[Rondec]

The K-1 looks fine to me. People are arguing about tracking auto focus. I expect it to be better than the K3 (which truthfully was fine for most normal purposes), but probably not D5 level. Once again, Pentax isn't selling to the sports market yet. And this camera won't be ideal for sports from a number of other issues -- particularly frame rate. My wife shoots with a K3 and it is fine for wedding photography. She has not had issues tracking the bride and groom -- biggest problem comes in situations where you can't use flash and pushing the iso up adds noise.

I am hoping that low iso is really good on this camera. Just lowering the bottom iso doesn't necessarily add a bunch of dynamic range. Canon sensors notoriously don't have great dynamic range at low iso and the A7s and A7r II actually don't have as good low iso dynamic range as the sensor in the D810 and A7r. I had heard that the sensor in the D810 was specifically manufactured for Nikon with deeper photon wells which allowed for better DR at low iso and actual lower base iso. I imagine we won't see that here.

We will see when the final specs come out and when (if ever) DXO Mark actually tests it, but to me, it looks like a good start and those whining really need to settle down a little.

[biz-engineer]

Not really on a sunny day keeping at say f/5.6 you can get iso 25, 1/100s that is fast enough for many shots. Couting that many lenses are at their best around f/5.6 that not too bad. I shoot many pictures in day light and most of the time you have quite some margin on the shutter speed for many subjects.

Yes, according to the sunny f16 rule (which practically works), in sunny weather, you could use ISO25 handheld! The only thing is that the best light for photography is not in the middle of a sunny day (deep shadows) and that's especially in sunny conditions that the dynamic range of the scenes largely exceed the dynamic range capability of digital sensors because of the huge difference in light intensity between the areas exposed to the sun and the shadowed areas. On top of this, light intensity does not place your lenses in the best conditions. So, if you want greatly defined tones (for example see what Rondec is doing early morning just before sunrise or at sunset times).
I get the best images in Autumn, early spring, when the sun is filtered out by the atmosphere or clouds. Sometimes the sun strikes through the clouds , so the amount of light is fairly good but I couldn't shoot at ISO25, best case ISO100 or 200.


---------- Post added 14-02-16 at 12:56 ----------

you don't get the same result by stacking shots than from a continuous exposure

Yes, you do, that's why people use both long exposures or stacking for photographying stars.

in most cases because one could do HDR and use ND filters

HDR does not give natual rendering. If you'd like to cope with high dynamic range scenes and still get naturally rendered photos, you could use Expose Fusion, that works very well.
ND filters? I suppose you wanted to say Gradual ND filters, because ND filters are used to increase exposure times.

In my opinion, even 10ev of DR is sufficient is you are a good photographer (i.e if you master the light). Then of course, for photogs who don't really know what you are doing, the sky is the limit of DR!

[OTH]

Yes, according to the sunny f16 rule (which practically works), in sunny weather, you could use ISO25 handheld! The only thing is that the best light for photography is not in the middle of a sunny day (deep shadows) and that's especially in sunny conditions that the dynamic range of the scenes largely exceed the dynamic range capability of digital sensors because of the huge difference in light intensity between the areas exposed to the sun and the shadowed areas. On top of this, light intensity does not place your lenses in the best conditions. So, if you want greatly defined tones (for example see what Rondec is doing early morning just before sunrise or at sunset times).
I get the best images in Autumn, early spring, when the sun is filtered out by the atmosphere or clouds. Sometimes the sun strikes through the clouds , so the amount of light is fairly good but I couldn't shoot at ISO25, best case ISO100 or 200.

Best regards

That's right if you use a camera with a Bayer-Sensor. I'm always impressed of my Foveon DP2 at very intense light conditions. This is a beast at that conditions and the quality in bright, intense light is stunning.

[jaad75]

This is nonsense. Dynamic range always depends on the spatial frequency and DxO does the only meaningful thing in their print score. "Screen DR" is irrelevant.

Nope. If you want to know the senor characteristic on sensel level you need to know the range from the saturation point to the black readout noise level (from the OBP region). It doesn't matter if it's measured in dB (like sensors manufacturers usually do) or EV. The DxO shows you that measurement in "Screen" tab and even those should be read with a grain of salt, cause those last two bits in the shadows are just a garbage with no use (not to mention that the real use of theoretical DR is even more limited by the banding in some cameras). The "Print" tab in case of DR is always biased towards hi-resolution sensors and it's not a "measurement" per se, it's more like a "score" they give using their own rules, not the real world use (who said their 8MPix print resolution is the one you want to compare the cameras at? who in the real world downscales pictures without any PP, which almost always includes some form of noise reduction?), and that's why you get some peculiar results like a 14+ EV range using the sensor with 14-bit ADCs built-in.

