Pentax 645D vs D800 Image Quality

[wy2lam]

If downsampling lowers the noise for you, that means you must not have printed anything. All you're doing is viewing images at 100% on your TN LCD panel.

i don't need chart to see the obvious.

I do .. I won't pass any judgment until there are studio shots in controlled setting with same subjects and lighting.

But MF is simply another planet for ultimate IQ at base iso. a

That has yet to be seen.

they should have kept the number lower and they would have had a better IQ.

No thanks. 36 MP are just fine. You can always downssample it later and lower the noise.

 

[awaldram]

No thanks. 36 MP are just fine. You can always downssample it later and lower the noise.

Garbage in garbage out, If you've already destroyed the detail with noise no amount of down-sampling will restore it.

And 36 MP captures the detail that 16 MP don't, so you are staring out from more detail. And you are just babbling nonsense. It's a well known fact that downsampling reduces noise.

Your really showing your complete lack of photographic experience.

Your statement (besides being deliberately offensive) is ridiculous and shows you have no credentials for the rubbish you talk.

Lesson 101 in noise for ET2

Noise caused by hiISO (sensor and op-amps) will raise the Noise floor above the signal level, Once this happens any signal below the noise floor will be gone.

Once the signal is gone no amount of post processing to reduce the noise floor will restore that lost signal.

So Down sampling will reduce noise but it WILL NOT restore the lost detail exactly as I stated.

My original statement made an assumptions on the photographic knowledge of the reader that was obviously well above your current level.

You should remember this is a Technical Photographic forum, I expected you to have a rudimentary understanding of the technology.!

GiGO = It is used primarily to call attention to the fact that computers will unquestioningly process the most nonsensical of input data ("garbage in") and produce nonsensical output ("garbage out").

I was referring to the fact once you've destroyed the data the computer cannot restore it whatever you do. i.e Garbage IN

--
My PPG
http://www.pentaxphotogallery.com/artists/andrewwaldram
My Photo Stream
http://www.flickr.com/photos/awaldram/
1x.com
http://1x.com/artist/awaldram/wall

 

[wy2lam]

Never mind about ET2, who'll be happy with his noiseless picture downsampled all the way to a single pixel. Smoothest, cleanest image ever!!

 

[awaldram]

Never mind about ET2, who'll be happy with his noiseless picture down-sampled all the way to a single pixel. Smoothest, cleanest image ever!!

Exactly my point maybe taking to extremes but accurate all the same.

re ET2 I always like to think there's time for people to grow nobody is a lost cause.

People on this forum have surprised me folks who've I've pegged as idiot come back with knowledge and ability that humbles me.

Mind you whilst ET2 spends all his time pontificating and none listening he has very littler chance of raising himself above bottom feeder a position he has placed himself in .

I do wonder though after all the comments he's had across every forum re his attitude, Sony bias and DXO love affair. He'd maybe think to look at his approach and amend it to a more socially acceptable format.

Maybe he enjoys the ridicule he courts ?

--
My PPG
http://www.pentaxphotogallery.com/artists/andrewwaldram
My Photo Stream
http://www.flickr.com/photos/awaldram/
1x.com
http://1x.com/artist/awaldram/wall

 

[Ian Stuart Forsyth]

This is not true. Even if D800 sensor tech is identical to D7000 (and let's assume it is), then D800, given it's 36 MP, will score much higher than D7000. That's because Dxomark scores are calculated by normalizing all cameras to 8 MP. Obviously, the gain from down-sampling 16 MP to 8 MP would be lower than down-sampling 36 MP to 8 MP. So D800, even if it performs identically to D7000 at pixel level, will score much higher than D7000 when both are normilzed to 8 MP. That's how Dxomark calculates scores; they normalize everything to 8 MP.

In fact, I won't be surprised if D800 becomes the highest rated camera ever on dxomark.

I can see the D800 closing the gap on the out-dated 645d and maybe under certain conditions beating the 645 in IQ. The $6000 dollar question is that gap with it.
--
The Camera is only a tool, photography is deciding how to use it.



http://ianstuartforsythphotography.com
http://www.flickr.com/photos/isfphotography/
http://www.pentaxphotogallery.com/artists/ianforsyth#/carousel

 

[awaldram]

Hey Ian

I think you missed some words out

"The $6000 dollar question is that gap with it. "

--
My PPG
http://www.pentaxphotogallery.com/artists/andrewwaldram
My Photo Stream
http://www.flickr.com/photos/awaldram/
1x.com
http://1x.com/artist/awaldram/wall

 

[ET2]

Your statement (besides being deliberately offensive) is ridiculous and shows you have no credentials for the rubbish you talk.

