K-5 IIs vs K-5 II - can you tell the difference?

adi3000 · 11-13-2012, 04:11 PM

Sharpening jpgs have no use for me, sorry. I have already an opinion. I did a lot of sharpening this days (from K5, K5II, K5IIs, 7D, Nex-5N, 5DmkII). I have enough. I just sell my K5 to go with K5IIs. We will see.
K5 was wonderful but no in low light focusing and for that he had to go.

NVSteve · 11-13-2012, 04:26 PM

Originally posted by JohnBee

Since the images are from different system, it wouldn't be very helpful to process them identically. That being said, we can however address the inherent difference between both systems. ie. in this case, we can counter the effects of the AA filter by deblurring the K-5 II RAW image to match the K-5 IIs. At which point both images can then be developed accordingly.

Let me know if this makes sense.

No, it does not make sense to me. Why spend the effort to get a pseudo-similar image when the more helpful of approaches would be to either present both the II and IIs unaltered, or both after post processing? They both use the same sensor & take the same lenses, so they are only different as to AA filter or no AA filter. Post processing both would be beneficial because then we could see if the II can come close to the IIs in real world usage & by using techniques the majority of us use when processing RAW.

JohnBee · 11-13-2012, 04:26 PM

Originally posted by adi3000

Sharpening jpgs have no use for me, sorry. I have already an opinion. I did a lot of sharpening this days (from K5, K5II, K5IIs, 7D, Nex-5N, 5DmkII). I have enough. I just sell my K5 to go with K5IIs. We will see.
K5 was wonderful but no in low light focusing and for that he had to go.

FTR. I have the original TIFF's for download for anyone interested in trying their hand at it.

JohnBee · 11-13-2012, 04:31 PM

Originally posted by NVSteve

No, it does not make sense to me. Why spend the effort to get a pseudo-similar image when the more helpful of approaches would be to either present both the II and IIs unaltered, or both after post processing? They both use the same sensor & take the same lenses, so they are only different as to AA filter or no AA filter. Post processing both would be beneficial because then we could see if the II can come close to the IIs in real world usage & by using techniques the majority of us use when processing RAW.

I'm thinking this may delve in a JPG vs RAW type scenario, though for a RAW shooter, adding a deblur adjustment in ones profile would be seamless. ie. many of us have been doing this with the K-5 classic due to it's rather heavy handed AA filter. However, it could see people investing in the IIs version if they were shooting OOC images(JPG) for example.

Though above all, I think it's comes down to whether or not people are willing to deal with the effects of moire and aliasing in JPG vs that of having the ultimate IQ in RAW that comes-up .

PS. I'm not trying ignore your point on final processing btw. And I agree that this would ultimately show what each system can do. However, since this particular exercise was designed to address the effects of the AA filter exclusively. This method gives everyone the opportunity to process the files according to their own likes and dislikes. - focusing on the effects and differences of the AA filter.

Last edited by JohnBee; 11-13-2012 at 05:10 PM.

RBellavance · 11-13-2012, 04:46 PM

Originally posted by JohnBee

I'm thinking this may delve in a JPG vs RAW type scenario, though for a RAW shooter, adding a deblur adjustment in ones profile would be seamless. ie. many of us have been doing this with the K-5 classic due to it's rather heavy handed AA filter. However, it could see people investing in the IIs version if they were shooting OOC images(JPG) for example.

Well, in all fairness, one could also change the sharpening settings of the OOC jpegs of the K-5 II as well...

Quote:

Though above all, I think its whether people want to deal with the effects of moire and aliasing in JPG vs that of having the ultimate IQ in RAW that comes-up .

I understand what you did and why you did it this way. I think it's a good idea. And it reinforces my leaning towards the regular K-5 II instead of the IIs. To me, the theorical advantage in resolution is not worth the risk of moire. Speaking of which, this particular scene + distance + lens/aperture seems to land right in the IIs' moire-producing sweet spot, doesn't it ?

JohnBee · 11-13-2012, 04:55 PM

Originally posted by RBellavance

Well, in all fairness, one could also change the sharpening settings of the OOC jpegs of the K-5 II as well...

Yes and I'd say that makes good sense.

And I'd add that the sharpening applied here(deconvoluted) isn't your typical type sharpening(see: deblurring). And so when we consider that the K-5 II image suffers from the effects of the AA filter(softness), we find that K-5 IIs doesn't benefit from the same adjustment.

