My K-7 alpha test report and blog

jeffkrol · 06-16-2009, 04:38 AM

Quanitity of images w/ moire will determine how weak or strong the AA filter is.
Resolution tests as well.
The weakest AA Pentax seems to be the K10, strongest *ist-D...

From the very early reports of images looking soft but sharpening well points in the direction of a strong AA filter since this was an inherent quality in the D.
Wormy patterns or maze artifacts are more software related (Bayer demosaicing related and jpg construction) then mechanical and are not particularly worrisome.
Unlockers (thanks btw) parrot shot also seems to point to a strong AA filter.

Last edited by jeffkrol; 06-16-2009 at 04:44 AM.

RBellavance · 06-16-2009, 05:10 AM

Originally posted by Class A

The minute I saw it I thought this is an unnecessary complication. I was amazed to see many report that they often found their dial to be in a different position after having stored the camera in the bag. This has never happened to me once and frankly I cannot imagine how you'd have to handle a camera for this to happen.

I feel the same. It seems to me like a solution in search of a problem. I cannot recall a single instance of the mode dial being moved unintentionally over the 2.5 years I've had my K10D.

Class A · 06-16-2009, 05:22 AM

Originally posted by falconeye

It happens if you heavily sharpen a noise floor.

I had this thought as well, but the structures are there before the sharpening already, just pretty faintly.
I'd be surprised if they weren't a demosaicing artefact. Perhaps you can try another converter (I've read one can modify PEFs so that they can be read by existing converters)?

Originally posted by jeffkrol

From the very early reports of images looking soft but sharpening well points in the direction of a strong AA filter since this was an inherent quality in the D.

Not sure I agree a strong AA filter is a plus. Unless I'm missing something you can emulate a stronger AA filter with PP blurring but you cannot (quite) undo the blurring effect of a strong AA.

EDIT: The difference between a standard Canon 5D and one with the AA filter removed (scroll down to see the after/before images side by side) is quite impressive; like that between an outstanding lens and an OK lens.

Unlocker · 06-16-2009, 05:30 AM

That's all going a bit too in depth for my liking!

Just posted some high ISO shots taken on Sunday, and won't post any crops until / if I get final firmware before hand back, but thought you might like to dissect them!

RBellavance · 06-16-2009, 05:52 AM

Originally posted by Class A

Not sure I agree a strong AA filter is a plus. Unless I'm missing something you can emulate a stronger AA filter with PP blurring but you cannot (quite) undo the blurring effect of a strong AA.

I don't think you're missing anything. It is definitely not a plus in my book, as it lowers the effective resolution.

jeffkrol · 06-16-2009, 05:55 AM

Originally posted by Class A

Not sure I agree a strong AA filter is a plus. Unless I'm missing something you can emulate a stronger AA filter with PP blurring but you cannot (quite) undo the blurring effect of a strong AA.

EDIT: The difference between a standard Canon 5D and one with the AA filter removed (scroll down to see the after/before images side by side) is quite impressive; like that between an outstanding lens and an OK lens.

I'll check your link but an AA filter is a necessity until we get pixel counts high enough to not need them (that could be around a 50+mp range)... Sigma doesn't use one and many complain about jaggies.
It's a tradeoff and one to, I suppose, base your camera choice on. On one hand soft looking images, on the other jaggies and moire.
High quality lenses will usually give you the best results w/ even strong AA filters..
Marginal lenses, on the other hand, go to mush....
as frustrating as a strong AA filter, it is more frustrating to get moire patterns in textiles (and much harder to deal with)......
Why Pentax shifted from the weak AA (at least in 1 direction) on the k10 is unknown but some pre-production images had some terrible moire in some blue jeans and other images had it in different things as well.
EDIT: checking the Nikon photo the hot-rodded one shows moire everywhere (stop sign, AC on truck, AC on wall) and some of the detail, I suspect, is just artifacts from conversion, not true detail.
Best case would be a removeable AA filter so we can choose (ala Kodak back "in the day")
Screen capture added some artifacts as well as the now second conversion to jpg but there is plenty of detail and no moire and little "stepping", at least none objectionable: AT 200X

falconeye · 06-16-2009, 06:26 AM

Originally posted by Class A

EDIT: The difference between a standard Canon 5D and one with the AA filter removed (scroll down to see the after/before images side by side) is quite impressive; like that between an outstanding lens and an OK lens.

Look at the Area 3 crop, the shadow boundaries below the radiator. See the purple fringe to the left and green fringe to the right? False colors. Can be seen in K-7 videos at max. resolution as well and just some find it disturbing ...

Also, the hot rod image top in that page looks jaggy, something some didn't like in K-7 max.resolution video samples...