Funny that is coming from someone who just yesterday listed as a "plus" for K-5 that his K-5 has ISO 512000 but D800 doesn't. I don't believe anyone ever posted something this stupid on this forum in the last 6 months.

Noise caused by hiISO (sensor and op-amps) will raise the Noise floor above the signal level, Once this happens any signal below the noise floor will be gone.

D800 and K-5 have identical pixel size, so on per pixel level, D800 can never be worse than K-5 for noise. However, since D800 files are 36 MP, downsizing D800 to 16 MP will mean that D800 will have significantly lower noise and better DR than K-5 (at equal size). Anyone who denies this is just outright stupid

 

[ET2]

Never mind about ET2, who'll be happy with his noiseless picture downsampled all the way to a single pixel. Smoothest, cleanest image ever!!

Posting one-liner like that doesn't refute what I posted. Dxomark has covered this topic extensively,

"Contrary to conventional wisdom, higher resolution actually compensates for noise"

http://www.dxomark.com/index.php/Publications/DxOMark-Insights/More-pixels-offset-noise%21

And read this article by Emil Martinec (Professor of physics at the university of Chicago)

http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/noise-p3.html

quote:

"Bottom line: Among the important measures of image quality are signal-to-noise ratio of the capture process, and resolution. It was shown that for fixed sensor format, the light collection efficiency per unit area is essentially independent of pixel size, over a huge range of pixel sizes from 2 microns to over 8 microns, and is therefore independent of the number of megapixels. Noise performance per unit area was seen to be only weakly dependent on pixel size. The S/N ratio per unit area is much the same over a wide range of pixel sizes. There is an advantage to big pixels in low light (high ISO) applications, where read noise is an important detractor from image quality, and big pixels currently have lower read noise than aggregations of small pixels of equal area. For low ISO applications, the situation is reversed in current implementations -- if anything, smaller pixels perform somewhat better in terms of S/N ratio (while offering more resolution). A further exploration of these issues can be found on the supplemental page. Rather than having strong dependence on the pixel size, the noise performance instead depends quite strongly on sensor size -- bigger sensors yield higher quality images, by capturing more signal (photons)."

Yes, he said smaller pixels provide better S/N ratio for lower ISOs than larger pixels ... so Nikon was correct in sticking with the 36 MP sensor. I will take the 36 MP over the 16 MP in D4. D4 (as a sport camera) needed a very high burst rate. That's why it has 16 MP sensor. Otherwise, just for pure IQ, especially at lower ISOs, the D800 sensor is probably superior than D4 sesnor (and I am sure dxomark will confirm that).

 

[hikenhi]

I've thought it over several times. If I go to something like a D-800E, with the comparable 3 lenses, my Pentax; DA 12-24 f4, DA* 16-50 f2.8, & DA* 50-135 f2.8, in the FF Nikon FX format.... What will I loose aside from going from excellent IQ to perhaps superior IQ? I'll tell you; A $4,000 Pentax set-up to a $10,000 Nikon setup that's twice the equipment volume, twice the weight, and will require me to upgrade my Gitzo Carbon Fiber Tripod from a 0 series to an expensive $1,000 2-series set-up that will add another 2 pounds. It's like carrying 2 of everything I've already got at 2 1/2 times the price, All for just a marginal IQ improvement and a useless 2-fold resolution for non-Ansel Adams-like work that I do. It's just not practical for what I do and what most people do(.). I'll stick with my K-5, "Way more bang for my buck(.)"

 

[awaldram]

Your statement (besides being deliberately offensive) is ridiculous and shows you have no credentials for the rubbish you talk.

Funny that is coming from someone who just yesterday listed as a "plus" for K-5 that his K-5 has ISO 512000 but D800 doesn't. I don't believe anyone ever posted something this stupid on this forum in the last 6 months.

I think you've managed many stupider comments see below being an example .

The statement I made was 'Fact' something like truth which escapes you totally.

It is a potential plus point as the top ISO has a bearing on the usable ISO (normally 2 -2.5 stops below quoted max)

Its not a xx camera is better the YY camera as you seem to think the world revolves around.

Noise caused by hiISO (sensor and op-amps) will raise the Noise floor above the signal level, Once this happens any signal below the noise floor will be gone.

D800 and K-5 have identical pixel size,

4.75 Vs 4.88 the d800 has an advantage, But the larger sensor means more thermal control required.

so on per pixel level, D800 can never be worse than K-5 for noise.