That being said, I think I might have been better off putting up a small paragraph to better explain the deblurring and its effects on the K-5 II image rather than using the term sharpening as I did. At any rate, if you deblur the K-5 IIs, image in the same way, it quickly degrades past the point of perceptual rendering into what I'd call a pixelated or over-amplification of contrast details. - if that makes sense.

close-enough · 11-13-2012, 06:02 PM

JohnBee, Have you tried focusing with Photo Ninja. Class A quoted their spiel saying that they had a special de-ant-alias blurring. They are made by the Noise Ninja folks. I see they have killed Noise Ninja and replaced it with an entire RAW developer. If you can't join Lightroom, beat it?

PictureCode home page: Photo Ninja
PictureCode home page: Photo Ninja

JohnBee · 11-13-2012, 06:33 PM

Originally posted by close-enough

JohnBee, Have you tried focusing with Photo Ninja. Class A quoted their spiel saying that they had a special de-ant-alias blurring. They are made by the Noise Ninja folks. I see they have killed Noise Ninja and replaced it with an entire RAW developer. If you can't join Lightroom, beat it?

PictureCode home page: Photo Ninja
PictureCode home page: Photo Ninja

Totally new to me.
But it looks exciting

I'm going to give it a whirl tonight.

JohnBee · 11-13-2012, 09:46 PM

I've since taken the liberty of sharpening both files using FocalBlade:

Pentax K-5 IIs

Pentax K-5 II

While I cannot speak for others, I honestly can't find any advantage to the K-5 IIs in this particular scene.
Though if we factor-in moire and false color, then I suppose the K-5 II gains the upper hand in this particular race.

JinDesu · 11-13-2012, 09:52 PM

Please give more information on what you did - again, when I used photoshop's unsharp mask with the same settings for both RAW files, I could see a difference. I would like to understand what you did with Focal Blade that resulted in practically no difference.

JohnBee · 11-13-2012, 10:08 PM

Originally posted by JinDesu

Please give more information on what you did - again, when I used photoshop's unsharp mask with the same settings for both RAW files, I could see a difference. I would like to understand what you did with Focal Blade that resulted in practically no difference.

Okay let me summarize from the top:

1) Developed both files in RawTherapee (yields more detail than ACR or LR4)
2) Applied Deconvoluted sharpening to the K-5 II image (compensate for AA filter blur). - This is a Raw Therapee feature, though Topaze Detail is an equivalent if you're limited to Photoshop. Settings are: 68 rad & 28 strength.
3) Sharpened both images using FocalBlade: i) K-5 IIs using default settings w/reduced strength just short of artifacts(halo). ii) K-5 II sharpened as follows:

Hope this helps.

PS. looking at the image again, I think the radius could go down a few notches on the K-5 II

JinDesu · 11-13-2012, 10:10 PM

So to be clear - you first sharpened the k-5II file, then sharpened both with Focal Blade?

JohnBee · 11-13-2012, 10:20 PM

Originally posted by JinDesu

So to be clear - you first sharpened the k-5II file, then sharpened both with Focal Blade?

The adjustment made to the K-5 II in RawTherapee is called Deconvoluted Sharpening. Which is a fancy name for deblurring. Not to be confused with unsharp mask type methods, which is why it is appropriate to apply this particular adjustment at the beginning rather than the end of the post processing workflow. That being said, if you're not using Raw Therapee, you can apply the same type of deblurring using Topaz Detail. After which, both files are subject to sharpening as usual. I used FocalBlade in this case, though I'm confident Smart sharpen would work just as well.

PS. I just tried something and discovered that you can forgo the deblurring altogether and render both files with FocalBlade alone. (never really considered this would work tbh). Which doesn't produce as fine a result as the deblur > sharpening method, though it certainly simplifies the process.

JohnBee · 11-14-2012, 12:24 PM

Okay so I had chance to play around some more today and I managed to adjust both images to better reflect their inherent capacity for detail. The first thing I noticed here is where the K-5 IIs is prone to creating false detail due to the effect of moire .

Having said that, I also noticed that though there seem to be inherent differences between the K-5 II and K-5 IIs, it appears as though these can all be addressed in post. Though as far as final detail goes, it looks as though whatever is lost to moire can never be regained, whereas what is obfuscated by the AA filter is easilly recoverable.

Click for full size

Click for full size

Click for full size

Click for full size

And for reference:
Full size image K-5 II
Full size image K-5 IIs

Hope this helps

wahid_satu · 11-15-2012, 03:47 AM

Originally posted by JohnBee

Though as far as final detail goes, it looks as though whatever is lost to moire can never be regained, whereas what is obfuscated by the AA filter is easilly recoverable.