Also, in a side by side comparison, you have to sharpen the AA sample because it is meant to be sharpened.

IMHO, there is an optimum AA filter strength, not too weak so enough color information is recorded and not too strong to maintain resolution.

Last edited by falconeye; 06-16-2009 at 06:31 AM.

Class A · 06-16-2009, 06:40 AM

Originally posted by RBellavance

I don't think you're missing anything.

Actually, I thought about it again (using an audio analogy) and you cannot faithfully emulate an AA filter with software. In general, you may get artefacts that will be very difficult to remove without losing more resolution than the AA filter would have lost you.

As I don't shoot fashion as a professional I'd prefer a rather weak AA filter but none at all would probably be too limiting in practise.

Class A · 06-16-2009, 06:49 AM

Originally posted by falconeye

See the purple fringe to the left and green fringe to the right? False colors.

Yes, could be chroma moire, but the shot to right also has a hint of it and a lot more of it on the vertical lines surrounding the radiator. Perhaps the colour artefacts are caused by lens CA? Or the AA filter of the 5D should be slightly stronger to completely avoid colour moire?

Area 5 is a good example in favour of little/no AA. But you are right, the image to the right should be capture-sharpened (and the image to the left treated for luminance and color moire?). Anyhow.

Originally posted by falconeye

IMHO, there is an optimum AA filter strength, not too weak so enough color information is recorded and not too strong to maintain resolution.

I agree.

falconeye · 06-17-2009, 08:50 AM

Breaking news:

I've just received the final firmware release aka v1.00

Disclaimer: I have been informed that the firmware has been optimized for the sensor as used in series production K-7. The alpha test K-7 are equipped with a pre-series sensor (which was said to be 95% identical) and final final noise tests will have to await production cameras. Anyway, tomorrow I start to look into noise and resolution.

Unlocker · 06-17-2009, 09:01 AM

PM Sent re firmware!

and · 06-17-2009, 09:03 AM

Any noticeable changes with the 1.0 firmware? in terms of features that is, anything new in the menus that was not previously available.

yipchunyu · 06-17-2009, 09:12 AM

Originally posted by falconeye

Breaking news:

I've just received the final firmware release aka v1.00

Disclaimer: I have been informed that the firmware has been optimized for the sensor as used in series production K-7. The alpha test K-7 are equipped with a pre-series sensor (which was said to be 95% identical) and final final noise tests will have to await production cameras. Anyway, tomorrow I start to look into noise and resolution.

so, you get a pre-production dslr (95% identical to the production one) with a production firmware?

Vaikis_ · 06-17-2009, 09:48 AM

Originally posted by falconeye

Breaking news:

I've just received the final firmware release aka v1.00

Disclaimer: I have been informed that the firmware has been optimized for the sensor as used in series production K-7. The alpha test K-7 are equipped with a pre-series sensor (which was said to be 95% identical) and final final noise tests will have to await production cameras. Anyway, tomorrow I start to look into noise and resolution.

wow, waiting for tests

falconeye · 06-17-2009, 12:03 PM

Originally posted by Unlocker

PM Sent re firmware!

done

Originally posted by and

Any noticeable changes with the 1.0 firmware? in terms of features that is, anything new in the menus that was not previously available.

No, I haven't seen any changes from browsing through.

The contrast AF doesn't feel quicker and the removal of CAF during movie recording turned out to persist (read: no AF during movie capture with v1.00).

It is true probably that most of the work for v1.00 was in the noise department.

Originally posted by yipchunyu

so, you get a pre-production dslr (95% identical to the production one) with a production firmware?

Seems so

06-16-2009, 05:22 AM	#33
Class A Pentaxian Join Date: Aug 2008 Location: Wellington, New Zealand Posts: 11,251	Originally posted by falconeye It happens if you heavily sharpen a noise floor. I had this thought as well, but the structures are there before the sharpening already, just pretty faintly. I'd be surprised if they weren't a demosaicing artefact. Perhaps you can try another converter (I've read one can modify PEFs so that they can be read by existing converters)? Originally posted by jeffkrol From the very early reports of images looking soft but sharpening well points in the direction of a strong AA filter since this was an inherent quality in the D. Not sure I agree a strong AA filter is a plus. Unless I'm missing something you can emulate a stronger AA filter with PP blurring but you cannot (quite) undo the blurring effect of a strong AA. EDIT: The difference between a standard Canon 5D and one with the AA filter removed (scroll down to see the after/before images side by side) is quite impressive; like that between an outstanding lens and an OK lens. Last edited by Class A; 06-16-2009 at 05:44 AM.