Probably but please quote your source for this "fact"

However, since D800 files are 36 MP, downsizing D800 to 16 MP will mean that D800 will have significantly lower noise and better DR than K-5 (at equal size). Anyone who denies this is just outright stupid

I deny 100% downsizing has never and never will increase DR , to the contrary anyone who says otherwise (see above) has made themselves look a complete idiot.

--
My PPG
http://www.pentaxphotogallery.com/artists/andrewwaldram
My Photo Stream
http://www.flickr.com/photos/awaldram/
1x.com
http://1x.com/artist/awaldram/wall

 

[zakaria]

I've thought it over several times. If I go to something like a D-800E, with the comparable 3 lenses, my Pentax; DA 12-24 f4, DA* 16-50 f2.8, & DA* 50-135 f2.8, in the FF Nikon FX format.... What will I loose aside from going from excellent IQ to perhaps superior IQ? I'll tell you; A $4,000 Pentax set-up to a $10,000 Nikon setup that's twice the equipment volume, twice the weight, and will require me to upgrade my Gitzo Carbon Fiber Tripod from a 0 series to an expensive $1,000 2-series set-up that will add another 2 pounds. It's like carrying 2 of everything I've already got at 2 1/2 times the price, All for just a marginal IQ improvement and a useless 2-fold resolution for non-Ansel Adams-like work that I do. It's just not practical for what I do and what most people do(.). I'll stick with my K-5, "Way more bang for my buck(.)"

one of the most interesting comment here .you got and counted it in easy way.
yes every thing here about d800 is double k5.
thanks

 

[Neil Morgan]

I have a D700 and had a K20D and can heartily say this is too true.

There was no K5 at the time I changed over but I sorely missed the size and weight advantage of the K20D not to mention the menu system which is still more functionally correct than the Nikon menus.

Bang for buck Pentax is extremely good, they have their faults too, but IQ is not one of them.

If you want to see the difference in sizes between the two systems then they are on my flickr pages in the gear section.

To me, the D800 offers an upgrade from where i am now, but the files from the D700 are still big enough for my needs and 36mp at the moment wont really help much. If it was 24mp i think it would have been a better thing.

As far as noise goes, I think its a lot of ho-ha about nothing. I have seen a lot of banter about comparisons to the canon d1x but its the 5d it should be compared against. DXO may mark it well through downsampling but this is the biggest load of $hite ive seen. If its so great downsampled why bother having 36mp, just give me REAL WORLD results. DR will not be improved on the K5/645D D3s etc, only marginal gains if any and it seems to me that at the moment that is about the limit

of DR today. Certainly even if you whack the biggest best processor in 36mp is still too much to deal with, especially seeing 4fps frame rate is as fast as the processor can handle.

The price is very good though and I'm sure that no one will be really disappointed.

Better than the 645D at base - doubt it. Maybe better at 6400iso and it definitely has better video but landscapers for example only want IQ so I think the 645D may still have an edge.

Cheers Neil
--
A Birth Certificate shows that we were born.
A Death Certificate shows that we died.
Pictures show that we lived!

http://www.flickr.com/photos/knumbnutz/

 

[ET2]

However, since D800 files are 36 MP, downsizing D800 to 16 MP will mean that D800 will have significantly lower noise and better DR than K-5 (at equal size). Anyone who denies this is just outright stupid

I deny 100% downsizing has never and never will increase DR , to the contrary anyone who says otherwise (see above) has made themselves look a complete idiot.

You can check Dxomark on that where 100% pixel size has lower DR than the 8 MP (normalized) print size. That applies to K-5's DR score too on Dxomark (13.6 DR score at 100% pixel, and 14.1 at 8 MP).

themselves look a complete idiot.

lol @ that's coming you ..

 

[bavarius6]

You can check Dxomark on that where 100% pixel size has lower DR than the 8 MP (normalized) print size. That applies to K-5's DR score too on Dxomark (13.6 DR score at 100% pixel, and 14.1 at 8 MP).

ET2, you are always going to disagree with the majority of people here as most here actually practise photography while you are poring over DXo charts and pixel peeping others test images.

If you think downsizing images is ever going to improve the quality of any image rather than just masking it's shortcomings then you obviously have little true comprehension of the subjects you preach on.

And your quote above is an absolute classic! You really believe that by downsizing an image AFTER capture can increase the true DR in that image therefore adding detail? Please! Thankyou to you and DXo for the entertainment.....brilliant! :0

 

[awaldram]

However, since D800 files are 36 MP, downsizing D800 to 16 MP will mean that D800 will have significantly lower noise and better DR than K-5 (at equal size). Anyone who denies this is just outright stupid

I deny 100% downsizing has never and never will increase DR , to the contrary anyone who says otherwise (see above) has made themselves look a complete idiot.

