So I just got my NEX-7 back yesterday evening from Lifepixel (I sent it out last Monday). Overall, I'm very satisfied with the service. They packed it well and even cleaned my NEX-7. The AA-less filter stack works right out of the box. docmass from dpreview had to send in his NEX-7 to Maxmax twice since the first time the result was actually worse.
There is definitely a tangible difference. Below is a quick comparison (left is AA-less) at 100% and 200% with 2sec delay on a sturdy tripod.
Considering the AA-less picture was taken under a lower contrast scene, the difference in the fence would be more noticeable under the same lighting. Regrettably, I couldn't borrow another NEX-7 to get the exact match. As for WB, I would have to shoot more.
Unfortunately, I don't have any "problematic" RF wide to test. I'm considering the CV 15 for a small wide on my NEX, but it's not the priority at the moment. However, I can definitely do an aperture series if you can send me yours to test .
Regardless of the lighting being so very different I would say that the level of detail between the two is negligible at best but the shots from the modified camera have a lot more noise.
If it were me I'd send it back in and have them restore it.
The noise could be easily attributed to the left image having more blue channel than the right, which may be a noisy channel on the NEX 7.
It would definitely be interesting to see strict A/B comparisons, but I can already make out the fence detail a bit better with the AA-less camera, which is pretty cool
Both were shot at ISO 100. As for the lighting, I guess until I can get another NEX-7, there is no clear conclusion. There was a firmware change as well. I wish there is another NEX-7 user in NM that we can meet .
What? That's ISO 100 for the Nex7? Wow... I was thinking it was around 800 or so given its reputation - or at least the reputation of the 5n which I know more about. Dang.
cputeq makes a good point. So I guess if it were me as I suggested, I'd wanna make some more tests before deciding anything.
Here's some samples from the GH2 and why I thought the NX7 must have been set to around ISO 800:
100% crop from the RAW all controls centered or zeroed in ACR convert to JPeg - No PP - ISO 400
100% crop from the RAW all controls centered or zeroed in ACR convert to JPeg - No PP - ISO 640
200% crop from the RAW all controls centered or zeroed in ACR convert to JPeg - No PP - ISO 400
200% crop from the RAW all controls centered or zeroed in ACR convert to JPeg - No PP - ISO 640
Sorry for the OOF... I used an autofocus lens and AF sucks nuggets IMO - especially on µ4/3 cameras - so it missed (on both actually). You can see the noise amount and size just the same tho.
Bifurcator, you can't make accurate noise comparisons between cameras based only on ISO. The exposure between the two cameras must be the same in the same light, to start.
Bifurcator wrote:
What? That's ISO 100 for the Nex7? Wow... I was thinking it was around 800 or so given its reputation - or at least the reputation of the 5n which I know more about. Dang.
cputeq makes a good point. So I guess if it were me as I suggested, I'd wanna make some more tests before deciding anything.
Actually, the NEX-7 doesn't have a noise-free ISO 100, especially in the shadow area (though no banding as the 5D Mark 2 and 3). And we are looking at 100%-200% so I don't think it would be a huge problem. Nonetheless, I would love to meet with another NEX-7 user to make a better comparison. I'm going to be in Houston from Wednesday to Sunday next week. So if any of you in the area would like to do the comparison, I would meet.
Yeah, not a huge problem for normal photography - I agree. And for sure NP for prints. But it would still bug the hell out of me!
For meeting people I find spending some time in a camera shop works well - at least for me. Or maybe ask to put up a bulletin board there? Adding a community element to a shop is usually really good for business.
Well, today manufacturers simply hope that every RAW developer has NR switched on and customers wont go under these values.
If you start using something DCraw based, you are going to get nasty suprise and sudden realisation than most today sensors are pretty much cr*p. There are cams that are truly clean at base ISO, but there isnt exactly plenty of them (mostly FF cams).
But then, what can be expected in this age.
Yea and those GH2 samples - at ISO 800 it has 26.1dB SNR, so it really cant look like this without a lot of NR. Most likely its done directly on RAW in camera, or another hidden auto-correction feature of ACR, which is quite far from "RAW" even if you zero everything.
And AA-less NEX-7 .. well fence is bit more visible. But simply try to shoot something normal (not test) and lets see? Tho you wont see much with regular lens, something Leica/Zeiss/Voigtlaender grade is needed.
For me, I'm fine with the result that I got so far with normal shooting. I think I even imagine that the image from the EVF is a bit crisper . So this is just to share (admittedly not conclusive). This was shot with the Zeiss 24/1.8.
Thank you for all the suggestions. I will post this on Flickr NEX group and see what would happen. I will also check the local brick and mortar store to see if they have any NEX-7 for testing purpose.
Even if it does exhibit more noise, so what. At least then you can control the noise reduction vs detail. The fence has substantially more detail. A better quality lens might possibly render even better detail. The NEX7 loves lots of light, there is a significant difference between the two images in the base lighting. In the pre AA removal shot, good luck recovering fence detail on that image.
The GH2 has a very noisy blue channel, substantially more noise in the b-channel vs the NEX. And yes I have shot both.
Andrew Gough wrote:
Even if it does exhibit more noise, so what. At least then you can control the noise reduction vs detail. The fence has substantially more detail. A better quality lens might possibly render even better detail. The NEX7 loves lots of light, there is a significant difference between the two images in the base lighting. In the pre AA removal shot, good luck recovering fence detail on that image.
The GH2 has a very noisy blue channel, substantially more noise in the b-channel vs the NEX. And yes I have shot both.
Mescalamba, there were no ISO 800 samples posted. Reread the post maybe?
Also this is about the same result I get with all RAW converters. This is just what (my copy of) the GH2 looks like at 400 and 640 respectively. Plastisonic definitely holds an eye towards unconventional "improvements" (read: cheating the process) so it wouldn't surprise me if there were some capture-time NR routines implemented but since no one can prove it and they certainly ain't copping to it, then I guess it has to remain far-side speculative and undiscussable.
douglasf13 wrote:
Bifurcator, you can't make accurate noise comparisons between cameras based only on ISO. The exposure between the two cameras must be the same in the same light, to start.
Sure I can. In fact, I just did. And it IS relevant too. Sure it's not scientifically perfect or anything but it's WAY good enough to communicate a general truth. Especially here where the lighting in my shot should have produced more noise than in the Nex7 outdoor shots - yet didn't.
Was confused by that statement at start. Sorry. Tho it still remains true, it should be way noisier then is. But truth is that whatever Panasonic is doing with NR, its working nicely at that ISO 400 and 640, no noise and it doesnt seem like detail is lost.
.
Wow... I was thinking it was around 800 or so given its reputation - or at least the reputation of the 5n which I know more about. Dang.