[infoomatic]

I am so looking forward to this beast. The specs seem so mouthwatering to me, I guess the K-1 will be my ultimate gear. I just stocked up my storage system with 4x8TB harddisks so enough space hehe

[jaad75]

Sigma, Nikon and Pentax are using the same image processor (Fujitsu Milbeaut). The lowest ISO of this new image processor is ISO 100.

The base sensitivity is a matter of the sensor they use more than a processor.

[biz-engineer]

Nope. If you want to know the senor characteristic on sensel level you need to know the range from the saturation point to the black readout noise level (from the OBP region). It doesn't matter if it's measured in dB (like sensors manufacturers usually do) or EV. The DxO shows you that measurement in "Screen" tab and even those should be read with a grain of salt, cause those last two bits in the shadows are just a garbage with no use (not to mention that the real use of theoretical DR is even more limited by the banding in some cameras). The "Print" tab in case of DR is always biased towards hi-resolution sensors and it's not a "measurement" per se, it's more like a "score" they give using their own rules, not the real world use (who said their 8MPix print resolution is the one you want to compare the cameras at? who in the real world downscales pictures without any PP, which almost always includes some form of noise reduction?), and that's why you get some peculiar results like a 14+ EV range using the sensor with 14-bit ADCs built-in.

Falk his right in the way that he refers to a theoretical reference model to compare the overall image quality of image sensors where basically, all pixels are merged into one large pixel, all noise of all pixels is summed up into one large pixel and the amount of signal is the sum of signal from all pixels merged into one large pixel. That model provide a metric to compare overall image quality of an image taken with different sensors regardless of the sensor size. The reason why Falk considers special frequency is because if you sum everything up into one pixel and you compare the overall DR of a K-3 versus K-5, the K-3 will deliver a better IQ, and this is actually true, because random contribution of the noises is averaged out in the final image, thanks to further oversampling performed by the K-3 sensor versus K-5 sensor.

If you consider only one of the pixels of a sensor, this pixel has its own dymanic range, but this pixel is only one of the contributors to the global image. As you know, the dynamic range of an individual pixel is the difference between the maximum voltage level that the PIN junction can handle without breaking down (if this is match the ADC max input level, that's optimal), and the noise floor (dark level). The maximum level is does not depends much on the pixel size, the noise floor depends on the pixel size, in other words, with CMOS tech, DR depends essentially on the noise floor of each pixel and how many pixels are used to form an image. For example, if each pixel have 10ev of DR, and you have 50Mpixels (=Canon 5DSR), the overall image quality is still way better than the quality of the same image taken with a K-5 pixel having 14ev of DR !

DXO metrics are not fully correct in the way that Sony sensors are favored against Canon sensors. So, if you look at DXO, you'd think that the 5DSr is not good, but if you look at real pictures , you realize that it is simply not true.

---

Last edited by biz-engineer; 02-14-2016 at 05:30 AM.

[Nicolas06]

Yes, according to the sunny f16 rule (which practically works), in sunny weather, you could use ISO25 handheld! The only thing is that the best light for photography is not in the middle of a sunny day (deep shadows) and that's especially in sunny conditions that the dynamic range of the scenes largely exceed the dynamic range capability of digital sensors because of the huge difference in light intensity between the areas exposed to the sun and the shadowed areas. On top of this, light intensity does not place your lenses in the best conditions. So, if you want greatly defined tones (for example see what Rondec is doing early morning just before sunrise or at sunset times).
I get the best images in Autumn, early spring, when the sun is filtered out by the atmosphere or clouds. Sometimes the sun strikes through the clouds , so the amount of light is fairly good but I couldn't shoot at ISO25, best case ISO100 or 200.


---------- Post added 14-02-16 at 12:56 ----------


Yes, you do, that's why people use both long exposures or stacking for photographying stars.

Yeah end of day light is alway missing, I wonder how one could ever work bellow iso 100

FA77, f/5.6, 1/8000, iso 100



DA15, f/11, 1/200, iso 100.



FA77, f/3.2, 1/500, iso 100



FA77, f/8, 1/160, iso 100



FA77, f/5.6, 1/1600, iso 400 (obviously an error)



And we all know that pictures in daylight look crapy if it is not sunrise or sunset. No color deph, no saturation...

Almost 2 hours before sunset:

DA15, f/4, 1/1000s, iso 100



Near noon:

DA15, f/8, 1/250s, iso100



Man please stop saying thing are impossible when we are all doing it all the time.

[biz-engineer]

Yeah end of day light is alway missing, I wonder how one could ever work bellow iso 100

You have very deep shadows, some large areas of you photos are completely in the dark. So, as I wrote those are not the best lighting conditions to deliver the best photographs. But that's fine if you like that. Do you think if you had shot the same photos at ISO25 the dark areas would also be as dark? Or perhaps there are cases when the dark side give more drama to the scene , then you actually don't need more DR.

