[falconeye]

By the way does anybody have any data on the dynamic range distribution? Is there a distribution curve up somewhere? How much of it is in shadow and highlight areas?

There is no such thing as dynamic range distribution. Dynamic range is a range, i.e., measuring the distance between shadows and highlights (with comparable properties).

What you may mean is the signal distributon curve, i.e., how much signal there is left at various degrees of luminosity.

DxO has full detail if you go out of compare mode. It is called "Full SNR". And it is in "Screen" normalisation so be careful when comparing.

Another good indication is the dynamic range over iso curve. The higher the iso, the more the result describes the shadow properties.

Last note: DPReview differentiates between highlight and shadow headroom. That is nonsense and if at all, only makes sense for a given JPG tone mapping curve.

[Agnostic]

There is no such thing as dynamic range distribution. Dynamic range is a range, i.e., measuring the distance between shadows and highlights (with comparable properties).

What you may mean is the signal distributon curve, i.e., how much signal there is left at various degrees of luminosity.

DxO has full detail if you go out of compare mode. It is called "Full SNR". And it is in "Screen" normalisation so be careful when comparing.

Another good indication is the dynamic range over iso curve. The higher the iso, the more the result describes the shadow properties.

Last note: DPReview differentiates between highlight and shadow headroom. That is nonsense and if at all, only makes sense for a given JPG tone mapping curve.

Thanks for the info. I'm not all that versed in the more intricate technicalities of sensor quality measurement. Yes I mean the signal distribution curve.
From experience I know that the K20D and K7 create very dark shadows and leave little detail and tonal variation in shadow areas, presumably to hide shadow noise though you can recover a bit of that from raw with ACR or DXO optics pro.

I have also been a bit surpised in the past by apparent discrepancies between DPReview and DXO results. Generally in my personal evaluations I tend to rely mostly on real world experience and how the raw files respond in pp as well as looking at lots of samples if I want to evaluate before buying.

It's only when I start running into real world limitations that annoy me that I start looking for technical info to understand what changes in gear and technique could help me to overcome those limitations. Since photography is my personal form of meditation and my exercise in staying in the here and now, I prefer to be out shooting.

I guess that's also part of the reason why I still like analog photography. It slows me down and for my taste negative scans need much less pp to make me feel satisfied with the results. I'm rarely very happy with b&w conversions from the K20D, even the K100D super was better in that regard.

[Rite]

Originally posted by rawr
...which is equivalent to $1600 USD. So stop complaining ...



That's exactly what it's selling for here. And I will not stop complaining.

==========================

Nor will I. Pentax won't see the K-5 bodies flying off the shelves @ $1,600.

[cheekygeek]

Is this "real world" enough for ya?

Angle of View :: Pentax K-5 and an Exposure "Oops"

As the writer says: "Wicked dynamic range".

[Abstract]

Is this "real world" enough for ya?

Angle of View :: Pentax K-5 and an Exposure "Oops"

As the writer says: "Wicked dynamic range".

O_O

.......

Just think of how many shots I could have saved.

[pentaxus]

DxOmark is great because it shows the the things that make up image quality that are hard to put into an advertisement. It seems that they really don't care much about pixel count. ( Nikon salesman training says that 6MP is enough for most people, which is why they were the last to abandon the 6MP SLR). Us ole film users who miss dynamic range are excited about the 14 stop dynamic range of the new Pentax cameras. Pentax has, as usual for them, chosen the truthful path and "only" has 14 bit color on these cameras. Apparently on the 16 bit cameras the bottom two bits have no information. It just looks good. Since film is at least 16 bit color, manufacturers want to use it to show that digital is fully equivalent to film. Anyway, after the rant, I'm excited to start saving up my sheckels to get a K-5.

[falconeye]

I don't get the rant but what you just said about pixels is a common misconeption of what DxO says.

What they really say is this:

We treat the sensor as an information capturing device with essentially two dimensions:

1. a surface area, which is nothing but the pixel count.
2. a depth per normalized surface area element, expressed in bits or f-stops or as they choose, DxOMarks

So, the only way they want their charts read is by the X/Y chart of DxOMarks vs. MPixels.

BTW, If you look at that, all the cameras at the outer edge of the band are interesting ones: S5Pro, K-5, D3s, D3x, P65+

[Class A]

I have also been a bit surpised in the past by apparent discrepancies between DPReview and DXO results.

The discrepancy arises because DxO measures sensors (RAW files) and DPR measures tone curves (JPEGs).

The DPR results are irrelevant (except for JPEG shooters). They are just about in-camera tone curve shapes. Their scores increasingly reward cameras the more they deviate from a standard sRGB tone curve. By flattening the tone curve at the top end you can fool DPR into believing that the camera has a lot of "highlight headroom".

In short, ignore the DPR results. The DxO measurements, on the other hand, give you solid data that is indicative of what you can achieve with a camera in the field. A lot of the final result is based on PP but the sensor quality measured by DxO defines the base point from on which PP can build on.

[Agnostic]

The discrepancy arises because DxO measures sensors (RAW files) and DPR measures tone curves (JPEGs).

The DPR results are irrelevant (except for JPEG shooters). They are just about in-camera tone curve shapes. Their scores increasingly reward cameras the more they deviate from a standard sRGB tone curve. By flattening the tone curve at the top end you can fool DPR into believing that the camera has a lot of "highlight headroom".

