Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
10-06-2016, 02:51 AM   #16
Pentaxian
amoringello's Avatar

Join Date: May 2008
Location: Virginia, USA
Photos: Gallery
Posts: 1,393
QuoteOriginally posted by MyTZuS Quote
Meaning, I know that NR is miles ahead of the K3 II when in FF mode. But is it just as good when the crop mode is used?

Thanks.
Comparing the K1 FF to K1 crop? Short answer is "No". Its a crop.
The physics don't change simply because you ignore some pixels, so neither does the quality.

But that only applies if you don't change distance to subject, focal length, aperture and/or viewing/print size to try to create an "equivalent" image for both full and crop. It also applies only if you're competing the K1 to itself. If you're actually comparing against the K3II, etc... the technology is different so that takes a role as well.
You need to define your environment or you're going to get people saying a larger sensor make a difference, as they are assuming the above is changing to make an "equivalent" image.... which is very important in answering your question correctly.

10-06-2016, 05:34 AM - 1 Like   #17
Pentaxian




Join Date: Sep 2011
Location: Nelson B.C.
Posts: 3,211
My informal test is shooting loons and mergansers eating the kokanee who are spawning. The light is awful, the distance is daunting. I couldn't get any with the K5 when I tried it a few years ago. The K3 was ok if they were close, but noise limited shutter speed. The K1 is a different world. I crop substantially in these situations.

K1 500 f4.5 Sigma iso 2500 1/1250. 1830x1464 2.7m The K3 would not have gotten this shot. The K5 neither.

Drat. I made the mistake of rotating the off level using Google photo and the results are awful. I will upload a new one.

If I can fill the frame on the K1 the results are shockingly amazing. The dynamic range, the color is great. But cropping, which I often do is better than the K3. Something like this loon I could crop severely up to about iso 800. Past that the detail disappeared in noise. This shot would have been motion blurred at 1/400.

The crop mode gives me about the same frame rate as the K3 when holding down the af button.

10-06-2016, 08:47 PM   #18
Loyal Site Supporter
clackers's Avatar

Join Date: Jul 2013
Location: Melbourne
Photos: Albums
Posts: 7,685
QuoteOriginally posted by Fogel70 Quote

You still do not seem to understand the difference between pixel level noise and image level noise.
Fogel, it's clear you don't understand the difference between a real measurement (which is the so-called 'screen' one), versus the speculation about re-sampling they put in the 'print' one - it is *not* real data, they never measured a second time.

They make this quite clear if you read the explanation.

There is also *no* difference in the SNR due to resizing any image as you claim - all RAW files are collections of pixels.

That's only possible if you additionally resample - delete info in the original image by averaging.

Edit:

In other words, the 'Screen' data is DxoMark's physical measurement of the sensor - you can't do better than that - and the 'Print' stuff is their guess of what postprocessing might do.

They are *not* the same thing.

The less said about their "Landscape", "Sports" scores etc, the better ...

Last edited by clackers; 10-07-2016 at 01:54 AM.
10-06-2016, 11:32 PM   #19
Pentaxian




Join Date: Apr 2007
Location: Sweden
Posts: 1,883
QuoteOriginally posted by clackers Quote
Fogel, it's clear you don't understand the difference between a real measurement (which is the so-called 'screen' one), versus the speculation about re-sampling they put in the 'print' one - it is *not* real data, they never measured a second time.
I have a stick that I measured to be one meter long.

Now, without measuring, I can tell you that the middle of that stick is 0.5 meter from either end. I know that even though I havenít actually measured the distance to the middle, so it is *not* real data. Yet I believe it to be true.

Math is strange in that way, it can tell you things without having to measure it.

10-06-2016, 11:52 PM   #20
Pentaxian
Fogel70's Avatar

Join Date: Sep 2009
Location: Stockholm, Sweden
Photos: Albums
Posts: 1,737
QuoteOriginally posted by clackers Quote
Fogel, it's clear you don't understand the difference between a real measurement (which is the so-called 'screen' one), versus the speculation about re-sampling they put in the 'print' one - it is *not* real data, they never measured a second time.

They make this quite clear if you read the explanation.

There is also *no* difference in the SNR due to resizing any image as you claim - all RAW files are collections of pixels.

That's only possible if you additionally resample - delete info in the original image by averaging.

In other words, the 'Print' data is DxoMark's physical measurement of the sensor - you can't do better than that - and the 'Screen' stuff is their guess of what postprocessing might do.

