DC Watch interview with K-1 designers

JPT · 03-21-2016, 11:13 PM - **1 Like**

DC Watch has posted a lengthy interview with the K-1 designers. It is in Japanese, so you'll need to see what you can understand through a translator. I don't have time to look at it now, but it seems to include a lot of geeky detail.

http://m.dc.watch.impress.co.jp/docs/news/interview_dcm/20160322_748845.html

Not a Number · 03-22-2016, 12:55 AM

Thanks, even with Google translate you can pull out some interesting items.

Such as the initial study group for the Full Frame concept was formed spring 2012 and the actual design teams started about a year later in 2013. One factoid I ran across on USB3 was that by 3rd quarter of 2013 only an estimated 1/4 of devices in use were equipped with USB3.

The K3ii has motion correction for pixel shift and is always on. Motion Correction doesn't work well with details like leaves in trees because it is difficult to detect motion.

The shutter is turned 180° in respect to those on APS-C bodies and had to be larger than the Full Frame to accommodate SR. SR should still be turned off for tripod use but SR and macro?

Kunzite · 03-22-2016, 01:16 AM

Thank you, very interesting.

**jbinpg** · 03-22-2016, 01:55 AM

Wow, a lot of info there. Thanks for the link!

Zen4Life · 03-22-2016, 08:39 AM

Originally posted by Not a Number

The K3ii has motion correction for pixel shift and is always on. Motion Correction doesn't work well with details like leaves in trees because it is difficult to detect motion.

That is also what I found testing out this feature recently. That said, do we have a feeling whether the K1 motion correction is an "improved" version of the K-3 II one? Or is it the same?

Tan68 · 03-25-2016, 11:56 AM

Zen4Life, perhaps better motion detection than with K3.

I watched a CP+ video in Japanese and posted some notes on what I thought I could understand here: CP+ video illustrates pixel shift & moving subject. I think: Pentax SLR Talk Forum: Digital Photography Review

JPT · 03-25-2016, 05:39 PM - **1 Like**

Not having ever tried the RRS function on the K-3 II, I can't compare how it works. But I did have a chance to test it at CP+

I can tell you what Kimio Tanaka is saying about the use of RRS and motion correction in that video.
- He believes the RRS function makes the image quality equivalent to 70mp
- You need to use a good lens to get the benefit
- He thinks the name "Real Resolution System" is underselling the function
- You should always use Raw + JPEG
- You should switch off motion correction if you are sure the image is still to avoid the system making any mistakes
- For long exposures like the water shot, RRS works well, but if the exposure time is too short, you can see artefacts

When I tried it, the Ricoh rep encouraged me to shoot right into the moving crowd with the motion correction on and a shutter speed of 1/200. This must be a real torture test for the system, but for the most part, it worked well. What I noticed was that the high contrast edges were easily detected and corrected for. For examples, people's heads against the background were correctly identified and image from the first frame applied to that area. On the other hand, low contrast areas, such as the folds in people's clothes, were not fully detected and there were some artefacts when you zoomed in. To indicate how effective the motion correction function is, the rep then made me take the same image with RRS on but motion correction off, and the result was a complete multi-colour mess. It definitely works well within it's limits, but it will take a fair amount of experimentation by someone to find out what those limits are.

All the pictures I took were handheld, which I'm sure is not advisable, but it was feasible because the sensor shift can still do its SR function while doing RRS. I'm not sure if that was the case for the K-3 II, but the Ricoh rep told me that it is true for the K-1, and my use of it seemed to confirm that. The SR on the K-1 seems to be very advanced.

Tan68 · 03-26-2016, 06:32 AM

Thanks for the notes. For the movement correction system to work well, sounds like more movement and more contrast is helpful and chances of success are greater. That makes sense.

Was there a reason given for always using RAW+JPG? Perhaps the recommendation is in case there is an unexpected result with the RAW, the JPG will be there to use..?

For multiple exposure, it is possible for all images to be compiled into one RAW that can be edited with any image editor.

I would like Pentax to develop a computer utility to combine the RRS files into one RAW (rather than export TIF from PDU or PCU..?) and then this one RAW could be edited in any image editor. This way, not all image editors would have to support the RRS files and it would be possible for the user to take advantage of RRS and editor of choice.. They could even allow masking of areas like the clothing in your samples. If that is too complicated, at least allowing the files to be compiled to a RAW would be fine enough :^)

+ ...maybe the idea of a utility is no good. maybe once all the RRS files are compiled any advantage will be lost... perhaps a RAW editor needs access to all the RRS files while the image is being manipulated. then the RAW editor can output to TIF the info you chose to use from the RRS files. ?

