The GetDPI Photography Forum
Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!
Lunchtime with the A900 and Zeiss 135
- Thread starter ryc
- Start date
Greg Seitz
New member
It's fine at lower ISOs but I disagree with Edwards claim that it has better looking noise regardless of ISO.
The noise on all of these 20+ Mpix looks reasonably good up to a point due to the very small pixel size but the more noise A vs B has the worse it will look. No getting around that fact because with more noise you lose more fine detail as shown above. Well, there is one exception and that is if you start getting pattern noise like banding which we perceive much more easily.
And with respect to the Nikon I haven't used a D3X but if every other Nikon that I've used is any indication the reason the files look washed out and lifeless at higher ISOs is simply due to the approach Capture NX 2 takes to suppress noise, not anything in the underlying file. Use another raw developer and you will see a very different outcome.
Come on guys, these are all great tools. The fact is they simply have different strengths and weaknesses, nothing wrong with that.
The noise on all of these 20+ Mpix looks reasonably good up to a point due to the very small pixel size but the more noise A vs B has the worse it will look. No getting around that fact because with more noise you lose more fine detail as shown above. Well, there is one exception and that is if you start getting pattern noise like banding which we perceive much more easily.
And with respect to the Nikon I haven't used a D3X but if every other Nikon that I've used is any indication the reason the files look washed out and lifeless at higher ISOs is simply due to the approach Capture NX 2 takes to suppress noise, not anything in the underlying file. Use another raw developer and you will see a very different outcome.
Come on guys, these are all great tools. The fact is they simply have different strengths and weaknesses, nothing wrong with that.
carstenw
Active member
Interestingly, if you downsize the two crops of the branch to 50%, a better test for what it would look like printed (and let's face it, when else do you need the full resolution?), then they look very similar, with the A900 having a tad more detail and a tad more definition in dark areas. There is still some large blotching, but they are much closer. The A900 crop is a touch lighter in general, so darkening it a little would bring them even closer.
I am not saying the A900 is as good as the 5D2 in dark situations at high ISO, but the difference in real-world terms is much smaller than a 100% crop would lead one to believe. In fact, it is so small that at least for me it would not be a decisive factor in choosing between the two. And I expect the situation at ISO 1600 to be even better.
I am not saying the A900 is as good as the 5D2 in dark situations at high ISO, but the difference in real-world terms is much smaller than a 100% crop would lead one to believe. In fact, it is so small that at least for me it would not be a decisive factor in choosing between the two. And I expect the situation at ISO 1600 to be even better.
Yet, in reality it will be much worse even at ISO 1000.
As Irakly constantly reminds me, you don't normally shoot ISO 1000 to 2000+ in nice light.
The real-world test is ugly light, or very contrasty light with fall off in the shadows.
I have pals that use the Canon 5DMKII ... and while it's not my cup of tea, I cannot fault it's performance at the higher ISOs. IMO the only better low light camera is the Nikon D700/D3 at the cost of resolution. But as you say, how often do you really need that kind of resolution ... which landscape shooters may disagree with, but I wouldn't for my wedding work which is 80% 8'X10" prints.
Properly used (AKA, sparingly), Nik Define 2 is the A900s best friend.
I have agree with you Greg.
"One size fits all" doesn't apply to camera gear. Use the tool for it's strengths, and avoid it's weaknesses.
IMO, the only wide range ISO, high meg camera on the market today is the Canon 5D-MKII. But it's a compromise through out that range ... with cameras like the A900 and D3X being better at lower ISOs, and the D700/D3 being better at high ISOs. But for a single camera solution the 5D-MKII comes the closest.
My preference is a 2 camera solution with less compromises for their intended use.
Personally, I don't think you are being critical Shelby ... you're simply defining the perimeters of use. It's sound judgement based on real-world experiences rather than fanciful wishfull thinking in some ill concieved attempt to justify a purchase decision, or whatever motivation that leads to unrealistic performance claims on the internet.