06-16-2009, 05:55 AM	#36
jeffkrol Veteran Member Join Date: Sep 2006 Location: Wisconsin USA Photos: Gallery Posts: 8,434	Originally posted by Class A Not sure I agree a strong AA filter is a plus. Unless I'm missing something you can emulate a stronger AA filter with PP blurring but you cannot (quite) undo the blurring effect of a strong AA. EDIT: The difference between a standard Canon 5D and one with the AA filter removed (scroll down to see the after/before images side by side) is quite impressive; like that between an outstanding lens and an OK lens. I'll check your link but an AA filter is a necessity until we get pixel counts high enough to not need them (that could be around a 50+mp range)... Sigma doesn't use one and many complain about jaggies. It's a tradeoff and one to, I suppose, base your camera choice on. On one hand soft looking images, on the other jaggies and moire. High quality lenses will usually give you the best results w/ even strong AA filters.. Marginal lenses, on the other hand, go to mush.... as frustrating as a strong AA filter, it is more frustrating to get moire patterns in textiles (and much harder to deal with)...... Why Pentax shifted from the weak AA (at least in 1 direction) on the k10 is unknown but some pre-production images had some terrible moire in some blue jeans and other images had it in different things as well. EDIT: checking the Nikon photo the hot-rodded one shows moire everywhere (stop sign, AC on truck, AC on wall) and some of the detail, I suspect, is just artifacts from conversion, not true detail. Best case would be a removeable AA filter so we can choose (ala Kodak back "in the day") Screen capture added some artifacts as well as the now second conversion to jpg but there is plenty of detail and no moire and little "stepping", at least none objectionable: AT 200X Last edited by jeffkrol; 06-16-2009 at 06:27 AM.

06-17-2009, 08:50 AM	#40
falconeye Veteran Member Join Date: Jan 2008 Location: Munich, Alps, Germany Photos: Gallery Posts: 6,871 Original Poster	K-7 final firmware released Breaking news: I've just received the final firmware release aka v1.00 Disclaimer: I have been informed that the firmware has been optimized for the sensor as used in series production K-7. The alpha test K-7 are equipped with a pre-series sensor (which was said to be 95% identical) and final final noise tests will have to await production cameras. Anyway, tomorrow I start to look into noise and resolution.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Sony Alpha 900	dylansalt	Canon, Nikon, Sony, and Other Camera Brands	18	10-27-2010 11:01 AM
Any field test report for the kr?	opiedog	Pentax DSLR Discussion	1	09-15-2010 11:47 AM
rawtherapee v3.0 alpha 1 is out	deejjjaaaa	Digital Processing, Software, and Printing	14	01-07-2010 09:41 PM
sigma 35-135mm alpha IV	pdxbmw	Pentax SLR Lens Discussion	3	12-29-2009 01:47 PM
Sony Alpha dSLR-A200	deejjjaaaa	Pentax News and Rumors	52	01-09-2008 03:30 PM

06-16-2009, 04:38 AM	#31
jeffkrol Veteran Member Join Date: Sep 2006 Location: Wisconsin USA Photos: Gallery Posts: 8,434	Quanitity of images w/ moire will determine how weak or strong the AA filter is. Resolution tests as well. The weakest AA Pentax seems to be the K10, strongest ist-D... From the very early reports of images looking soft but sharpening well points in the direction of a strong AA filter since this was an inherent quality in the D. Wormy patterns or maze artifacts are more software related (Bayer demosaicing related and jpg construction) then mechanical and are not particularly worrisome. Unlockers (thanks btw) parrot shot also seems to point to a strong AA filter. Last edited by jeffkrol; 06-16-2009 at 04:44 AM.*

06-16-2009, 05:10 AM	#32
RBellavance Veteran Member Join Date: Nov 2006 Location: Near Montréal, Canada Photos: Gallery Posts: 1,716	Originally posted by Class A The minute I saw it I thought this is an unnecessary complication. I was amazed to see many report that they often found their dial to be in a different position after having stored the camera in the bag. This has never happened to me once and frankly I cannot imagine how you'd have to handle a camera for this to happen. I feel the same. It seems to me like a solution in search of a problem. I cannot recall a single instance of the mode dial being moved unintentionally over the 2.5 years I've had my K10D.

06-16-2009, 05:30 AM	#34
Unlocker Senior Member Join Date: Jan 2008 Location: UK Posts: 169	That's all going a bit too in depth for my liking! Just posted some high ISO shots taken on Sunday, and won't post any crops until / if I get final firmware before hand back, but thought you might like to dissect them!

06-16-2009, 05:52 AM	#35
RBellavance Veteran Member Join Date: Nov 2006 Location: Near Montréal, Canada Photos: Gallery Posts: 1,716	Originally posted by Class A Not sure I agree a strong AA filter is a plus. Unless I'm missing something you can emulate a stronger AA filter with PP blurring but you cannot (quite) undo the blurring effect of a strong AA. I don't think you're missing anything. It is definitely not a plus in my book, as it lowers the effective resolution.