They are *not* the same thing.

The less said about their "Landscape", "Sports" scores etc, the better ...
Thank you for proving my point that you quoted.
10-07-2016, 02:00 AM   #21
Loyal Site Supporter
clackers's Avatar

Join Date: Jul 2013
Location: Melbourne
Photos: Albums
Posts: 7,685
QuoteOriginally posted by Gimbal Quote
I have a stick that I measured to be one meter long.

Now, without measuring, I can tell you that the middle of that stick is 0.5 meter from either end. I know that even though I haven’t actually measured the distance to the middle, so it is *not* real data. Yet I believe it to be true.

Math is strange in that way, it can tell you things without having to measure it.
But of course, in your example, Gimbal, there is only one set of circumstances, and no guesswork about the outcome.

In going to the Print extrapolation, they do not measure, say, the use of the Lightroom or Raw Therapee noise reduction routines, nor admit that this is all scene dependent. When you do not want to destroy fine detail or introduce artifacts, you want very little noise reduction - you should get the cleanest possible image from your sensor in the first place. Instead, they just assume averaging to postprocess away noise, but never actually measure it.

Call me old fashioned, but they should stick to collecting data and let us as consumers work out the ramifications. Other review sites don't try this speculation when they do cameras or lenses.

Last edited by clackers; 10-07-2016 at 02:06 AM.
10-07-2016, 02:04 AM   #22
Loyal Site Supporter
clackers's Avatar

Join Date: Jul 2013
Location: Melbourne
Photos: Albums
Posts: 7,685
QuoteOriginally posted by Fogel70 Quote
Thank you for proving my point that you quoted.
Well, I do thank you, Fogel, for letting me know so I could correct my post. You didn't comment further, so I'll match your dignity and not add anything further myself.
10-07-2016, 02:40 AM   #23
Pentaxian
Fogel70's Avatar

Join Date: Sep 2009
Location: Stockholm, Sweden
Photos: Albums
Posts: 1,737
QuoteOriginally posted by clackers Quote
But of course, in your example, Gimbal, there is only one set of circumstances, and no guesswork about the outcome.
There is no guess work on DxOmark. They know how the pixels perform and can easily upscale that score for the full sensor as they know the the size and numbers of pixels on the sensor. It's just simple mathematics.

QuoteQuote:
In going to the Print extrapolation, they do not measure, say, the use of the Lightroom or Raw Therapee noise reduction routines, nor admit that this is all scene dependent. When you do not want to destroy fine detail or introduce artifacts, you want very little noise reduction - you should get the cleanest possible image from your sensor in the first place. Instead, they just assume averaging to postprocess away noise, but never actually measure it.
DxOmark do not measure resolution, so the averaging do affect the data they present. The averaging they do it to be able to present data that can easily be compared between cameras.
The same averaging happens when you view images on screen at the same size or print them at the same size.

QuoteQuote:
Call me old fashioned, but they should stick to collecting data and let us as consumers work out the ramifications. Other review sites don't try this speculation when they do cameras or lenses.
The problem is that it does not work, as all do not understand what the data represent.
Like FI when you read the "screen" data on DxOmark, and use that to compare sensor performance.

Some other may believe that a APS-C lens (used on a APS-C camera) that score higher LP/mm will outperform a FF lens with lower LP/mm used on a FF camera.

If data is not presented in a directly comparable format it will easy be misunderstood.

10-07-2016, 04:01 AM   #24
Veteran Member
MJKoski's Avatar

Join Date: Aug 2016
Posts: 691
Sony 16MP crop sensor (K5x) is from the same era as the 36MP sensor. They are similar in performance when driving K-1 in APS-C mode. This was the way with Nikon as well (D800 crop vs D7000).

That is about it. Now, 1:1 seems to have changed things somewhat as those crop lenses seem to work quite nicely with it.
10-08-2016, 07:18 AM   #25
Pentaxian




Join Date: Sep 2011
Location: Nelson B.C.
Posts: 3,211
QuoteOriginally posted by Fogel70 Quote
There is no guess work on DxOmark. They know how the pixels perform and can easily upscale that score for the full sensor as they know the the size and numbers of pixels on the sensor. It's just simple mathematics.

DxOmark do not measure resolution, so the averaging do affect the data they present. The averaging they do it to be able to present data that can easily be compared between cameras.
The same averaging happens when you view images on screen at the same size or print them at the same size.

