Have Pentax IBIS capabilities improved with successive generations?

Fogel70 · 10-02-2020, 01:43 AM - **1 Like**

SR has improved over the years, but cameras with higher pixel pitch also require more efficient image stabilization, if you want to be able to use the higher resolution.

The standard for measuring image stabilization has changed over time. I believe it was up until K3 all camera manufacturers had their own standard for testing the efficiency. So it could vary between manufacturers.
Later is had been a CIPA standard that all manufacturers have been using for image stabilization test. But is it a bit limited as the CIPA standard only tested with 2 axis of motion. It does not test in circumstances where 3- or 5-axis stabilization offer an advantage.

So the specification on image stabilization efficiency (number of stops) is not fully comparable on older and newer cameras. And it may be a difference in image stabilization between cameras or lenses even when the specified number of stops is the same. In general after adopting the CIPA standard most manufactures had an increase of 0.5 -1.0 stop in efficiency and Canon and Nikon change specification on their lenses. But I do not think camera manufacturers changed the specification on cameras no longer in production.

Last edited by Fogel70; 10-02-2020 at 02:00 AM.

**savoche** · 10-02-2020, 01:55 AM

Originally posted by swanlefitte

It won't get better than 6.3 stops. That is unless you can slow the earth rotation.
Earth’s Rotation Limits IBIS Performance to 6.3 Stops – The Center Column

That's for a 24 MP full frame sensor. If I got my numbers right the limit would be about 5.8 for a 24 MP APS-c sensor using the same formula (and 6.05 for a 36 MP full frame sensor).

But, as @Breakfastographer says, if the camera knows its location and which direction it is pointing it should be possible to compensate for that rotation. I have no idea if the Astrotracer would be sufficiently accurate for the task.

So now we know what we want from the K-new, right?

Fogel70 · 10-02-2020, 04:49 AM - **1 Like**

BTW Canon claim 8 stops of image stabilization on R6 and R5 when IBIS is used together with an optical stabilized lens, or with the 28-70/2 lens without optical stabilization.

swanlefitte · 10-02-2020, 07:27 AM

Originally posted by Fogel70

BTW Canon claim 8 stops of image stabilization on R6 and R5 when IBIS is used together with an optical stabilized lens, or with the 28-70/2 lens without optical stabilization.

I wonder how. My guess is by defining it as acceptable image quality.

"So if the image circle is really big, if the image quality on it is really good edge-to-edge, you're allowing the sensor to move around a lot, which will give you the image stabilization. And obviously that will vary depending on focal lengths and all the rest of it as well."
5 lenses that get 8 stops of stabilization on the Canon EOS R5 and R6 | Digital Camera World

It sounds like they are saying a great lens with 8 stops will still be as good as a good lens with 8 stops more light. Or at 1/1000 you might get 2500 LW/PH and at 1/4 you can still get 2200 LW/PH or something like that.

vonBaloney · 10-02-2020, 07:43 AM

Originally posted by Breakfastographer

Huh? To get beyond 6.3 stops, all you need is 2D electronic balance, compass, and GPS. Neural networks have nothing to do with it. "Neural networks" is not a buzzword for "software":
Artificial neural network - Wikipedia

I didn't say it was a buzzword. But neural networks are software. i.e. you can clean up shaky images after the fact much better than a few years ago with neural networks (among other things that the camera could also do like saving the SR sensor data for later processing). So I'm saying software will continue to allow improvement in shake reduction even beyond a hardware limit that might be imposed by the rotation of the earth or whatever.

Breakfastographer · 10-02-2020, 08:27 AM

Originally posted by vonBaloney

I didn't say it was a buzzword. But neural networks are software. i.e. you can clean up shaky images after the fact much better than a few years ago with neural networks (among other things that the camera could also do like saving the SR sensor data for later processing). So I'm saying software will continue to allow improvement in shake reduction even beyond a hardware limit that might be imposed by the rotation of the earth or whatever.