You can check Dxomark on that where 100% pixel size has lower DR than the 8 MP (normalized) print size. That applies to K-5's DR score too on Dxomark (13.6 DR score at 100% pixel, and 14.1 at 8 MP).

That is becasue DXO define DR as

Dynamic range is defined as the ratio between the highest and lowest gray luminance a sensor can capture. However, the lowest gray luminance makes sense only if it is not drowned by noise, thus this lower boundary is defined as the gray luminance for which the SNR is larger than 1. The dynamic range is a ratio of gray luminance; it has no defined unit per se, but it can be expressed in Ev, or f-stops.

So as weve already established the statistical SNR will improve by downsampling then their DR figure will improve.

This does not alter the fast that the Dynamic range (i.e the difference between the lightest and darkest point on an image) cannot alter due to Downsizng.

And must cast some doubt as to the veracity of DXO's figures for non technical people to use..

The downsampled image will have no better detail than the original and no different DR yet DXO figures will show improvement for both.

No action on the image can create data that wasn't originally caught by the sensor to imagine otherwise is fairytale material, exceeding Nyquest limit to generate false postives is the usual reason why figures of 'lab sites' exceed the sensors bounderies.

Technically they are correct the signal that remains is now lifted further above the noise floor and becaues of that the reported DR will be better.

This is why understanding what figures mean is considerably more important than being able to quote them, and why observational data is often at odds to the statisical evidence.

Whatever DXO mark the d800 achieves will not alter the fact that in current examples the 645d is Leagues above it in IQ.

Would I buy a d800 above a 645d you betcha as the 645d would not be able to take images in the enviroment I normally shoot.
That doesn't alter anything though.

themselves look a complete idiot.

lol @ that's coming you ..

--
My PPG
http://www.pentaxphotogallery.com/artists/andrewwaldram
My Photo Stream
http://www.flickr.com/photos/awaldram/
1x.com
http://1x.com/artist/awaldram/wall

 

[ET2]

That is becasue DXO define DR as

So we agree if D800's per pixel performance is same as K-5, then D800's DR score will be better than K-5 on Dxomark . That's because D800 is 36 MP, and down-sampling 36 MP to 8 MP will result in bigger DR (and lowlight) gains on Dxomark

K-5 already scores better DR than 645D on dxomark.

Whatever DXO mark the d800 achieves will not alter the fact that in current examples the 645d is Leagues above it in IQ.

This claim hasn't been proven till we see proper studio shots and reviews.

 

[awaldram]

That is becasue DXO define DR as

So we agree if D800's per pixel performance is same as K-5, then D800's DR score will be better than K-5 on Dxomark . That's because D800 is 36 MP, and down-sampling 36 MP to 8 MP will result in bigger DR (and lowlight) gains on Dxomark

Yes DXOmark scores are an invalid reference to a cameras real world abilities and only provide a guide to the cameras potential.

The scoring is heavily weighted towards a General purpose Jack of all trades camera rather than one that excels in a single genre.
Even their sub classes Portrait ,landscape and sport are misnomers.

Lanscape score takes no account of colour fidelity or tonal range !!

Portraits score is about number of colours rather than how accurate skin tones can be reproduced.

Sports well according to DXO the 645d is a better camera for sport than the d7000!

K-5 already scores better DR than 645D on dxomark.

Probably indicative that the K5 has a better dark level then the 645d rather than real luminance range advantage.

The RAW noise modification utilised by Pentax will give the k5 a boost in both DR and SNR using DXOmark standard parameters.

Given the way the NR was done it will have very little if any affect on detail blurring but will affect DXO mark scores.

The cynical may imagine Pentax sticking two fingers up at DXO at this point

Whatever DXO mark the d800 achieves will not alter the fact that in current examples the 645d is Leagues above it in IQ.

This claim hasn't been proven till we see proper studio shots and reviews.

Yes it has, read what I said not what you imagine, current examples as provide by Nikon (I would hope the best they can deliver).

Even when studio shots and reviews are available the 'current samples' will be the same so the results that the 645d outshine the 'current samples' will remain true.

--
My PPG
http://www.pentaxphotogallery.com/artists/andrewwaldram
My Photo Stream
http://www.flickr.com/photos/awaldram/
1x.com
http://1x.com/artist/awaldram/wall

 

[ET2]

Yes DXOmark scores are an invalid reference to a cameras real world abilities

No, DXOmark scores are a valid reference to camera's sensor RAW potential.

