Canon EOS R5C

[Emanuel]

@kye The digital stabilization can vary from a device or method to another. Higher resolution can also play a determinant role there, I guess. I wish it could be as simple as the exposure triangle is ;- )

[kye]

3 minutes ago, Emanuel said:

@kye The digital stabilization can vary from a device or method to another. Higher resolution can also play a determinant role there, I guess. I wish it could be as simple as the exposure triangle is ;- )

It shifts, scales, and rotates each frame to make them line up.

That's it.

It's probably simpler than the exposure triangle.

[Emanuel]

Maybe, you're right on this one : ) I am not engineer but I believe the degree of complexity may bring a different outcome accordingly the implementation of it. I've appreciated what I've seen from this new entry. At least for my needs :- )

[kye]

23 minutes ago, Emanuel said:

Maybe, you're right on this one : ) I am not engineer but I believe the degree of complexity may bring a different outcome accordingly the implementation of it. I've appreciated what I've seen from this new entry. At least for my needs :- )

Regardless of how you implement it, there are two fundamental principles:

1. it happens after the frame has been exposed
2. you can't un-blur an image

Therefore, IBIS / OIS = frames not so blurry, Digital IS = frames hella blurry

 

[Emanuel]

I saw your test. Well done : ) I just wonder on modern digital stabilization mainly this or upcoming cameras to introduce the most advanced technology to gradually provide better results from there.

[Video Hummus]

13 hours ago, kye said:

1Gbps is fine for 8K RAW, but if you were shooting a documentary with literally hundreds (maybe even 1000+) hours of footage, halving the bitrate is a significant thing and can make a huge difference.  

Then use one of the 30 other bitrates the camera offers including some in 8K HEVC! Not that hard. Use the 8K RAW when it benefits you. Use the more compressed formates when storage space is a concern. It's not like this cameras has 3 bitrates options to choose from.

13 hours ago, kye said:

Ummm..... how?

Digital stabilisation can line up the frames with each other, but OIS / IBIS actually stabilises DURING THE EXPOSURE OF EACH FRAME.  If you have ANY motion blur in ANY frame then it's there forever and digital stabilisation can't do a single thing about it.

Even worse though, is that once the digital stabilisation has worked its magic, the shot looks smooth but there will be random blurring of frames without any corresponding change to the overall shot.  In other words, digital stabilisation without having very short shutter speeds will look worse than no stabilisation at all.

Yes trade offs for each system. I think you underestimate the sophistication of gyro and image processing but with regards to practical applications you are going to be using a stabilizer of some kind anyway. 

[Video Hummus]

In the ProAV video the stationary handheld OIS + Digital IS versus IBIS and the results for the digital was just as good if not better without weird warping corners which you would have to crop to get rid of anyway.

I'm not saying Digital IS is going to completely replace IBIS. They each have their trade offs and IMHO optical IS + Digital IS seems to offer less artifacts at wider focal lengths than IBIS can in video. Until we have IBIS mechanisms that can be completely lock down then there will be a market for both. IBIS is clearly beneficial but its not going to suddenly make your camera into a gimbal stabilized beast, not yet anyway.

[herein2020]

2 hours ago, kye said:

Regardless of how you implement it, there are two fundamental principles:

1. it happens after the frame has been exposed
2. you can't un-blur an image

Therefore, IBIS / OIS = frames not so blurry, Digital IS = frames hella blurry

 

 

What I am seeing is worse than even this video shows, if you look closely at handheld C70 footage you will see weird jumping at times that makes the video hard to watch. Mainly when things with a lot of details are in the shot kind of like the details from the first frame and the details from the second frame were aligned too quickly and the process added the jittery feeling that I keep seeing.

IBIS does not do this, it is a much more natural gradual axis correction process whereas digital IS tries to instantly fix what it perceives as problems but to my eyes the fine details are jittery because the speed at which it corrects the image is unnatural. At first I thought it was YouTube compression artifacts but I've never seen that look in any other footage on YouTube. Also, my GoPro's digital IS doesn't do that at all, it looks much more like gimbal footage. 

I really think this jitter could be fixed by a FW upgrade; if they manage to achieve the same look I am getting out of my GoPro then I'd be perfectly fine with no IBIS. Canon's digital IS looks fine when just hand holding or very slowly moving the camera, anything more with lots of detail in the frame or rapid changes in camera direction is when the jitter occurs.

[ac6000cw]

8 hours ago, kye said:

The digital stabilisation was done in Resolve, which has better digital stabilisation than ANY camera will EVER have, because Resolve can see into the future and cameras can't.

IBIS / OIS stabilises DURING the exposure of each frame, digital stabilisation does not and therefore the frames will have motion blur.  

You're not comparing IBIS to in-camera digital image stabilisation in your video, which is what is relevant to the R5 versus R5c situation - you are comparing IBIS to post-production DIS.

AFAIK modern in-camera DIS normally uses in-camera motion sensors to control it (as does sensor-shift IBIS), whereas stabilisation in post normally has to derive the motion information just from the video frame content (motion estimation/prediction). One exception to that is Sony with some A7 models that add camera motion sensor data as metadata to the video files, so their software can perform post-production DIS based on real camera movement data.

I agree with you that you can't remove motion blur due to unintentional camera movement in post DIS - that needs OIS/IBIS. But I suspect that lens OIS + good in-camera DIS can get pretty close to what OIS + IBIS can achieve in a lot of situations, if it's properly integrated and implemented i.e. lens and body from same manufacturer.

Personally, based on my experience with Panasonic GX80, G80 and G9, I almost always have the additional EIS enabled because (for my use cases) I think the minimal image quality degradation is outweighed by the increased stability and lower warping artefacts at wide angles.