In RawTherapee we can choose between 7 demosaic algorithm. I can't remember when I use any algorithm other than the default Amaze but when tinkering with DPR K-5IIs raws I found that moire can be reduced by using the hphd algorithm and setting the false color suppression steps to 5. I found that this will substantially reduce moire on the sides of Cesar's (I think that's him) head and the VW beetles' back. Moire on the coin face above mickey and the horizontal line card left of Baileys are only slightly reduced, though. This method has no effect on sharpness, and slight saturation reduction on some colors.

IMHO this shows there are variance of moire(s?), and that depending on what type of moire we encounter the post processing tax of removing moire manually can be reduced considerably when we shoot raw. Also, that there may well be better Algorithm in the future which will help reduce moire even more, as far fetched as it may seem.

11-13-2012, 04:55 PM	#36
JohnBee Veteran Member Join Date: Sep 2008 Location: Newrfoundland Photos: Albums Posts: 4,667 Original Poster	Originally posted by RBellavance Well, in all fairness, one could also change the sharpening settings of the OOC jpegs of the K-5 II as well... Yes and I'd say that makes good sense. And I'd add that the sharpening applied here(deconvoluted) isn't your typical type sharpening(see: deblurring). And so when we consider that the K-5 II image suffers from the effects of the AA filter(softness), we find that K-5 IIs doesn't benefit from the same adjustment. That being said, I think I might have been better off putting up a small paragraph to better explain the deblurring and its effects on the K-5 II image rather than using the term sharpening as I did. At any rate, if you deblur the K-5 IIs, image in the same way, it quickly degrades past the point of perceptual rendering into what I'd call a pixelated or over-amplification of contrast details. - if that makes sense. Last edited by JohnBee; 11-13-2012 at 05:12 PM.

11-13-2012, 09:46 PM	#39
JohnBee Veteran Member Join Date: Sep 2008 Location: Newrfoundland Photos: Albums Posts: 4,667 Original Poster	I've since taken the liberty of sharpening both files using FocalBlade: Pentax K-5 IIs Pentax K-5 II While I cannot speak for others, I honestly can't find any advantage to the K-5 IIs in this particular scene. Though if we factor-in moire and false color, then I suppose the K-5 II gains the upper hand in this particular race. Last edited by JohnBee; 11-13-2012 at 09:56 PM.

11-13-2012, 10:08 PM	#41
JohnBee Veteran Member Join Date: Sep 2008 Location: Newrfoundland Photos: Albums Posts: 4,667 Original Poster	Originally posted by JinDesu Please give more information on what you did - again, when I used photoshop's unsharp mask with the same settings for both RAW files, I could see a difference. I would like to understand what you did with Focal Blade that resulted in practically no difference. Okay let me summarize from the top: 1) Developed both files in RawTherapee (yields more detail than ACR or LR4) 2) Applied Deconvoluted sharpening to the K-5 II image (compensate for AA filter blur). - This is a Raw Therapee feature, though Topaze Detail is an equivalent if you're limited to Photoshop. Settings are: 68 rad & 28 strength. 3) Sharpened both images using FocalBlade: i) K-5 IIs using default settings w/reduced strength just short of artifacts(halo). ii) K-5 II sharpened as follows: Hope this helps. PS. looking at the image again, I think the radius could go down a few notches on the K-5 II Last edited by JohnBee; 11-13-2012 at 10:34 PM.