I looked at that functionality in Cyberlink PhotoDirector 11 and arrived at the opposite of your conclusion.

Classic deconvolution is still best for fixing blur in my experience.

Fogel70 · 10-02-2020, 08:30 AM

Originally posted by swanlefitte

I wonder how. My guess is by defining it as acceptable image quality.

"So if the image circle is really big, if the image quality on it is really good edge-to-edge, you're allowing the sensor to move around a lot, which will give you the image stabilization. And obviously that will vary depending on focal lengths and all the rest of it as well."
5 lenses that get 8 stops of stabilization on the Canon EOS R5 and R6 | Digital Camera World

It sounds like they are saying a great lens with 8 stops will still be as good as a good lens with 8 stops more light. Or at 1/1000 you might get 2500 LW/PH and at 1/4 you can still get 2200 LW/PH or something like that.

It may also be because of other new more efficient tech. FI latest motion sensors with faster response and higher accuracy. Automotive and mobile industry has put high demand on this type och tech that camera manufacturers can benefit from. These thing may also be easier to implement more efficiently on a newly designe camera system that most likely have more computing power and faster internal communiccation protocols.

Pentax DSLRs may suffer a bit from having a system where the electronic design origins from 1980s and 1990s, where there may be bottlenecks in the system for keeping backward compatibility with the original design.

Breakfastographer · 10-02-2020, 08:43 AM

Originally posted by Fogel70

BTW Canon claim 8 stops of image stabilization on R6 and R5 when IBIS is used together with an optical stabilized lens, or with the 28-70/2 lens without optical stabilization.

As the link from @swanlefitte (does that work?) or my earlier compilation shows, only five lenses are claimed by Canon to provide 8 stops of stabilisation on R5/R6 cameras.

vonBaloney · 10-02-2020, 08:49 AM

Originally posted by Breakfastographer

I looked at that functionality in Cyberlink PhotoDirector 11 and arrived at the opposite of your conclusion.

Classic deconvolution is still best for fixing blur in my experience.

I've only used Topaz. Sometimes it is excellent, sometimes it doesn't work out. In any case, the technology is definitely improving and deep learning has made major strides in the last few years in general.

**TedH42** · 10-02-2020, 09:38 AM - **2 Likes**

Originally posted by ALLCAPS

Happy 6 year k5iis user here. One of the things that drew me to the brand was the in body stabilization. I have "nervous hands," and my recent freehand use of longer focal lengths has been a challenge.

For those of you who have used the older cameras as well as the latest generations of Pentax gear, have you noticed improvements in the effectiveness of the IBIS systems? A more effective IBIS system and better AF in the k-new would almost clinch a purchase from me.

I have had K100D, K200D, K-5, K-3, and KP (also K-1ii, which has similar IBIS to KP). The IBIS has continuously, and quite noticeably, improved through this series. When I got the KP, and using it with lenses up to 300mm (the DA 55-300 PLM is a favorite), I just got the feeling that I could throw away my tripod (I actually didn't, but the additional freedom from the tripod for lower shutter speeds was wonderful).

In addition to better IBIS, the KP also maintains better noise characteristics to higher ISO values, which permits higher shutter speeds, also reducing the effect of shake.

Pentax IBIS technology has steadily, and significantly improved over camera generations. My personal expectation is that K-new IBIS will be at least better than KP/K-1 level, which far better than K-5 capability. And it will have better higher ISO noise characteristics than the K-5iis, so your shake experience should be noticeably better than what you have currently. But it won't be cheap.

que es tu · 10-03-2020, 12:09 PM

Slow the earth's rotation, neural networks, classic deconvolution?? I'm not even sure if you all are even being serious or just pulling my leg!!

ALLCAPS · 10-03-2020, 05:48 PM

Originally posted by que es tu

Slow the earth's rotation, neural networks, classic deconvolution??