The problem is that it does not work, as all do not understand what the data represent.
Like FI when you read the "screen" data on DxOmark, and use that to compare sensor performance.

Some other may believe that a APS-C lens (used on a APS-C camera) that score higher LP/mm will outperform a FF lens with lower LP/mm used on a FF camera.

If data is not presented in a directly comparable format it will easy be misunderstood.
The extrapolations don't represent reality. I did calculations based on the extrapolations that dxo does for the quality of cropping the K1 vs the K3. That is the type of shooting I do. If I crop the apsc capture to half the size, how does the quality compare to the full frame capture cropped down to the same size? My calculations were that slightly smaller than apsc the quality decreased more on the K1.

But in reality, it isn't the case. Maybe at iso 100-400, but above that the noise characteristics of the K3 are the limiting factor for cropping. Even if there are more pixels, there is less detail. The K1 at 3200 is about the same as K3 at 800. Increased shutter speed, stopped down lenses means that the image is of better quality images in the field.

I don't fault dxo. They provide a data point, and all else being equal gives a valid comparison. But all else isn't equal, especially between formats, and their data curves aren't representative of reality.
10-08-2016, 07:29 AM   #26
Pentaxian
normhead's Avatar

Join Date: Jun 2007
Location: Near Algonquin Park
Photos: Gallery | Albums
Posts: 23,715
I was out shooting yesterday with my K-3 and the day before with my K-1.

A few comparisons for you... one k-1 image, one k-3 image. The photographers advantage is not that one is better than the other. The photographers advantage is knowing which is better than the other in which situation, and using each appropriately. Both cameras can achieve excellent results.









You can make up all kinds of stuff about why the K-1 should be better, but from my perspective, you just shoot with the appropriate camera. IQ< refers to image quality. You cannot judge IQ from test charts, only from images. This whole test chart thing is a colossal waste of time an energy and analyzing DR and noise numbers. Get out and learn to use your gear. Find out by experience.

And honestly, 90% of the time, the camera I'll use will be the one that's in my hand. It's not like the differences matter, unless you walk into a situation where there's an image you can sell for big bucks. Then there are a few times where it might be worth your while to switch bodies.

I'll switch from a K-3 to a K-1 for sunsets or sunrises where the extra DR and better noise reduction might matter, and I'll switch to APS_c shooting birds at a distance, where more magnification and subject resolution using the same lens matters. Just taking snapshots of most day in day out kinds of shots, take the camera out of the bag you anticipate being the most appropriate, and use what's in your hand.

Shooting with the appropriate camera, you may get better IQ on some images, but even then, it's not a given. It's surprising how often little things, the low noise at 3200 on the K-1 or the wider DoF and subject magnification using APs-c jump up and bite you.

Last edited by normhead; 10-08-2016 at 07:53 AM.
10-08-2016, 08:45 AM   #27
Pentaxian
Fogel70's Avatar

Join Date: Sep 2009
Location: Stockholm, Sweden
Photos: Albums
Posts: 1,737
QuoteOriginally posted by derekkite Quote
The extrapolations don't represent reality. I did calculations based on the extrapolations that dxo does for the quality of cropping the K1 vs the K3. That is the type of shooting I do. If I crop the apsc capture to half the size, how does the quality compare to the full frame capture cropped down to the same size? My calculations were that slightly smaller than apsc the quality decreased more on the K1.

But in reality, it isn't the case. Maybe at iso 100-400, but above that the noise characteristics of the K3 are the limiting factor for cropping. Even if there are more pixels, there is less detail. The K1 at 3200 is about the same as K3 at 800. Increased shutter speed, stopped down lenses means that the image is of better quality images in the field.

I don't fault dxo. They provide a data point, and all else being equal gives a valid comparison. But all else isn't equal, especially between formats, and their data curves aren't representative of reality.
DxOmark data also show that K1 has better performance on APS-C crop than K3 II.
Their data show that FI K1 has around 0.5 stop advantage in APS-C crop on noise performance vs K3 II.
All test images I have seen show the same thing.

It would be interesting to see your images to compare. There are many things that can differ and affect the result, from individual camera that don't perform the same, software processing that are not exactly the same, performance on lenses...