11-13-2012, 10:20 PM	#43
JohnBee Veteran Member Join Date: Sep 2008 Location: Newrfoundland Photos: Albums Posts: 4,667 Original Poster	Originally posted by JinDesu So to be clear - you first sharpened the k-5II file, then sharpened both with Focal Blade? The adjustment made to the K-5 II in RawTherapee is called Deconvoluted Sharpening. Which is a fancy name for deblurring. Not to be confused with unsharp mask type methods, which is why it is appropriate to apply this particular adjustment at the beginning rather than the end of the post processing workflow. That being said, if you're not using Raw Therapee, you can apply the same type of deblurring using Topaz Detail. After which, both files are subject to sharpening as usual. I used FocalBlade in this case, though I'm confident Smart sharpen would work just as well. PS. I just tried something and discovered that you can forgo the deblurring altogether and render both files with FocalBlade alone. (never really considered this would work tbh). Which doesn't produce as fine a result as the deblur > sharpening method, though it certainly simplifies the process. Last edited by JohnBee; 11-14-2012 at 12:50 PM.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
K-5 K-5 II K-5 IIs, e dial issue, Noise	Eder	Pentax K-5 & K-5 II	17	06-03-2014 12:33 AM
Pentax K-5 II and K-5 IIs Now In Stock!	Adam	Pentax K-5 & K-5 II	24	10-25-2012 03:38 PM
Suggestion K-5 forum renamed to K-5, K-5 II, and K-5 IIs?	Buschmaster	Site Suggestions and Help	2	10-16-2012 02:17 PM
Pentax could we get the K-5 II / K-5 IIs in color please?	Hegemon	Pentax DSLR Discussion	17	10-14-2012 01:59 AM
Facts and rumors regarding the 16.28 Megapixel sensor in K- 5 II and K-5 IIs	F-pix	Pentax K-5 & K-5 II	11	09-28-2012 08:17 AM

11-13-2012, 04:11 PM	#31
adi3000 New Member Join Date: Dec 2010 Location: Bucharest Posts: 20	Sharpening jpgs have no use for me, sorry. I have already an opinion. I did a lot of sharpening this days (from K5, K5II, K5IIs, 7D, Nex-5N, 5DmkII). I have enough. I just sell my K5 to go with K5IIs. We will see. K5 was wonderful but no in low light focusing and for that he had to go.

11-13-2012, 04:26 PM	#32
NVSteve Junior Member Join Date: Oct 2012 Location: Salt Lake City Posts: 47	Originally posted by JohnBee Since the images are from different system, it wouldn't be very helpful to process them identically. That being said, we can however address the inherent difference between both systems. ie. in this case, we can counter the effects of the AA filter by deblurring the K-5 II RAW image to match the K-5 IIs. At which point both images can then be developed accordingly. Let me know if this makes sense. No, it does not make sense to me. Why spend the effort to get a pseudo-similar image when the more helpful of approaches would be to either present both the II and IIs unaltered, or both after post processing? They both use the same sensor & take the same lenses, so they are only different as to AA filter or no AA filter. Post processing both would be beneficial because then we could see if the II can come close to the IIs in real world usage & by using techniques the majority of us use when processing RAW.

11-13-2012, 04:26 PM	#33
JohnBee Veteran Member Join Date: Sep 2008 Location: Newrfoundland Photos: Albums Posts: 4,667 Original Poster	Originally posted by adi3000 Sharpening jpgs have no use for me, sorry. I have already an opinion. I did a lot of sharpening this days (from K5, K5II, K5IIs, 7D, Nex-5N, 5DmkII). I have enough. I just sell my K5 to go with K5IIs. We will see. K5 was wonderful but no in low light focusing and for that he had to go. FTR. I have the original TIFF's for download for anyone interested in trying their hand at it.

11-13-2012, 04:31 PM	#34
JohnBee Veteran Member Join Date: Sep 2008 Location: Newrfoundland Photos: Albums Posts: 4,667 Original Poster	Originally posted by NVSteve No, it does not make sense to me. Why spend the effort to get a pseudo-similar image when the more helpful of approaches would be to either present both the II and IIs unaltered, or both after post processing? They both use the same sensor & take the same lenses, so they are only different as to AA filter or no AA filter. Post processing both would be beneficial because then we could see if the II can come close to the IIs in real world usage & by using techniques the majority of us use when processing RAW. I'm thinking this may delve in a JPG vs RAW type scenario, though for a RAW shooter, adding a deblur adjustment in ones profile would be seamless. ie. many of us have been doing this with the K-5 classic due to it's rather heavy handed AA filter. However, it could see people investing in the IIs version if they were shooting OOC images(JPG) for example. Though above all, I think it's comes down to whether or not people are willing to deal with the effects of moire and aliasing in JPG vs that of having the ultimate IQ in RAW that comes-up . PS. I'm not trying ignore your point on final processing btw. And I agree that this would ultimately show what each system can do. However, since this particular exercise was designed to address the effects of the AA filter exclusively. This method gives everyone the opportunity to process the files according to their own likes and dislikes. - focusing on the effects and differences of the AA filter. Last edited by JohnBee; 11-13-2012 at 05:10 PM.