Or google searches the FBI uncovers when searching a mad scientist villain's computer........

bvg · 11-11-2020, 03:26 AM

Originally posted by Fogel70

BTW Canon claim 8 stops of image stabilization on R6 and R5 when IBIS is used together with an optical stabilized lens, or with the 28-70/2 lens without optical stabilization.

It seems Panasonic S1H beats Canon R5:

Youtu.be

yucafrita · 11-11-2020, 03:34 AM

I wonder if this IBIS race is because the photographers are getting older, shakier hands or try to shoot drunken. If you imagine that Pentax was one of the first and the industry as a whole denied the necessity of it (because it's in the lens). Now everyone is raving about it. Strange new world.

**Rondec** · 11-11-2020, 03:42 AM

The whole 8 stops of IBIS seems over inflated. If you are shooting a 24-70 lens at 40mm, it feels like even with stabilization that is great, it would be a struggle to hand hold at 6 or 7 seconds, which is what that would be. My experience with Pentax is that IBIS works well till you get to shutter speeds that are slower than 1/2 second and then it doesn't matter what the focal length is, I can't really hand hold it. So, I can hand hold 60mm at 1/6 second and get sharp images, but at 15mm I can only hand hold to about 1/4 second before shake starts to show up.

10-02-2020, 04:49 AM - 1 Like	#18
Fogel70 Pentaxian Join Date: Sep 2009 Location: Stockholm, Sweden Photos: Albums Posts: 3,062	BTW Canon claim 8 stops of image stabilization on R6 and R5 when IBIS is used together with an optical stabilized lens, or with the 28-70/2 lens without optical stabilization. Last edited by Fogel70; 10-02-2020 at 04:57 AM.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Question about properly using Pentax IBIS - has it improved ?	aslyfox	Troubleshooting and Beginner Help	17	02-22-2020 02:56 PM
About IBIS - Which company started IBIS in DSLR's ? Curiosity question.	jpzk	Pentax DSLR Discussion	30	06-13-2019 04:56 PM
How could I have improved this?	Zephos	Photographic Technique	19	02-11-2015 09:12 AM
Nature Ibis, Ibis, everywhere	photolady95	Post Your Photos!	9	05-11-2011 08:38 AM
Any bags that have a drying capabilities?	Pentaxtic	Pentax Camera and Field Accessories	7	05-18-2009 04:52 PM

10-02-2020, 01:43 AM - 1 Like	#16
Fogel70 Pentaxian Join Date: Sep 2009 Location: Stockholm, Sweden Photos: Albums Posts: 3,062	SR has improved over the years, but cameras with higher pixel pitch also require more efficient image stabilization, if you want to be able to use the higher resolution. The standard for measuring image stabilization has changed over time. I believe it was up until K3 all camera manufacturers had their own standard for testing the efficiency. So it could vary between manufacturers. Later is had been a CIPA standard that all manufacturers have been using for image stabilization test. But is it a bit limited as the CIPA standard only tested with 2 axis of motion. It does not test in circumstances where 3- or 5-axis stabilization offer an advantage. So the specification on image stabilization efficiency (number of stops) is not fully comparable on older and newer cameras. And it may be a difference in image stabilization between cameras or lenses even when the specified number of stops is the same. In general after adopting the CIPA standard most manufactures had an increase of 0.5 -1.0 stop in efficiency and Canon and Nikon change specification on their lenses. But I do not think camera manufacturers changed the specification on cameras no longer in production. Last edited by Fogel70; 10-02-2020 at 02:00 AM.

10-02-2020, 01:55 AM	#17
savoche Master of the obvious Join Date: Jun 2011 Location: Lowlands of Norway Posts: 18,311	Originally posted by swanlefitte It won't get better than 6.3 stops. That is unless you can slow the earth rotation. Earth’s Rotation Limits IBIS Performance to 6.3 Stops – The Center Column That's for a 24 MP full frame sensor. If I got my numbers right the limit would be about 5.8 for a 24 MP APS-c sensor using the same formula (and 6.05 for a 36 MP full frame sensor). But, as @Breakfastographer says, if the camera knows its location and which direction it is pointing it should be possible to compensate for that rotation. I have no idea if the Astrotracer would be sufficiently accurate for the task. So now we know what we want from the K-new, right?