06-16-2009, 06:26 AM	#37
falconeye Veteran Member Join Date: Jan 2008 Location: Munich, Alps, Germany Photos: Gallery Posts: 6,871 Original Poster	Originally posted by Class A EDIT: The difference between a standard Canon 5D and one with the AA filter removed (scroll down to see the after/before images side by side) is quite impressive; like that between an outstanding lens and an OK lens. Look at the Area 3 crop, the shadow boundaries below the radiator. See the purple fringe to the left and green fringe to the right? False colors. Can be seen in K-7 videos at max. resolution as well and just some find it disturbing ... Also, the hot rod image top in that page looks jaggy, something some didn't like in K-7 max.resolution video samples... Also, in a side by side comparison, you have to sharpen the AA sample because it is meant to be sharpened. IMHO, there is an optimum AA filter strength, not too weak so enough color information is recorded and not too strong to maintain resolution. Last edited by falconeye; 06-16-2009 at 06:31 AM.

06-16-2009, 06:40 AM	#38
Class A Pentaxian Join Date: Aug 2008 Location: Wellington, New Zealand Posts: 11,251	Originally posted by RBellavance I don't think you're missing anything. Actually, I thought about it again (using an audio analogy) and you cannot faithfully emulate an AA filter with software. In general, you may get artefacts that will be very difficult to remove without losing more resolution than the AA filter would have lost you. As I don't shoot fashion as a professional I'd prefer a rather weak AA filter but none at all would probably be too limiting in practise.

06-16-2009, 06:49 AM	#39
Class A Pentaxian Join Date: Aug 2008 Location: Wellington, New Zealand Posts: 11,251	Originally posted by falconeye See the purple fringe to the left and green fringe to the right? False colors. Yes, could be chroma moire, but the shot to right also has a hint of it and a lot more of it on the vertical lines surrounding the radiator. Perhaps the colour artefacts are caused by lens CA? Or the AA filter of the 5D should be slightly stronger to completely avoid colour moire? Area 5 is a good example in favour of little/no AA. But you are right, the image to the right should be capture-sharpened (and the image to the left treated for luminance and color moire?). Anyhow. Originally posted by falconeye IMHO, there is an optimum AA filter strength, not too weak so enough color information is recorded and not too strong to maintain resolution. I agree.

06-17-2009, 09:01 AM	#41
Unlocker Senior Member Join Date: Jan 2008 Location: UK Posts: 169	PM Sent re firmware!

06-17-2009, 09:03 AM	#42
and Veteran Member Join Date: Mar 2007 Photos: Gallery Posts: 2,476	Any noticeable changes with the 1.0 firmware? in terms of features that is, anything new in the menus that was not previously available.

06-17-2009, 09:12 AM	#43
yipchunyu Senior Member Join Date: Mar 2007 Location: HK Posts: 171	Originally posted by falconeye Breaking news: I've just received the final firmware release aka v1.00 Disclaimer: I have been informed that the firmware has been optimized for the sensor as used in series production K-7. The alpha test K-7 are equipped with a pre-series sensor (which was said to be 95% identical) and final final noise tests will have to await production cameras. Anyway, tomorrow I start to look into noise and resolution. so, you get a pre-production dslr (95% identical to the production one) with a production firmware?

Thread Tools	Search this Thread
	Search this Thread: Advanced Search

06-17-2009, 09:48 AM	#44
Vaikis_ Senior Member Join Date: Oct 2008 Posts: 213	Originally posted by falconeye Breaking news: I've just received the final firmware release aka v1.00 Disclaimer: I have been informed that the firmware has been optimized for the sensor as used in series production K-7. The alpha test K-7 are equipped with a pre-series sensor (which was said to be 95% identical) and final final noise tests will have to await production cameras. Anyway, tomorrow I start to look into noise and resolution. wow, waiting for tests

06-17-2009, 12:03 PM	#45
falconeye Veteran Member Join Date: Jan 2008 Location: Munich, Alps, Germany Photos: Gallery Posts: 6,871 Original Poster	Originally posted by Unlocker PM Sent re firmware! done Originally posted by and Any noticeable changes with the 1.0 firmware? in terms of features that is, anything new in the menus that was not previously available. No, I haven't seen any changes from browsing through. The contrast AF doesn't feel quicker and the removal of CAF during movie recording turned out to persist (read: no AF during movie capture with v1.00). It is true probably that most of the work for v1.00 was in the noise department. Originally posted by yipchunyu so, you get a pre-production dslr (95% identical to the production one) with a production firmware? Seems so