[Nicolas06]

You have very deep shadows, some large areas of you photos are completely in the dark. So, as I wrote those are not the best lighting conditions to deliver the best photographs. But that's fine if you like that. Do you think if you had shot the same photos at ISO25 the dark areas would also be as dark? Or perhaps there are cases when the dark side give more drama to the scene , then you actually don't need more DR.

So basically because you don't like dark shadows and some pictures samples (not all) here were done on purpose with theses then low iso setting are useless. And one of the picture is HDR, some have the sun in the frame... That's no accident, thanks.

Really man understand that not everybody will do the same as you, understand that some manage to take picture handled in daylight, even sunrise and sunset and still have some margin.

The thing is if you shoot a low contrast scene to begin with you don't need that much dynamic range. You want to say theses are the only picture worth taking, well that only apply to yourself.. But theses are the less demanding picture in term of gear. With a tripod, low dynamic range a compact camera or a smartphone could do just fine.

You just try to argue. First this is not possible. Then no daylight is not good, must use sunrise or sunset. Then sunrise or sunset pictures are still with too much contrast and too deep blacks (so too much dynamic range apparently) so it is still not the right settings. Even of course this is not the cases of all pictures shown Ahahah... You would not admit the sky is blue !

[Simen1]

This is ignoring physics. Dynamic range is basically a property of light (photon shot noise in shadows vs. highlights). For it to be large, you need much light, i.e. low iso.
A camera can do worse than physics alone allows, but e.g., a D810 already comes close.

QE at ISO 100 is around 50%, but it doesn't say anything about read noise and thus DR. Read noise is still notable higher then photon noise and can be physically improved by more then one stop. Well depth is also a notable restriction that may be increased. Physically Its nothing wrong by wishing for 20 stops DR (at ISO 1 or so) or wanting several stops better super high ISO performance.

[jaad75]

For example, if each pixel have 10ev of DR, and you have 50Mpixels (=Canon 5DSR), the overall image quality is still way better than the quality of the same image taken with a K-5 pixel having 14ev of DR !

DXO metrics are not fully correct in the way that Sony sensors are favored against Canon sensors. So, if you look at DXO, you'd think that the 5DSr is not good, but if you look at real pictures , you realize that it is simply not true.

No. The 5Ds has a better resolution, which is quite obvious, but it doesn't have wider dynamic range. You can pull the shadows up, downscale both to the same resolution and they still don't match:
Image comparison: Digital Photography Review
No K-5 there, but you have Fuji with (almost) the same IMX071.

[Mistral75]

I am quoting here this very interesting message from DragonLord, which recently drew my attention and which is perfectly relevant re. K-1's autofocus:

There is one thing I must call to attention regarding the AF array.

Notice that there is a significant gap between either side point and the main AF area. There are seven columns of selectable AF points, but nine columns of physical AF sensors:



(...)The extra columns appear to be unused.

—DragonLord

The two extra columns would be enabled on K-1, each for three AF points, thus giving a total of 27 + 3 + 3 = 33 AF points as observed.

Judging by the spacing of the two brackets inscribed on the focusing screen of K-1, as seen by reflection on the mirror and which most likely correspond to the outer AF points (as with other Pentax cameras), it seems that Ricoh Imaging has found a way to enlarge the AF points coverage compared to what the K-3 II module would spontaneously produce on K-1. IIRC Ricoh had filed a patent on this.

[monochrome]

The K-1 looks fine to me. People are arguing about tracking auto focus. I expect it to be better than the K3 (which truthfully was fine for most normal purposes), but probably not D5 level. Once again, Pentax isn't selling to the sports market yet. And this camera won't be ideal for sports from a number of other issues -- particularly frame rate. My wife shoots with a K3 and it is fine for wedding photography. She has not had issues tracking the bride and groom -- biggest problem comes in situations where you can't use flash and pushing the iso up adds noise.

I am hoping that low iso is really good on this camera. Just lowering the bottom iso doesn't necessarily add a bunch of dynamic range. Canon sensors notoriously don't have great dynamic range at low iso and the A7s and A7r II actually don't have as good low iso dynamic range as the sensor in the D810 and A7r. I had heard that the sensor in the D810 was specifically manufactured for Nikon with deeper photon wells which allowed for better DR at low iso and actual lower base iso. I imagine we won't see that here.

We will see when the final specs come out and when (if ever) DXO Mark actually tests it, but to me, it looks like a good start and those whining really need to settle down a little.

As always your comment is level, reasonable and to the point.

[Nicolas06]

No. The 5Ds has a better resolution, which is quite obvious, but it doesn't have wider dynamic range. You can pull the shadows up, downscale both to the same resolution and they still don't match:
Image comparison: Digital Photography Review
No K-5 there, but you have Fuji with (almost) the same IMX071.

And Canon said it officially, their 5DS is not made for high iso, it is made for resolution. Choose what matter most to you.