I don't drag out my Hasselblad H3D-II/39 to shoot stuff it wasn't designed for either.
Fact is, for most people, in most situations, the A900 is a stellar performer ... maybe the best solution for the money available today given the resolution, SSS and the Zeiss AF optics. Can't fault enthusasism for that.
jonoslack
Active member
I quite agree - pretending it's something it isn't (i.e. a fast autofocusing, low light, speed demon) belittles the thing's it IS good for - i.e. good light, fantastic lenses, lovely files.
I have to shoot a concert in a church tonight, and I have a couple of weddings at the end of the month, I'll make the A900 work for me (together with the M8's) , and the 135 f1.8 will certainly help this evening, but I'd rather be using a D700 (at least for the concert). I don't do enough of this kind of stuff to warrant having a separate system, and the A900 is better for what I do most of, but I don't have any illusions that it (or any other system) is going to be the best for everything.
edwardkaraa
New member
Georg,
When I first saw your files, I thought the second one was from the A900, as this is what I usually get at iso 3200 processed in IDC with NR off and then removing the chroma noise only in noise ninja. I am not sure if this particular to raw developer, but I definitely do not see any of my high iso files looking like this. In any case, similar settings do not give fair results as files of different cameras require different settings on the same converter.
I would be very interested to have a look at the A900 raw file if possible.
Well, hate to burst that bubble ... but preliminary tests I just did to get up to speed on the D3X is pretty much in conflict with everything you say here.
Did 20 different side-by-side shots with both cameras using the 85/1.4 lenses from each system (a disadvantage for the Nikon given that the 85/1.4 isn't the sharpest knife in the Nikon drawer : -) . Used different real life lighting scenarios (from crappy to okay), and different ISOs. Here are some high ISO shots ... where, contrary to internet chatter, the D3X smokes the A900 hands down ... and these are examples where the scenarios favored the A900, so it did a little better than it did on others.
Didn't do much of anything to any of the files except crop in on them ... while I compared and inspected them in C1, I processed them in CS4 ... in C1 the Nikon comparatively does even better especially in the shadow noise. I did note that the noise structure itself is almost identical between the two cameras (no surprise there since it's the same sensor maker). Color out of the cameras was pretty similar ... with the Nikon a bit more true to the actual subject ... in other words, leaning towards a more neutral feel.
I'd say, at the higher ISOs, the D3X will respond to Nik Define-II better than the A900 because the noise is less blotchy.
One odd thing I noted was how each slightly differed in metered values when using aperture preferred. So to even the playing field, I metered manually. That way, all the camera settings were identical.
I'd say that for my applications and use, the D3X is the best 35mm DSLR available today. Now, that doesn't take anything away from the A900 given the price and feature set, or someone else's application criteria.
Greg Seitz
New member
Sure, here you go, this contains the Sony and Canon raw files:
http://www.mediafire.com/file/mdmq2rzqncz/A900-5D2-3200.zip
Thanks,
Greg
Greg Seitz
New member
Marc,
Thanks for showing those. I'd say based on the relative differences I've been seeing and the same that you've shown you might be able to infer that the 5D2 and the D3X would be very close indeed in the upper ISO range and there is certainly plenty of evidence that shows the D3X is cleaner at low ISOs. Now where is that D700X?
Thanks,
Greg
Thanks for showing those. I'd say based on the relative differences I've been seeing and the same that you've shown you might be able to infer that the 5D2 and the D3X would be very close indeed in the upper ISO range and there is certainly plenty of evidence that shows the D3X is cleaner at low ISOs. Now where is that D700X?
Thanks,
Greg
Yep, and I just started working with the files. I'm sure there's more in there ... just takes time.
No matter what, it has to be a simple processing proceedure .... 500 files from a wedding dictates that can't be some complex ordeal ... I just don't have the time to "noise control" every freaking shot from a dark reception hall. Which is why I kept the D3 to work with the D3X.