See FI below result from dpr where K1 (FF) has ~1.5 stop NR advantage over K3 II (APS-C). On APS-C crop K1 would have ~0.5 stop advantage over K3 II.
Attached Images
 
10-08-2016, 10:47 AM   #28
Pentaxian
normhead's Avatar

Join Date: Jun 2007
Location: Near Algonquin Park
Photos: Gallery | Albums
Posts: 23,715
QuoteQuote:
DxOmark data also show that K1 has better performance on APS-C crop than K3 II
And yet my own actual field tests show practically no difference between a K-3 and a K-1 image, as posted above.
You're doing the "how many angles fit on the head of a pin?" thing. You're arguing doctrine, and dogma not experience.
No one gives a crap about photographic religious doctrine or your DxO bible.
10-08-2016, 11:08 AM   #29
Pentaxian
Fogel70's Avatar

Join Date: Sep 2009
Location: Stockholm, Sweden
Photos: Albums
Posts: 1,737
QuoteOriginally posted by normhead Quote
And yet my own actual field tests show practically no difference between a K-3 and a K-1 image, as posted above.
You're doing the "how many angles fit on the head of a pin?" thing. You're arguing doctrine, and dogma not experience.
No one gives a crap about photographic religious doctrine or your DxO bible.
My post was meant to comfirm that DxOmark show the same thing as derekkite experienced when using his cameras.
If scientific tests are properly performed they will mirror the result from real usage.

Your result is hardly any surprise either as the K1 was shot at ISO1600 and the K3 was shot at ISO400.
Which basically just conform what I said, as in these condition there will be difficult to see any difference in IQ between the shots. Unless you print very large where the extra resolution from K1 can make a difference.
10-08-2016, 11:15 AM   #30
Pentaxian
normhead's Avatar

Join Date: Jun 2007
Location: Near Algonquin Park
Photos: Gallery | Albums
Posts: 23,715
QuoteOriginally posted by Fogel70 Quote
My post was meant to comfirm that DxOmark show the same thing as derekkite experienced when using his cameras.
If scientific tests are properly performed they will mirror the result from real usage.

Your result is hardly any surprise either as the K1 was shot at ISO1600 and the K3 was shot at ISO400.
Which basically just conform what I said, as in these condition there will be difficult to see any difference in IQ between the shots. Unless you print very large where the extra resolution from K1 can make a difference.
I clearly didn't read your post carefully enough, my apologies.

But right off th stop, the high ISO difference isn't as much as one might think. The k-3 is shot at ISO 400, but to get the same DoF on the K-1 I would have had to shoot at ƒ8 and 800 ISO. So the actual improvement is about 1 stop. In these examples the K-3 image is actually the better image because shot at ƒ5.6 compared to ƒ4.5 (wide open) on the K-1 image the birds beak is in focus, but out of focus on the K-1 image. So there are times when the lower noise values of the K-1 aren't enough to compensate for the advantages the K-3 presents. Every set up presents it's challenges.

ISO 100 , and ƒ5.6, K-3 with my DA*200, K-1 with my Tamron SP AF 300, the K-3 set up will have more depth of field, and that may be enough to push the K-3 image ahead of the K-1 image, in some cases.

Last edited by normhead; 10-08-2016 at 11:26 AM.
Reply

Bookmarks
  • Submit Thread to Facebook Facebook
  • Submit Thread to Twitter Twitter
  • Submit Thread to Digg Digg
Tags - Make this thread easier to find by adding keywords to it!
crop, crop mode, dslr, ff, full frame, full-frame, iq comparison, k-1, k1, mode, pentax k-1
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
FF vs crop mode Timd Pentax K-1 5 06-08-2016 03:16 PM
FF mode - DA15 - crop factors & crop sizes acoufap Pentax K-1 6 05-09-2016 05:52 PM
How does a FF work in crop mode? Albermarle Pentax Full Frame 54 12-29-2015 11:28 AM
FF in crop mode for astro Conqueror Pentax Full Frame 6 09-15-2015 10:36 AM
(un-)importance of crop mode for FF ?!? Arne Bo Pentax Full Frame 54 03-21-2015 02:11 PM



All times are GMT -7. The time now is 06:06 AM. | See also: NikonForums.com, part of our network of photo forums!
  • Red (Default)
  • Green
  • Gray
  • Dark
  • Dark Yellow
  • Dark Blue
  • Old Red
  • Old Green
  • Old Gray
  • Dial-Up Style
Hello! It's great to see you back on the forum! Have you considered joining the community?
register
Creating a FREE ACCOUNT takes under a minute, removes ads, and lets you post! [Dismiss]
Top