Originally posted by JPT

Not having ever tried the RRS function on the K-3 II, I can't compare how it works. But I did have a chance to test it at CP+

...
- For long exposures like the water shot, RRS works well, but if the exposure time is too short, you can see artefacts

...
For examples, people's heads against the background were correctly identified and image from the first frame applied to that area. On the other hand, low contrast areas, such as the folds in people's clothes, were not fully detected and there were some artefacts when you zoomed in. ...

Gimbal · 03-26-2016, 06:56 AM

Originally posted by Tan68

I would like Pentax to develop a computer utility to combine the RRS files into one RAW (rather than export TIF from PDU or PCU..?) and then this one RAW could be edited in any image editor. This way, not all image editors would have to support the RRS files and it would be possible for the user to take advantage of RRS and editor of choice.. They could even allow masking of areas like the clothing in your samples. If that is too complicated, at least allowing the files to be compiled to a RAW would be fine enough :^)

+ ...maybe the idea of a utility is no good. maybe once all the RRS files are compiled any advantage will be lost... perhaps a RAW editor needs access to all the RRS files while the image is being manipulated. then the RAW editor can output to TIF the info you chose to use from the RRS files. ?

I don't think it would work to combine the RRS file to a raw, since a RRS picture has all the RGB values ready, while a normal RAW file doesn't. So a RAW converter would be confused if you feed it a RAW file with all values already done.

A compressed 16bit TIFF file would be great though. Perhaps the K-1 already can do that? I don't know.

03-21-2016, 11:13 PM - 1 Like	#1
JPT Pentaxian Join Date: Jan 2012 Location: Tokyo Photos: Gallery Posts: 2,821	DC Watch interview with K-1 designers DC Watch has posted a lengthy interview with the K-1 designers. It is in Japanese, so you'll need to see what you can understand through a translator. I don't have time to look at it now, but it seems to include a lot of geeky detail. http://m.dc.watch.impress.co.jp/docs/news/interview_dcm/20160322_748845.html

03-26-2016, 06:32 AM	#8
Tan68 Veteran Member Join Date: Sep 2014 Photos: Albums Posts: 501	Thanks for the notes. For the movement correction system to work well, sounds like more movement and more contrast is helpful and chances of success are greater. That makes sense. Was there a reason given for always using RAW+JPG? Perhaps the recommendation is in case there is an unexpected result with the RAW, the JPG will be there to use..? For multiple exposure, it is possible for all images to be compiled into one RAW that can be edited with any image editor. I would like Pentax to develop a computer utility to combine the RRS files into one RAW (rather than export TIF from PDU or PCU..?) and then this one RAW could be edited in any image editor. This way, not all image editors would have to support the RRS files and it would be possible for the user to take advantage of RRS and editor of choice.. They could even allow masking of areas like the clothing in your samples. If that is too complicated, at least allowing the files to be compiled to a RAW would be fine enough :^) + ...maybe the idea of a utility is no good. maybe once all the RRS files are compiled any advantage will be lost... perhaps a RAW editor needs access to all the RRS files while the image is being manipulated. then the RAW editor can output to TIF the info you chose to use from the RRS files. ? Originally posted by JPT Not having ever tried the RRS function on the K-3 II, I can't compare how it works. But I did have a chance to test it at CP+ ... - For long exposures like the water shot, RRS works well, but if the exposure time is too short, you can see artefacts ... For examples, people's heads against the background were correctly identified and image from the first frame applied to that area. On the other hand, low contrast areas, such as the folds in people's clothes, were not fully detected and there were some artefacts when you zoomed in. ... Last edited by Tan68; 03-26-2016 at 06:38 AM.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
PF video interview with Ricoh - K-1 has no EFC shutter??	2351HD	Pentax Full Frame	20	02-28-2016 02:38 PM
Ricoh Imaging Interview on DC Watch website	JPT	Pentax News and Rumors	242	01-18-2015 11:57 AM
The designers of K-S1	Unregistered User	Pentax DSLR Discussion	65	09-02-2014 11:54 AM
DC Watch Interview with Q7 Designers	JPT	Pentax Q	18	07-25-2013 01:01 PM
DC Watch: Interview with Toshiyuki Kitazawa, Pentax's Head of Business Development	Mistral75	Pentax News and Rumors	181	05-04-2013 12:59 AM

03-22-2016, 12:55 AM	#2
Not a Number Moderator Join Date: Mar 2012 Location: Venice, CA Posts: 10,526	Thanks, even with Google translate you can pull out some interesting items. Such as the initial study group for the Full Frame concept was formed spring 2012 and the actual design teams started about a year later in 2013. One factoid I ran across on USB3 was that by 3rd quarter of 2013 only an estimated 1/4 of devices in use were equipped with USB3. The K3ii has motion correction for pixel shift and is always on. Motion Correction doesn't work well with details like leaves in trees because it is difficult to detect motion. The shutter is turned 180° in respect to those on APS-C bodies and had to be larger than the Full Frame to accommodate SR. SR should still be turned off for tripod use but SR and macro?