Still love the Sony ... my Summer-time camera ...
Thanks for posting all these high iso comparison shots between the A900, D3x and 5d2 as it is nice to know and see the differences in a fair way. There is no doubt that the D3x is doing a lot of noise removal in camera even on the RAW file which I don't like because I would prefer to do the noise removal in post with noise-ninja and have choices in the tradeoff between losing detail and removing noise. It would be interesting if you ran noise-ninja on the 5d2 and A900 files to see how they compare to the D3x file on the same scene with lots of tiny detail. The candlestick shot only has detail in the candlestick holder part of the picture and the rest of the picture doesn't have alot of fine detail. Maybe D3x's noise removal "secret sauce" is better than programs like noise-ninja but I am sceptical till I see a comparison shot done like I mention above.
Also, can't you just batch run the 500 pictures through noise-ninja, don't know how long that would take so maybe that is not an option if you need fast turn around.
Also, can't you just batch run the 500 pictures through noise-ninja, don't know how long that would take so maybe that is not an option if you need fast turn around.
S
Shelby Lewis
Guest
Just thought I'd add that I don't think the a900 is terrible at 3200... great, no.... terrible, no. It's just not up to the standards we're all coming to expect. However, at moderate print sizes 3200 is actually quite usable. You HAVE to take resolution into account.
An example.
100% crop... not beautiful for sure:
but at roughly 5x7 or 8x10 it would be just fine in print... this is only resized and corrected (somewhat) for color with color noise reduction in (gasp) ACR ... but you HAVE to expose correctly... and I think RD's detail retrieval is actually detriment to the a900 in high-iso situations as it make grain overly sharp and apparent:
Headin' out to shoot... later.
Shelby
An example.
100% crop... not beautiful for sure:
but at roughly 5x7 or 8x10 it would be just fine in print... this is only resized and corrected (somewhat) for color with color noise reduction in (gasp) ACR ... but you HAVE to expose correctly... and I think RD's detail retrieval is actually detriment to the a900 in high-iso situations as it make grain overly sharp and apparent:
Headin' out to shoot... later.
Shelby
I have the advantage of viewing the full screen shots, and the D3X isn't losing detail at High ISOs ... the wick in the candle shot is just as detailed as in the A900 shot.
When you take the file waaaaay up there isn't any more mushy pixels than there is in the A900 shots at lower ISOs.
My intention isn't to use the D3X for high ISO work ... I have a D3 for that. But it's good to know what the boundries are just in case.
The ISO 800/1000 files from the D3X are really very good. Which is an important ISO for my work.
No, applying some noise reduction program wholesale is not an option. It just smashes the pixels unless it's applied sparingly and selectively. A couple of problem files, sure ... a whole batch ... nope.
edwardkaraa
New member
Marc, you didn't burst any bubble. I have never claimed the A900 had less noise. Still to my eyes, your A900 files look better, they just need a slight touch of chroma noise reduction. I understand the misunderstanding, you are looking for noise free files, for me noise by itself is not the criteria to judge a file's quality. Once chroma noise is removed, I actually like the luminance noise because it looks like film grain. As you have seen, Shelby is adding film grain to many of his wedding shots to make them look real.
edwardkaraa
New member
My version (by the way, whatever lens you used on the Canon, it seems sharper in the corners than the ZA 24-70 at 70mm )
Edit: Sorry the photo seems to have been automatically downsized when uploaded to the post.
View attachment 17667
)Edit: Sorry the photo seems to have been automatically downsized when uploaded to the post.
View attachment 17667
Sony A900 3200 ISO 100% crop
To this:
Canon 5D Mk II 3200 ISO 100% crop
Both processed in Raw Developer with the same settings - the same small amount of chroma noise reduced on both. Not seeing the plasticky Canon files everyone keeps talking about. I'd be happy to provide the raw files for both if you'd like to take a crack at them.
Last edited:
edwardkaraa
New member
Here is IDC vs. DPP:
