DXO Grades the K5!

[night86mare]

:bsmilie:

Aya the K-r is good enough no point upgrading... which in a way i dun think its an upgrade when you dun need all the extra features of the K-5 and K-r is already serving ur needs...

seriously the sensor on the K-x/K-r is already very good... DR is more than enough... keep the money for something else...

yar, better save up for flat...

wait no more engagement like this guy's brother here :bsmilie::bsmilie:

 

[night86mare]

you know, someone just asked me why i was comparing cameras with TWO YEARS' TECHNOLOGY DIFFERENCE.

so i felt bad and went to compare the k-5 to d3s + 1d mark IV on dxomark... :bsmilie:

 

[pinholecam]

anyways, as a side note, i would still take all these figures with skeptism.

we had previously (as a group) slammed DXO's ratings of the older pentax cameras... i'm not going to sing a different song because it favours the newer bunch of cameras.

especially when the site lists k-r as 12 fps.... can't say they pay much attention to accuracy.

Exactly.
12fps certainly caught my eye. :bsmilie: Once it reaches 25fps, we can say goodbye to photography (all take video).

Was never a believer of their results which does not fully relate to the photos being taken (eg. detail retention vs noise). Which is why I but in a disclaimer right on the 1st post.


But many ppl certainly do that them as gospel.
(I just wonder if they change their tune now that the tables are turned :think

 

[tsammyc]

I think the full frame dxomark comparison simply confirms observations on a thread in pentaxforums that the K5 exceeds 2-yr old FF cameras. Technology most often helps newbies. I recall not being a very good photographer on film SLR cameras in terms of manual focus and exposure (but good on composition). With the K5, I can hardly do anything wrong. The camera is so smart and focusses so precisely.

 

[night86mare]

I think the full frame dxomark comparison simply confirms observations on a thread in pentaxforums that the K5 exceeds 2-yr old FF cameras. Technology most often helps newbies. I recall not being a very good photographer on film SLR cameras in terms of manual focus and exposure (but good on composition). With the K5, I can hardly do anything wrong. The camera is so smart and focusses so precisely.

how about one year old full frame, flagship cameras? :bsmilie:

 

[alantkh]

I think the full frame dxomark comparison simply confirms observations on a thread in pentaxforums that the K5 exceeds 2-yr old FF cameras.

I think that thread was on high ISO performance right? Most of the hype was on high ISO performance and DXOMARK actually confirm that 2-yr old FF is still better in terms of high ISO performance.

DXOMARK shows that k-5 wins in terms of its ISO80 performance which I feel is more impt than high ISO.

 

[tsammyc]

I think that thread was on high ISO performance right? Most of the hype was on high ISO performance and DXOMARK actually confirm that 2-yr old FF is still better in terms of high ISO performance.

Two things. First, I was saying 2 year old FF cameras, which means the D3, not the D3s. While its true that the "Sports Score" is higher on the FF cameras, we need to dig a little deeper because its not clear how DXOMARK's composite "Sports Score" is put together. For example, what is the definition of high ISO since the K5 can go to higher ISOs than the 2-yr old FFs. All is revealed when you look at the SNR chart, which I reproduce below:

If the "Sports Score" considers ISO 800 and 1600 "high ISO", then its clear the K5 is weaker. However, above 1600 things look different. The K5 lags the FFs slightly until ISO 3200, where it jumps up to match them. In fact, its interesting that the K5's ISO 3200 is almost as good as its ISO 1600. At ISO 12800, it appears to pass the Canon. I don't think the Pentaxforums thread was talking about ISO 800 and 1600.

I shoot a lot more indoor than outdoor, so the high ISO is slightly more important than ISO 80 to me.

 

[tsammyc]

Here is the DXO Mark Review for the Pentax

http://dxomark.com/index.php/en/Our-publications/DxOMark-reviews/DxOMark-review-for-the-Pentax-K5

An exerpt:

No need for suspense: this new 16.3 MP sensor is simply the best APS-C we have tested so far, sometimes able to compete even with very high-end full-frame cameras.

The overall score of the K5 puts it in the lead with 82 points — more than 9 points better than the D90 or the Alpha 55, and 16 points ahead of the Canon 7D or 60D. The K5 is literally the best APS-C performer for each segment, even in low ISO.

 

[night86mare]

the online photographer on the k-5 and dxomark ratings

The sensor in the new Pentax K-5 has taken over the #1 ranking for APS-C sized sensors on DxOmark.com, and not by a small margin. DxOmark is unambiguous in its praise: under the heading "The best APS-C in all tested fields," they write, "No need for suspense: this new 16.3 MP sensor is simply the best APS-C we have tested so far, sometimes able to compete even with very high-end full-frame cameras." It comes in sixth overall—nestled in between the Nikon D3s and D3 and surrounded by nothing but full-frame and medium format cameras. You have to go all the way down to #19 on the list to find the next APS-C sensor camera.

Calling its dynamic range performance "wonderful," DxOmark says this area is "clearly where the K5 struts its stuff...even the [Nikon] D3X’s full-frame sensor is not [as] good."