11-13-2012, 04:46 PM	#35
RBellavance Veteran Member Join Date: Nov 2006 Location: Near Montréal, Canada Photos: Gallery Posts: 1,716	Originally posted by JohnBee I'm thinking this may delve in a JPG vs RAW type scenario, though for a RAW shooter, adding a deblur adjustment in ones profile would be seamless. ie. many of us have been doing this with the K-5 classic due to it's rather heavy handed AA filter. However, it could see people investing in the IIs version if they were shooting OOC images(JPG) for example. Well, in all fairness, one could also change the sharpening settings of the OOC jpegs of the K-5 II as well... Quote: Though above all, I think its whether people want to deal with the effects of moire and aliasing in JPG vs that of having the ultimate IQ in RAW that comes-up . I understand what you did and why you did it this way. I think it's a good idea. And it reinforces my leaning towards the regular K-5 II instead of the IIs. To me, the theorical advantage in resolution is not worth the risk of moire. Speaking of which, this particular scene + distance + lens/aperture seems to land right in the IIs' moire-producing sweet spot, doesn't it ?

11-13-2012, 06:02 PM	#37
close-enough Junior Member Join Date: Apr 2009 Location: Portland, Oregon Posts: 44	JohnBee, Have you tried focusing with Photo Ninja. Class A quoted their spiel saying that they had a special de-ant-alias blurring. They are made by the Noise Ninja folks. I see they have killed Noise Ninja and replaced it with an entire RAW developer. If you can't join Lightroom, beat it? PictureCode home page: Photo Ninja PictureCode home page: Photo Ninja

11-13-2012, 06:33 PM	#38
JohnBee Veteran Member Join Date: Sep 2008 Location: Newrfoundland Photos: Albums Posts: 4,667 Original Poster	Originally posted by close-enough JohnBee, Have you tried focusing with Photo Ninja. Class A quoted their spiel saying that they had a special de-ant-alias blurring. They are made by the Noise Ninja folks. I see they have killed Noise Ninja and replaced it with an entire RAW developer. If you can't join Lightroom, beat it? PictureCode home page: Photo Ninja PictureCode home page: Photo Ninja Totally new to me. But it looks exciting I'm going to give it a whirl tonight.

11-13-2012, 09:52 PM	#40
JinDesu Veteran Member Join Date: Jun 2011 Location: New York City Photos: Gallery Posts: 5,638	Please give more information on what you did - again, when I used photoshop's unsharp mask with the same settings for both RAW files, I could see a difference. I would like to understand what you did with Focal Blade that resulted in practically no difference.

11-13-2012, 10:10 PM	#42
JinDesu Veteran Member Join Date: Jun 2011 Location: New York City Photos: Gallery Posts: 5,638	So to be clear - you first sharpened the k-5II file, then sharpened both with Focal Blade?

11-14-2012, 12:24 PM	#44
JohnBee Veteran Member Join Date: Sep 2008 Location: Newrfoundland Photos: Albums Posts: 4,667 Original Poster	Okay so I had chance to play around some more today and I managed to adjust both images to better reflect their inherent capacity for detail. The first thing I noticed here is where the K-5 IIs is prone to creating false detail due to the effect of moire . Having said that, I also noticed that though there seem to be inherent differences between the K-5 II and K-5 IIs, it appears as though these can all be addressed in post. Though as far as final detail goes, it looks as though whatever is lost to moire can never be regained, whereas what is obfuscated by the AA filter is easilly recoverable. Click for full size Click for full size Click for full size Click for full size And for reference: Full size image K-5 II Full size image K-5 IIs Hope this helps

11-15-2012, 03:47 AM	#45
wahid_satu Forum Member Join Date: Oct 2010 Location: Jakarta Photos: Gallery Posts: 96	Originally posted by JohnBee Though as far as final detail goes, it looks as though whatever is lost to moire can never be regained, whereas what is obfuscated by the AA filter is easilly recoverable. In RawTherapee we can choose between 7 demosaic algorithm. I can't remember when I use any algorithm other than the default Amaze but when tinkering with DPR K-5IIs raws I found that moire can be reduced by using the hphd algorithm and setting the false color suppression steps to 5. I found that this will substantially reduce moire on the sides of Cesar's (I think that's him) head and the VW beetles' back. Moire on the coin face above mickey and the horizontal line card left of Baileys are only slightly reduced, though. This method has no effect on sharpness, and slight saturation reduction on some colors. IMHO this shows there are variance of moire(s?), and that depending on what type of moire we encounter the post processing tax of removing moire manually can be reduced considerably when we shoot raw. Also, that there may well be better Algorithm in the future which will help reduce moire even more, as far fetched as it may seem.