10-02-2020, 07:27 AM	#19
swanlefitte Pentaxian Join Date: May 2015 Location: Minneapolis Photos: Gallery Posts: 4,068	Originally posted by Fogel70 BTW Canon claim 8 stops of image stabilization on R6 and R5 when IBIS is used together with an optical stabilized lens, or with the 28-70/2 lens without optical stabilization. I wonder how. My guess is by defining it as acceptable image quality. "So if the image circle is really big, if the image quality on it is really good edge-to-edge, you're allowing the sensor to move around a lot, which will give you the image stabilization. And obviously that will vary depending on focal lengths and all the rest of it as well." 5 lenses that get 8 stops of stabilization on the Canon EOS R5 and R6 \| Digital Camera World It sounds like they are saying a great lens with 8 stops will still be as good as a good lens with 8 stops more light. Or at 1/1000 you might get 2500 LW/PH and at 1/4 you can still get 2200 LW/PH or something like that.

10-02-2020, 07:43 AM	#20
vonBaloney Pentaxian Join Date: Oct 2011 Location: Albuquerque, NM Posts: 6,029	Originally posted by Breakfastographer Huh? To get beyond 6.3 stops, all you need is 2D electronic balance, compass, and GPS. Neural networks have nothing to do with it. "Neural networks" is not a buzzword for "software": Artificial neural network - Wikipedia I didn't say it was a buzzword. But neural networks are software. i.e. you can clean up shaky images after the fact much better than a few years ago with neural networks (among other things that the camera could also do like saving the SR sensor data for later processing). So I'm saying software will continue to allow improvement in shake reduction even beyond a hardware limit that might be imposed by the rotation of the earth or whatever.

10-02-2020, 08:27 AM	#21
Breakfastographer Pentaxian Join Date: Nov 2018 Posts: 673	Originally posted by vonBaloney I didn't say it was a buzzword. But neural networks are software. i.e. you can clean up shaky images after the fact much better than a few years ago with neural networks (among other things that the camera could also do like saving the SR sensor data for later processing). So I'm saying software will continue to allow improvement in shake reduction even beyond a hardware limit that might be imposed by the rotation of the earth or whatever. I looked at that functionality in Cyberlink PhotoDirector 11 and arrived at the opposite of your conclusion. Classic deconvolution is still best for fixing blur in my experience.

10-02-2020, 08:30 AM	#22
Fogel70 Pentaxian Join Date: Sep 2009 Location: Stockholm, Sweden Photos: Albums Posts: 3,062	Originally posted by swanlefitte I wonder how. My guess is by defining it as acceptable image quality. "So if the image circle is really big, if the image quality on it is really good edge-to-edge, you're allowing the sensor to move around a lot, which will give you the image stabilization. And obviously that will vary depending on focal lengths and all the rest of it as well." 5 lenses that get 8 stops of stabilization on the Canon EOS R5 and R6 \| Digital Camera World It sounds like they are saying a great lens with 8 stops will still be as good as a good lens with 8 stops more light. Or at 1/1000 you might get 2500 LW/PH and at 1/4 you can still get 2200 LW/PH or something like that. It may also be because of other new more efficient tech. FI latest motion sensors with faster response and higher accuracy. Automotive and mobile industry has put high demand on this type och tech that camera manufacturers can benefit from. These thing may also be easier to implement more efficiently on a newly designe camera system that most likely have more computing power and faster internal communiccation protocols. Pentax DSLRs may suffer a bit from having a system where the electronic design origins from 1980s and 1990s, where there may be bottlenecks in the system for keeping backward compatibility with the original design.