Where high ISO is concerned, being a "small" sensor, the K-5 is still "roughly a stop behind" the $780-more-expensive Nikon D700 and $900-more-expensive Canon 5D Mark II*. However, the K-5's overall sensor score of 82 trounces the APS-C competition, "more than 9 points better than the D90 or the Alpha 55, and 16 points ahead of the Canon 7D or 60D," and is top-ranked for high ISO among its same-size peers. (Note that the scores for the new Nikon D7000 are not posted yet, so Nikon still has a shot to retain the high-ISO crown.)

 

[alantkh]

However, above 1600 things look different. The K5 lags the FFs slightly until ISO 3200, where it jumps up to match them. In fact, its interesting that the K5's ISO 3200 is almost as good as its ISO 1600. At ISO 12800, it appears to pass the Canon. I don't think the Pentaxforums thread was talking about ISO 800 and 1600.

I shoot a lot more indoor than outdoor, so the high ISO is slightly more important than ISO 80 to me.

DXOMARK already confirm that pentax cameras do NR at 3200 and above so the SNR results are not accurate, ie artificially inflated. ie you have to look at ISO1600 and below for the actual performance.

There are a LOT of aritcles on this so I am surprised that people still don't know about it. Even DXOMARK explicitly mention the NR by using the term smoothing in the SNR curves.

That is a minus is what pentax does, which is forced NR on its raw files for ISO3200 and above. That is why I am skeptical of all the super clean high ISO pics people are posting claiming it is without NR when obiously you can see NR artifacts at 100% crop.

Still the k-5 sensor is REALLY REVOLUTIONARY. can't wait for a FF version of that sensor.

 

[creampuff]

alantkh, so what difference does it make if there is NR above ISO 3200? So what?
You seem to think it is an issue but can you actually tell us what practical effect does it have on picture taking. To argue for argument's sake is one thing but it is kinda meaningless to talk about "SNR results are not accurate, ie artificially inflated" because you haven't established what this so called negative effect of NR being applied is all about.

Bottomline for a less than tech savvy person like me is that the Pentax engineers have tweaked the image processing to deliver a stunning image quality across a broad spectrum of ISO range.

 

[poppyer]

Completely no idea what you are talking about.
I shoot at JPG at ISO12800, get bad results. Then I shoot at RAW and get acceptable images. That's it. What am I supposed to do more? use a screwdriver to screw up the sensor inside the camera directly? or a screwdriver with canon's non-NR sensor help here?

DXOMARK already confirm that pentax cameras do NR at 3200 and above so the SNR results are not accurate, ie artificially inflated. ie you have to look at ISO1600 and below for the actual performance.

There are a LOT of aritcles on this so I am surprised that people still don't know about it. Even DXOMARK explicitly mention the NR by using the term smoothing in the SNR curves.

That is a minus is what pentax does, which is forced NR on its raw files for ISO3200 and above. That is why I am skeptical of all the super clean high ISO pics people are posting claiming it is without NR when obiously you can see NR artifacts at 100% crop.

Still the k-5 sensor is REALLY REVOLUTIONARY. can't wait for a FF version of that sensor.

 

[pinholecam]

DXOMARK already confirm that pentax cameras do NR at 3200 and above so the SNR results are not accurate, ie artificially inflated. ie you have to look at ISO1600 and below for the actual performance.

There are a LOT of aritcles on this so I am surprised that people still don't know about it. Even DXOMARK explicitly mention the NR by using the term smoothing in the SNR curves.

That is a minus is what pentax does, which is forced NR on its raw files for ISO3200 and above. That is why I am skeptical of all the super clean high ISO pics people are posting claiming it is without NR when obiously you can see NR artifacts at 100% crop.

Still the k-5 sensor is REALLY REVOLUTIONARY. can't wait for a FF version of that sensor.

I don't see the point of repeating this over and over again on various threads. I think everyone knows some sort of noise reduction/suppression is done on high ISO. What matters most is that the o/p is very good. Those who have the camera know what they are seeing from the camera o/p and like it. Many have even posted their works and bothered to put up 100% crops. Do they care what is being done on camera? Can they change the way it is done? No

What you are purporting is like saying a Bayer sensor is not 'true' enough, and cameras should be shooting with a sensor that has photosites that register all 3 colors with equal capability. What is true?
Doesn't matter. So long as the o/p is desirable (ie. detail retention; fine grain noise; ease to cleanup with NR program).
Or should sensors be optical based rather than electronic based?
Maybe what we really need is a 'true' biological eye type sensor/camera?
Lets leave the science to the scientists and science fiction writers :bsmilie:


In two threads, I've asked you to post photos from what you think an 'ideal' and 'true' camera would give, but up till now, I only see repeated statements of ("obvious NR at high ISO).

 

[poseur]

Bottomline for a less than tech savvy person like me is that the Pentax engineers have tweaked the image processing to deliver a stunning image quality across a broad spectrum of ISO range.