By the way, I find it funny how most of the fanboys were touting k-5's dxomark scores (especially DR score) for last 1.5 year, and claiming that K-5 is better than FF cameras (and they were ignoring the fact that most of these FF cameras were 3 to 5 years old), but now, all of the sudden, Dxomark has become invalid again -- just it was invalid in K-7 era when K-7 scored badly on Dxomark.

This is funny.

This claim hasn't been proven till we see proper studio shots and reviews.

Yes it has, read what I said not what you imagine, current examples as provide by Nikon (I would hope the best they can deliver).

Nonsense. Sample "photos" (jpegs, not even RAW) taken by different cameras of different subject in different lighting in different countries (probably on different continents) by different photographers prove absolutely nothing about the sensor and IQ.

You (and Johnny) can repeat that claim a million times (till you turn blue), but I will wait for studio samples to make a judgement.

As for Dxomark (and absolute RAW performance of the sensor), I have no doubt that D800 will utterly destroy 645D on dxomark.

 

[awaldram]

Yes DXOmark scores are an invalid reference to a cameras real world abilities

No, DXOmark scores are a valid reference to camera's sensor RAW potential.

Rubbish you don't even understand what the scores mean

In your world .....
Phase One IQ180 with its .7FPS is a better Sports sensor than the eos rebel !!
966 Vs 793

and the d7000 is a better landscape sensor than the MF camera !!
13.9 Vs 13.6

By the way, I find it funny how most of the fanboys were touting k-5's dxomark scores (especially DR score) for last 1.5 year, and claiming that K-5 is better than FF cameras (and they were ignoring the fact that most of these FF cameras were 3 to 5 years old), but now, all of the sudden, Dxomark has become invalid again -- just it was invalid in K-7 era when K-7 scored badly on Dxomark.

This is funny.

Your point with reference to this discussion ??

This claim hasn't been proven till we see proper studio shots and reviews.

Yes it has, read what I said not what you imagine, current examples as provide by Nikon (I would hope the best they can deliver).

Nonsense. Sample "photos" (jpegs, not even RAW) taken by different cameras of different subject in different lighting in different countries (probably on different continents) by different photographers prove absolutely nothing about the sensor and IQ.

re Jpegs not even RAW ??? how do you propose to examine the RAW file !!

Let me get this right you want to make a Judgement based on one camera under controlled lighting (not sure about why you mentioned countries and continents are you xenophobic) by one photographer all without any credentials?.

Tell me just who is DXOmarks photographer and how qualified is he to make the images ??

Wow you really are living in cloud cuckoo land.

So you know more than Nikon now !! maybe you should tell them how they've got it all wrong.

http://imaging.nikon.com/lineup/dslr/d800/sample01.htm

Taken by the following pro photographers maybe you should provide them some pointers how to take a picture with this camera ??
Rob Van Petten - Fashion
Jim Brandenburg - Nature
Benjamin Antony Monn - Architectural
Cliff Mautner -Wedding

You (and Johnny) can repeat that claim a million times (till you turn blue), but I will wait for studio samples to make a judgement.

Unfortunately your judgement will only be of interest to one person.

As for Dxomark (and absolute RAW performance of the sensor), I have no doubt that D800 will utterly destroy 645D on dxomark.

This bothers me because .??
Do you own a camera, Have you ever taken a picture?

--
My PPG
http://www.pentaxphotogallery.com/artists/andrewwaldram
My Photo Stream
http://www.flickr.com/photos/awaldram/
1x.com
http://1x.com/artist/awaldram/wall

 

[tcom]

Yes DXOmark scores are an invalid reference to a cameras real world abilities

No, DXOmark scores are a valid reference to camera's sensor RAW potential.

The measures of dxomark do at least give some indication as to the sensor's potential or the ability of the camera manufacturer to bring out most of the sensor. But when I buy a 36MP camera, I want to know how it performs at 36MP, I am not interested to know how it would perform if I reduce it to 8MP.

As for Dxomark (and absolute RAW performance of the sensor), I have no doubt that D800 will utterly destroy 645D on dxomark.

What you do not seem to understand, there is more to image quality than resolution, SNR and DR. Image quality is also defined by color rendition, tonal details, microcontrast and bokeh, but these parameters can not be measured. The dxo gives some key numbers to compare cameras, but the true image quality is not within these numbers, it needs the eyes to look at the images.

The D800 will certainly destroy the 645D above 800iso, I have no doubt about that. But at base ISO, I do not need the numbers from dxo, from what I was able to see from the published D800 samples, I simply can not see it playing in the same image quality league than the 645D.
--
Dominique

International Press Association
http://ipaimpress.com/author/dominique/

http://www.pbase.com/tcom
http://www.flickr.com/photos/dschreckling/