03-22-2016, 01:16 AM	#3
Kunzite Pentaxian Join Date: Apr 2007 Location: Romania Posts: 15,132	Thank you, very interesting.

03-22-2016, 01:55 AM	#4
jbinpg Site Supporter Join Date: Mar 2008 Location: Prince George, BC Photos: Gallery \| Albums Posts: 3,546	Wow, a lot of info there. Thanks for the link!

03-22-2016, 08:39 AM	#5
Zen4Life Forum Member Join Date: Feb 2012 Posts: 63	Originally posted by Not a Number The K3ii has motion correction for pixel shift and is always on. Motion Correction doesn't work well with details like leaves in trees because it is difficult to detect motion. That is also what I found testing out this feature recently. That said, do we have a feeling whether the K1 motion correction is an "improved" version of the K-3 II one? Or is it the same?

03-25-2016, 11:56 AM	#6
Tan68 Veteran Member Join Date: Sep 2014 Photos: Albums Posts: 501	Zen4Life, perhaps better motion detection than with K3. I watched a CP+ video in Japanese and posted some notes on what I thought I could understand here: CP+ video illustrates pixel shift & moving subject. I think: Pentax SLR Talk Forum: Digital Photography Review

03-25-2016, 05:39 PM - 1 Like	#7
JPT Pentaxian Join Date: Jan 2012 Location: Tokyo Photos: Gallery Posts: 2,821 Original Poster	Not having ever tried the RRS function on the K-3 II, I can't compare how it works. But I did have a chance to test it at CP+ I can tell you what Kimio Tanaka is saying about the use of RRS and motion correction in that video. - He believes the RRS function makes the image quality equivalent to 70mp - You need to use a good lens to get the benefit - He thinks the name "Real Resolution System" is underselling the function - You should always use Raw + JPEG - You should switch off motion correction if you are sure the image is still to avoid the system making any mistakes - For long exposures like the water shot, RRS works well, but if the exposure time is too short, you can see artefacts When I tried it, the Ricoh rep encouraged me to shoot right into the moving crowd with the motion correction on and a shutter speed of 1/200. This must be a real torture test for the system, but for the most part, it worked well. What I noticed was that the high contrast edges were easily detected and corrected for. For examples, people's heads against the background were correctly identified and image from the first frame applied to that area. On the other hand, low contrast areas, such as the folds in people's clothes, were not fully detected and there were some artefacts when you zoomed in. To indicate how effective the motion correction function is, the rep then made me take the same image with RRS on but motion correction off, and the result was a complete multi-colour mess. It definitely works well within it's limits, but it will take a fair amount of experimentation by someone to find out what those limits are. All the pictures I took were handheld, which I'm sure is not advisable, but it was feasible because the sensor shift can still do its SR function while doing RRS. I'm not sure if that was the case for the K-3 II, but the Ricoh rep told me that it is true for the K-1, and my use of it seemed to confirm that. The SR on the K-1 seems to be very advanced.

03-26-2016, 06:56 AM	#9
Gimbal Pentaxian Join Date: Apr 2007 Location: Sweden Posts: 2,106	Originally posted by Tan68 I would like Pentax to develop a computer utility to combine the RRS files into one RAW (rather than export TIF from PDU or PCU..?) and then this one RAW could be edited in any image editor. This way, not all image editors would have to support the RRS files and it would be possible for the user to take advantage of RRS and editor of choice.. They could even allow masking of areas like the clothing in your samples. If that is too complicated, at least allowing the files to be compiled to a RAW would be fine enough :^) + ...maybe the idea of a utility is no good. maybe once all the RRS files are compiled any advantage will be lost... perhaps a RAW editor needs access to all the RRS files while the image is being manipulated. then the RAW editor can output to TIF the info you chose to use from the RRS files. ? I don't think it would work to combine the RRS file to a raw, since a RRS picture has all the RGB values ready, while a normal RAW file doesn't. So a RAW converter would be confused if you feed it a RAW file with all values already done. A compressed 16bit TIFF file would be great though. Perhaps the K-1 already can do that? I don't know.