In short, ignore the DPR results. The DxO measurements, on the other hand, give you solid data that is indicative of what you can achieve with a camera in the field. A lot of the final result is based on PP but the sensor quality measured by DxO defines the base point from on which PP can build on.

Thanks. I'll ignore the image quality results from DPR from now on. I never shoot JPG and the response I can expect from the raws in pp is really what matters to me.
Those recovered underexposed files from the K5 that several people refered to are really impressive. They seem to indicate you can expose for the highlights and still recover lots of shadow detail in pp without loosing much detail or introducing excessive noise.

[dosdan]

By the way does anybody have any data on the dynamic range distribution? Is there a distribution curve up somewhere? How much of it is in shadow and highlight areas?

Probably does not matter too much with raw. According to GordonBGood the K-5 seems to be exposing fairly heavily to the right (+2/3rd EV overexposure in the example he discussed):

http://forums. dp review .com/forums/read.asp?forum=1036&message=36897725

(remove spaces in URL)

So this will probably reduce clipping headroom. But if you apply a little bit of negative EVcomp in camera, you still have an awful lot of DR to work with in PP when you boost EV, manipulate the tone curve or dodge.

Dan.

[Rite]

THis have been fiscussed already . DXO is base in France and that the Euro price converted in USD
In Canada , I am buying the K-5 for $1420 but our $ in stranger than the USD vs Yen

Probably does not matter too much with raw. According to GordonBGood the K-5 seems to be exposing fairly heavily to the right (2/3rd EV in the example he discussed):

http://forums. dp review .com/forums/read.asp?forum=1036&message=36897725

(remove spaces in URL)

So this will probably reduce clipping headroom. But if you apply a little bit of negative EVcomp in camera, you still have an awful lot of DR to work with in PP when you boost EV, manipulate the tone curve or dodge.

Dan.

===============================

Pentax ownership bias aside, if you objectively looked at the K-5, D7000, and 60D, which one would you go for?

[dgaies]

===============================

Pentax ownership bias aside, if you objectively looked at the K-5, D7000, and 60D, which one would you go for?

IQ is going be fantastic using any of those bodies, so I would never select based on the tiny differences you have to pixel peep to see. For me, it would come down to the differences in features between the bodies and the selection of lenses available for system. I like small primes and in-body SR, so I would (and did) select the K-5. If I was into longer glass, more zoom choices and/or flash photography, I might be more inclined to try the Canon or Nikon.

[m8o]

IQ is going be fantastic using any of those bodies, so I would never select based on the tiny differences you have to pixel peep to see. For me, it would come down to the differences in features between the bodies and the selection of lenses available for system. I like small primes and in-body SR, so I would (and did) select the K-5. If I was into longer glass, more zoom choices and/or flash photography, I might be more inclined to try the Canon or Nikon.

Voice of Reason Alert!

[Rite]

IQ is going be fantastic using any of those bodies, so I would never select based on the tiny differences you have to pixel peep to see. For me, it would come down to the differences in features between the bodies and the selection of lenses available for system. I like small primes and in-body SR, so I would (and did) select the K-5. If I was into longer glass, more zoom choices and/or flash photography, I might be more inclined to try the Canon or Nikon.

==============================

Excellent answer. IMO, the K-5 has the advantage in the primes and in-body SR, just as you stated. To get the same mag alloy body in the Canon, you have to go to the 7D (which is only $100 more than the K-5). The 7000D DOES already have the mag alloy body.

This year really is a buyer's paradise.

I know you will enjoy immensely your K-5.

[UnknownVT]

Interesting that the Pentax K-5 tops the DxOMark ranking for APS-C sensors followed very closely by the also new Nikon D7000.

DxOMark ratings:
(Low-Light ISO) (Dynamic range)
Pentax K-5 1162 ISO; 14.1 Evs
Nikon D7000 1167 ISO; 13.9 Evs
Pentax K-x 811 ISO; 12.5 Evs
Pentax K-r 755 ISO; 12.4 Evs

It would appear that the Pentax K-5 and Nikon D7000 are pretty close;
the K-r and K-x are also close to each other.

A few points might be measurable - but they are unlikely to make much practical difference in real-world photography - however the K-5 and Nikon D7000 both seem to have significant improvement according to the DxO figures over the K-r and K-x.

It's a pity that Imaging-Resource have not published their standardized test images for the K-5 (or K-r) so we can see some results that we can compare directly with known cameras - however they have recently put up the test samples for the Nikon D7000.

I know this is not comprehensive - but I just wanted to see if the images had any significant improvement at higher ISO JPG, as that is what interests me using the Nikon D7000 as a sort of "substitute" for the K-5 compared to the existing K-x which had outstanding High ISO performance for its time to see if the K-5 is likely to have noticeable improvement at high ISO over the K-x (hope that makes some sense) -

ISO 6400



ISO12800



It seems to me that visually there is no major significant improvement for High ISO images of the Nikon D7000 (K-5 likely to be similar?) over the existing K-x - other than more pixels. But please remeber this is a very limited view using JPGs - one could download the RAW versions and process them and do a similar comparison - but this quick and dirty comparison shows me for the time being at least that this new generation of dSLRs (K-5 and Nikon D7000) are improvement for 16+Mp but not a significant visual improvement at Higher ISO than the K-x (and probably the K-r) despite the significant improvement in the DxO figures