I agree with creampuff, It is not a big deal if NR kicks in at ISO 3200 6400 or even 12800....it only show Pentax engineer created value by improving the NR algorithm without incuring extra hardware cost. The saving will be pass to consumers like us.

with this i leave behind a classic engineering story.
QC was badly needed in a soap factory when soap was misaligned to the plastic soap box, Many empty boxes packed and shipped out causing customer unhappiness.
Many engineers came up with solution requiring heavy investment in equipment like lazer alignment machines etc etc. one young engineer simply walk in took a high power industrial fan place it at the production line and blew away all empty box....

 

[night86mare]

alantkh, so what difference does it make if there is NR above ISO 3200? So what?
You seem to think it is an issue but can you actually tell us what practical effect does it have on picture taking. To argue for argument's sake is one thing but it is kinda meaningless to talk about "SNR results are not accurate, ie artificially inflated" because you haven't established what this so called negative effect of NR being applied is all about.

Bottomline for a less than tech savvy person like me is that the Pentax engineers have tweaked the image processing to deliver a stunning image quality across a broad spectrum of ISO range.

dunno, at the end of the day, if the pictures work, doesn't matter right.

if any camera can whack the picture hard with noise processing and still give me a clean picture with loads of detail, i couldn't give a rat's ass as to whether noise processing was added or not, or if a monkey had been running around inside the sensor painting a picture instead of the camera doing it. what i care about is whether the results are usable, instead of sitting around to quibble about what was done, what was not, or whether a monkey had shat out what i see in front of me, instead of the camera. :bsmilie:

 

[creampuff]

Ah I think I got it checked out... our dear CS member is a Nikon D700 user mah!
Thanks for stopping by and finding out more about the Pentax K-5. :lovegrin:

 

[tsammyc]

Haha. No wonder.

 

[alantkh]

alantkh, so what difference does it make if there is NR above ISO 3200? So what?
You seem to think it is an issue but can you actually tell us what practical effect does it have on picture taking. To argue for argument's sake is one thing but it is kinda meaningless to talk about "SNR results are not accurate, ie artificially inflated" because you haven't established what this so called negative effect of NR being applied is all about.

Bottomline for a less than tech savvy person like me is that the Pentax engineers have tweaked the image processing to deliver a stunning image quality across a broad spectrum of ISO range.

?? There is nothing negative about it PROVIDED you KNOW it is being done. If someone buys a k-5 thinking it will give high ISO performance as good as D700, he will be disappointed. Simple as that.

I was just trying to educate someone who did not know about pentax forced NR. TsammyC DID think that the k-5 performs as well as the D700 at high ISOs which is WRONG.

Yes I am a D700 user (and previously owned and used s3pro, KM7D, olympus E1 E330, samsung NX10) and I am looking for a 2nd body to mount a UWA lens for my travels. I am used to D700 high ISO performance so I was very happy to see a lighter APS-C sensor looking so good.

you guys need to be less defensive about brands. I keep saying good things about pentax k-5 but all you guys jump on are the bad stuff. I am thinking of using the k-5 side by side my D700 cuz I think it is really that good and lighter.

 

[felixcat8888]

?? There is nothing negative about it PROVIDED you KNOW it is being done. If someone buys a k-5 thinking it will give high ISO performance as good as D700, he will be disappointed. Simple as that.

I was just trying to educate someone who did not know about pentax forced NR. TsammyC DID think that the k-5 performs as well as the D700 at high ISOs which is WRONG.

Yes I am a D700 user (and previously owned and used s3pro, KM7D, olympus E1 E330, samsung NX10) and I am looking for a 2nd body to mount a UWA lens for my travels. I am used to D700 high ISO performance so I was very happy to see a lighter APS-C sensor looking so good.

you guys need to be less defensive about brands. I keep saying good things about pentax k-5 but all you guys jump on are the bad stuff. I am thinking of using the k-5 side by side my D700 cuz I think it is really that good and lighter.

Pls dun think of us like that. We appreciate a Nikon user giving comments too. We welcome anyone to get a Pentax and even those without one can join our monthly outings.

For most of us who are not that technical, whatever comes out of the camera is very good enough for us. As we have seen with the various photos posted, the K5 owners are very happy with the output.

I too am happy with my K7 as I only switched over from the K20D in March this year.

Pls do join us in our outings so you too can see the K5 first hand.

 

[poppyer]

Now I am really really curious how good high ISO performance TsammyC can get if he switch to a D700.

?? There is nothing negative about it PROVIDED you KNOW it is being done. If someone buys a k-5 thinking it will give high ISO performance as good as D700, he will be disappointed. Simple as that.

I was just trying to educate someone who did not know about pentax forced NR. TsammyC DID think that the k-5 performs as well as the D700 at high ISOs which is WRONG.

Yes I am a D700 user (and previously owned and used s3pro, KM7D, olympus E1 E330, samsung NX10) and I am looking for a 2nd body to mount a UWA lens for my travels. I am used to D700 high ISO performance so I was very happy to see a lighter APS-C sensor looking so good.

you guys need to be less defensive about brands. I keep saying good things about pentax k-5 but all you guys jump on are the bad stuff. I am thinking of using the k-5 side by side my D700 cuz I think it is really that good and lighter.