[manishved]

I really believe that the lens that I rented was one of the issues.

Very much possible. My 16-50 is very sharp and the colors are awesome.

[Sandy Hancock]

My 16-50 is very sharp and the colors are awesome.

+1
My most used event lens. SDM issues (which I have now bypassed) aside, it is a wonderful lump of glass.

[dtmateojr]

OK....

K-3


D700


D700 pixel level


K-3 same size..



I have no idea where this great difference is you're seeing.

And you're giving up about 1000 lw/ph, 2700, 1700. Thats a lot of resolution. About the same difference as a K-3 to D800. I wouldn't do it. I can barely see the difference at 3200 ISO, and giving up 1000 lw/ph, that's giving up too much.

That's well and good. The only issue is that high ISO noise becomes more visible in areas of less light. Have a look at this comparison from dpreview. In low light, the D700 shines

[DSims]

I know of a few photographers that actually prefer using their lighter and smaller APSC camera with a small DX zoom for weddings or long photoshoots. But it doesn't scream I'M A PHOTOGRAPHER as much as a D3 with a 70-200 on it.

All you need is a battery grip and a DA*50-135 with the lens hood on! People will believe you.

---------- Post added 03-04-14 at 01:39 AM ----------

Sometimes I wonder if newer sensors are really getting "better".

Me too.

---------- Post added 03-04-14 at 01:42 AM ----------

On those two things:

- Is it paying enough?

- Why spend the money on an old piece off electronics (instead off a new D610/D800/Df)?

Good points; why should you spend the money on a nearly obsolete body, and will the work pay you enough to justify buying anything?

---------- Post added 03-04-14 at 01:45 AM ----------

I have no idea where this great difference is you're seeing.

And you're giving up about 1000 lw/ph, 2700, 1700. Thats a lot of resolution. About the same difference as a K-3 to D800. I wouldn't do it. I can barely see the difference at 3200 ISO, and giving up 1000 lw/ph, that's giving up too much.

Thanks for the examples; I saw it right away. I thought that was why you were posting them until I saw your comment at the bottom!

[bdery]

One last question to keep my k30 as a backup, is the k3 worth the extra money or do you find that the k5ii is pretty compatible with the k3?

Just to answer this. Regarding noise, on print they will compare well, on a pixel by pixel level the K-3 is a bit noisier (but there are more pixels . Regarding AF, the K-3 is faster. Regarding general operation, the K-3 is faster. Regarding features, the K-3 has more. Regarding price, the K-5 ii is hard to beat.

[Pepe Le Pew]

Regarding price, the K-5 ii is hard to beat.

Or the K-30

[raider]

Hmmm..if the comparison is between k3 and another apsc camera i would suggest you look at k3 seriously because it is really a good camera but if u are asking about the d3 obviously you are fascinated by a ff camera and its capabilities....in that case i would vote for u to go with the d3...life is too short. Get the d3 and be happy :-)...my 2 cents

[jsherman999]

That's well and good. The only issue is that high ISO noise becomes more visible in areas of less light. Have a look at this comparison from dpreview. In low light, the D700 shines

Good comparison, but keep in mind that's showing different magnifications - you need to upsample one or downsample the other to see how they compare at the same display sizes. (DxOMark's low-light ISO score normalizes this way.)

[MJSfoto1956]

all these online comparisons are so damn subjective. So here's another one:



and a link to download the high-res versions:All sizes | What Noise??? | Flickr - Photo Sharing!

My take: if you use the camera properly with good lenses and good technique, the K-3 is quite competitive with full frame. Slap a cheap kit zoom lens on it and compare it to an FX camera with pro lenses and you'll come to a different conclusion.

Michael

P.S. the "kit" 50mm f/1.8 is a pretty damn good lens!

[dtmateojr]

Good comparison, but keep in mind that's showing different magnifications - you need to upsample one or downsample the other to see how they compare at the same display sizes. (DxOMark's low-light ISO score normalizes this way.)

Why resample? They are all showing 100% magnifications. Noise per unit area is as is for each sensor.

[Rondec]

This is iso 4000 on the K3 with no noise reduction.




I find it very usable at iso 6400 without much work and I have seen plenty of demonstrations of higher iso shots than that with good results.

[dtmateojr]

Even a m43 sensor can do that as long as there is enough light. SNR is very important.

[jsherman999]

Why resample? They are all showing 100% magnifications. Noise per unit area is as is for each sensor.

For the same reason DxOmark re samples when they make their comparison scores - if you don't do that, it doesn't represent perceived image nose at the same display sizes. Basically, if you were to make a print each from the d800 and d700, same size, the D800 wouldn't look any noisier.

Same idea with k3 vs k5 noise - viewing each at 100% doesn't give you an accurate picture of image noise.

[dtmateojr]

For the same reason DxOmark re samples when they make their comparison scores - if you don't do that, it doesn't represent perceived image nose at the same display sizes. Basically, if you were to make a print each from the d800 and d700, same size, the D800 wouldn't look any noisier.

Same idea with k3 vs k5 noise - viewing each at 100% doesn't give you an accurate picture of image noise.

Resampling isn't really fair. It has always been a trade off of noise vs resolution. You can't have both. You are supposed to view each image in their intended size. If I am to upsize 12mp to 36mp then it follows that I will need a larger sensor for the 12mp to maintain the same sensor sampling rate. Their methodology is wrong.

A fair comparison is a K5 vs D800. Same sampling rate (resolution).

---------- Post added 03-05-14 at 08:30 PM ----------

Just to follow up on this because it looks like you are mistakenly made to believe that higher res has an advantage over lower res in terms of resolution and noise. You lose data when you downsample (definitely) and you decrease noise (highly questionable)?!

[jsherman999]

Resampling isn't really fair.

It actually is quite a bit more fair than showing a higher-resolution sensor at it's greater native magnification, and then trying to extrapolate what overall image noise would look like from that.

The only time it wouldn't be fair is if you bought the higher-MP camera to shoot larger - then it would be accurate to look closer to 100% to get an indication of what to expect.

If you don't print or display at larger sizes than before, you need to re-sample to get an accurate view.

It has always been a trade off of noise vs resolution. You can't have both. You are supposed to view each image in their intended size.

Yes, intended size - like 8x10, 11x16, or some screen resolution like 2048p-width, whatever size you choose - When you do that, you're re-sampling one, or the other (or both.) This is the comparison you want to make, not a 100% pixel view. And this is why DxoMark does that. You can read more about it here: Contrary to conventional wisdom, higher resolution actually compensates for noise.

If I am to upsize 12mp to 36mp then it follows that I will need a larger sensor for the 12mp to maintain the same sensor sampling rate. Their methodology is wrong.

No, it's not wrong, and they show SNR measurements for both methods (area sample = 'screen', normalized sample = 'print',) and their score is based on the 'print', as that's normalized to an equal display size and is exactly indicative of what photographers do. This isn't anything controversial, and their methodology and reasons are widely accepted - please read that link.

I have personal experience with this - I had a D700 and bought a D800, intending to keep both. I was going to use the D700 as my 'low light' shooter because I expected it to have better noise performance, and that my D800 would be my 'outdoor' body. I kept noticing that at the same display sizes, the D800 really didn't show any more noise, and depending on the sampling, retained more detail still. What's more, the noise it had was finer-grained, just looked more natural and better to me. I still kept the D700 for a year because it's a fantastic camera and I liked having two bodies with me at times.

Here's a typical example:


.

I have some of my own examples as well.




.