would get from a 5-8 Megapixel AA-less camera.
Here are two equivalent M9 shots. The images have been uprezzed to 30 megapixel (6631 x 4413 pixels) before I cropped them (the native M9 resolution is 18 megapixel, 5186x3451.
I don't have a 50 year old lens, but I have a 30 year old, also 90 mm - the 90 Summicron-M (V3).
Interesting comparison Heinz and Luka, but I am curious about one thing: The final crop of Heinz' shot is exactly 1024 wide, which makes me wonder if it was actually upscaled by the software used to crop, rather than exported at its native pixel size? It also looks like it might represent a smaller part of the image than Luka's crops. This would make the comparison somewhat troubled.
carstenw wrote:
Interesting comparison Heinz and Luka, but I am curious about one thing: The final crop of Heinz' shot is exactly 1024 wide, which makes me wonder if it was actually upscaled by the software used to crop, rather than exported at its native pixel size? It also looks like it might represent a smaller part of the image than Luka's crops. This would make the comparison somewhat troubled.
As said, my crop is 0,22% of the original image size, so it represents a much much smaller area as Lukas crop. My mistake was, and I thought about that in the car whilst driving home, to post a magnification of the crop, blowing up 319px to 1024px.
Yes, an original sized crop would be preferable. Posting a 300 Megapixel crop is a bit overkill If you set it to 18 megapixels then we can directly compare it with the M9's native resolution.
In the mean time, two not previously posted 90 Cron shots from Bodo/Lofoten:
The 90 Cron has a very nice look to it, but for your uses, I feel that it is really missing a bit of bite. Have you been able to try a 90 APO Cron ASPH yet?
Carsten, yep, I tried one for a day a couple of weeks ago. Optically it's certainly better, but I wasn't sold on it. Funny you should mention 'bite' - that's what I thought the APO lacked - i.e high contrast in low and medium spatial frequency components.
I'm not 100% sure but I think that is what I found lacking with the 75 Cron ASPH. The use of 'lacking' is rather nonsensical as the problem can be described as too high/even contrast at all spatial frequencies stopped down. Good edge contrast depends on the mid spatial frequencies to have a lower contrast than the low frequencies. If they are equally good the distinction between coarse detail (edges etc) and mid level detail gets lost.
So I don't currently feel too much of an urge to get the 90 Cron AA. My main objection against the V3 pre-ASPH was strong CA in the edges, but Lightroom cleans that up very nicely.
denoir wrote:
Carsten, yep, I tried one for a day a couple of weeks ago. Optically it's certainly better, but I wasn't sold on it. Funny you should mention 'bite' - that's what I thought the APO lacked - i.e high contrast in low and medium spatial frequency components.
I'm not 100% sure but I think that is what I found lacking with the 75 Cron ASPH. The use of 'lacking' is rather nonsensical as the problem can be described as too high/even contrast at all spatial frequencies stopped down. Good edge contrast depends on the mid spatial frequencies to have a lower contrast than the low frequencies. If they are equally good the distinction between coarse detail (edges etc) and mid level detail gets lost....Show more →
Well, I see what you are driving at, but I feel that if the resolution is there, which it is to a greater extent in the APO 90, then it would be a question of finding good default LR3 or CS5 settings to give the pop where you want it. Of course it is great to just get the file from the camera as you want it, but not every lens can be a 50 Lux ASPH.
So I don't currently feel too much of an urge to get the 90 Cron AA. My main objection against the V3 pre-ASPH was strong CA in the edges, but Lightroom cleans that up very nicely.
That, and I think you mentioned that it never gets completely sharp stopped down, right?
Excellent Heinz, thanks for the DNG. I think I have a better comparison now.
We start at 18 Megapixels for both (your DNG resized to 18 from 30). I'm using this image from the M9:
Comparison:
Ok, so 18 Megapixel was waay too much for the 135 film scan. Let's try 2000 px width instead (2.5 megapixel). I'm using a different image from the M9 in order to have objects that are of a comparable size. This is the image I'm using this time:
And the comparison:
The M9 still has an edge here at 18 vs 2.5 Megapixel, but it's much closer now. So I'd guess that the 135 shot has the resolution of about 2 megapixel when maintaining the per pixel sharpness of the M9. What do you think?
Edit: I made an error in the first version of the post where the first film image was the 2.5 megapixel version and not the 18 megapixel. I've fixed that now but you may want to reload the page if you did not see this edit the first time.
I think your tongue is somewhere deep in your cheek
I think you need to match lenses and light conditions. 30 second night exposures are of a very different nature than bright daylight conditions, and this needs to be equalized. I also think that since the M9 is natively 18MP, it isn't fair to just resize the film scan to that. Some attention should be given to careful sharpening etc., or perhaps both should be sized to 24MP instead. I didn't even mention the 50 year-old lens yet. Try using the same lenses, guys
Pretty much what I think, Luka. Your examples where also without any sharpening, I asume.
If you like, lets do the same comparison on another image. I'll upload another DNG tomorrow.
Edit: the other one is daylight and looks way better at 100% than the nightshot, Carsten. Still, I thinks it is pretty much ok what came out under the really bad conditions the image was taken.
carstenw wrote:
Well, I see what you are driving at, but I feel that if the resolution is there, which it is to a greater extent in the APO 90, then it would be a question of finding good default LR3 or CS5 settings to give the pop where you want it. Of course it is great to just get the file from the camera as you want it, but not every lens can be a 50 Lux ASPH.
I think the problem is when all the different detail levels have very high contrast - at least with the type of web resize & sharpening that I know and use. A 75 Cron (and 90 AA for that matter) shot looks better at 100% than the 90 Cron V3, but once I start sharpening and resizing I'm amplifying the finer detail in each step to be preserved in the smaller image. It tends to work well for course and fine detail, but in my experience if the mid level detail is exceptionally high the end result gets messy. You simply cram too much high (micro)contrast detail on a small area.
It's possible that there is some sort of PP workaround and that a clever use of multi level USM can solve it, but with my current technique I end up with images (at web size) that I'm not quite pleased with.
That, and I think you mentioned that it never gets completely sharp stopped down, right?
It never gets as sharp as for instance the 75 Cron, but you have to engage into some serious pixel peeping to notice it. In terms of resolution stopped down, the difference between the 35/2 Biogon and the 28/2 Cron are on the same order as the difference between the 75 Cron and 90 Cron V3. It does affect the drawing style, but I don't consider the sharpness of the 28 Cron to be problematic, just because it doesn't quite reach the Biogon. On the contrary, the difference in the contrast of the finest detail give the lenses somewhat different drawing styles - which is a good thing. The same thing with the 90 Cron V3.
CA was the real issue - now that I've found a good software fix for it I don't really have the slightest urge to get rid of the lens.
If you set it to 18 megapixels then we can directly compare it with the M9's native resolution.