10-02-2020, 08:43 AM	#23
Breakfastographer Pentaxian Join Date: Nov 2018 Posts: 673	Originally posted by Fogel70 BTW Canon claim 8 stops of image stabilization on R6 and R5 when IBIS is used together with an optical stabilized lens, or with the 28-70/2 lens without optical stabilization. As the link from @swanlefitte (does that work?) or my earlier compilation shows, only five lenses are claimed by Canon to provide 8 stops of stabilisation on R5/R6 cameras.

10-02-2020, 08:49 AM	#24
vonBaloney Pentaxian Join Date: Oct 2011 Location: Albuquerque, NM Posts: 6,029	Originally posted by Breakfastographer I looked at that functionality in Cyberlink PhotoDirector 11 and arrived at the opposite of your conclusion. Classic deconvolution is still best for fixing blur in my experience. I've only used Topaz. Sometimes it is excellent, sometimes it doesn't work out. In any case, the technology is definitely improving and deep learning has made major strides in the last few years in general.

10-02-2020, 09:38 AM - 2 Likes	#25
TedH42 F/8 & Somewhere Join Date: Dec 2011 Location: Colorado Photos: Gallery Posts: 1,412	Originally posted by ALLCAPS Happy 6 year k5iis user here. One of the things that drew me to the brand was the in body stabilization. I have "nervous hands," and my recent freehand use of longer focal lengths has been a challenge. For those of you who have used the older cameras as well as the latest generations of Pentax gear, have you noticed improvements in the effectiveness of the IBIS systems? A more effective IBIS system and better AF in the k-new would almost clinch a purchase from me. I have had K100D, K200D, K-5, K-3, and KP (also K-1ii, which has similar IBIS to KP). The IBIS has continuously, and quite noticeably, improved through this series. When I got the KP, and using it with lenses up to 300mm (the DA 55-300 PLM is a favorite), I just got the feeling that I could throw away my tripod (I actually didn't, but the additional freedom from the tripod for lower shutter speeds was wonderful). In addition to better IBIS, the KP also maintains better noise characteristics to higher ISO values, which permits higher shutter speeds, also reducing the effect of shake. Pentax IBIS technology has steadily, and significantly improved over camera generations. My personal expectation is that K-new IBIS will be at least better than KP/K-1 level, which far better than K-5 capability. And it will have better higher ISO noise characteristics than the K-5iis, so your shake experience should be noticeably better than what you have currently. But it won't be cheap.

10-03-2020, 12:09 PM	#26
que es tu Pentaxian Join Date: Feb 2020 Location: Smoky Mountains, NC Photos: Gallery \| Albums Posts: 1,602	Slow the earth's rotation, neural networks, classic deconvolution?? I'm not even sure if you all are even being serious or just pulling my leg!!

10-03-2020, 05:48 PM	#27
ALLCAPS Forum Member Join Date: May 2018 Posts: 52 Original Poster	Originally posted by que es tu Slow the earth's rotation, neural networks, classic deconvolution?? Or google searches the FBI uncovers when searching a mad scientist villain's computer........

11-11-2020, 03:34 AM	#29
yucafrita Veteran Member Join Date: Oct 2015 Posts: 459	I wonder if this IBIS race is because the photographers are getting older, shakier hands or try to shoot drunken. If you imagine that Pentax was one of the first and the industry as a whole denied the necessity of it (because it's in the lens). Now everyone is raving about it. Strange new world.

11-11-2020, 03:42 AM	#30
Rondec Loyal Site Supporter Join Date: Mar 2009 Location: Gladys, Virginia Photos: Gallery Posts: 27,653	The whole 8 stops of IBIS seems over inflated. If you are shooting a 24-70 lens at 40mm, it feels like even with stabilization that is great, it would be a struggle to hand hold at 6 or 7 seconds, which is what that would be. My experience with Pentax is that IBIS works well till you get to shutter speeds that are slower than 1/2 second and then it doesn't matter what the focal length is, I can't really hand hold it. So, I can hand hold 60mm at 1/6 second and get sharp images, but at 15mm I can only hand hold to about 1/4 second before shake starts to show up.