[Emanuel]

1 hour ago, ac6000cw said:

You're not comparing IBIS to in-camera digital image stabilisation in your video, which is what is relevant to the R5 versus R5c situation - you are comparing IBIS to post-production DIS.

AFAIK modern in-camera DIS normally uses in-camera motion sensors to control it (as does sensor-shift IBIS), whereas stabilisation in post normally has to derive the motion information just from the video frame content (motion estimation/prediction). One exception to that is Sony with some A7 models that add camera motion sensor data as metadata to the video files, so their software can perform post-production DIS based on real camera movement data.

I agree with you that you can't remove motion blur due to unintentional camera movement in post DIS - that needs OIS/IBIS. But I suspect that lens OIS + good in-camera DIS can get pretty close to what OIS + IBIS can achieve in a lot of situations, if it's properly integrated and implemented i.e. lens and body from same manufacturer.

Thanks for your information to match my idea, it's exactly my impression, hence my apples to oranges comment :- )

[Video Hummus]

2 hours ago, herein2020 said:

What I am seeing is worse than even this video shows, if you look closely at handheld C70 footage you will see weird jumping at times that makes the video hard to watch.

I think the EIS is linked to AF or a combination of things because you can have a the camera on a tripod and get it to jump suddenly. I think it is a bug with the current implementation. Also, in one of the interviews Canon Rep mentioned the EIS is using gyro in the camera. I wonder if we will be seeing future firmware updates to improve its effectiveness?

[herein2020]

52 minutes ago, Video Hummus said:

I think the EIS is linked to AF or a combination of things because you can have a the camera on a tripod and get it to jump suddenly. I think it is a bug with the current implementation. Also, in one of the interviews Canon Rep mentioned the EIS is using gyro in the camera. I wonder if we will be seeing future firmware updates to improve its effectiveness?

That could be the problem. If they do manage to improve it then I am all for ditching the IBIS, but until they do I will wait and see.

[herein2020]

In this video they stated something very interesting that I have not seen anywhere else which is that the existing cages for the R5 will work with the R5c because all of the ports and size is exactly the same. That is pretty interesting to me because it means no waiting for months for your favorite cage maker to make a cage for the R5c. 

CineD I think really asked the hard questions in this interview. I think he also almost got the Canon rep to admit the DR of the R5 is not good enough for CLOG2. The IBIS questions were my favorite and it really remains to be seen if Canon's electronic stabilization is up to IBIS standards. He really got into the AF as well, and for me coming from the S5 and GH5 I have no doubt that the AF in the R5c is more than enough for my needs.

I do wish someone would ask Canon about 5.9K and adding higher compression options for that resolution, that would be way more useful than 8K60FPS. I also wish someone would ask them if the DR in video is worse than the DR for photography specifically for the R5c.

 

 

[Emanuel]

2 hours ago, Video Hummus said:

I think the EIS is linked to AF or a combination of things because you can have a the camera on a tripod and get it to jump suddenly. I think it is a bug with the current implementation. Also, in one of the interviews Canon Rep mentioned the EIS is using gyro in the camera. I wonder if we will be seeing future firmware updates to improve its effectiveness?

  

4 minutes ago, herein2020 said:

CineD I think really asked the hard questions in this interview. I think he also almost got the Canon rep to admit the DR of the R5 is not good enough for CLOG2. The IBIS questions were my favorite and it really remains to be seen if Canon's electronic stabilization is up to IBIS standards. He really got into the AF as well, and for me coming from the S5 and GH5 I have no doubt that the AF in the R5c is more than enough for my needs.

 

This is exactly what I'm talking about...

The ideal would be to have both worlds but if some technology will prove to match my needs from its implementation and improvement, I don't give a damn for anything else, what we know and don't know about, important is that fills the cup if so : ) Not much other counts when our user experience and outcome makes us to praise the joy to see our finishing line to be reached and we just become happy campers :- )

[webrunner5]

I don't see why you can't have both. There has to be a way to lock the sensor block and not have it move, and just use electronic when you want and vice versa. No doubt they both have their place at times. Trouble is you are always going to loose real estate using electronic stabilization.

[Davide Roveri]

11 minutes ago, webrunner5 said:

I don't see why you can't have both. There has to be a way to lock the sensor block and not have it move, and just use electronic when you want and vice versa. No doubt they both have their place at times. Trouble is you are always going to loose real estate using electronic stabilization.

This is exactly what I was thinking! I mean.. pretty much every camera that I've ever used equipped with IBIS has a sensor cleaning mode that does just that.. it locks the sensor in a neutral position so you can clean it easily without it moving all over the place. Surely it'll never be as strong as a sensor which is permanently held in place but those magnets are a lot stronger than I thought so I'm wondering why one couldn't shoot with the sensor in that mode.

Of course I'm saying this purely as an end user without any knowledge of the engineering behind it so it's entirely possible there are very good reasons why this can't work 😅

[kye]

8 hours ago, Emanuel said:

I saw your test. Well done : ) I just wonder on modern digital stabilization mainly this or upcoming cameras to introduce the most advanced technology to gradually provide better results from there.

You cannot un-blur an image.

This seems to be the fundamental thing that you're getting stuck on.  There are billions of dollars waiting for the company that can process an image and restore detail.  Law enforcement would be using it the way it's shown in the movies, but it's just not possible.

Digital stabilisation occurs AFTER the blurry frame has already been acquired.  and if that frame is blurred, there is no way to un-blur it.

It doesn't matter if the digital stabilisation is being done in-camera, in Resolve, in-country, or in-cognito, the only place this is possible is in-fantasyland.

8 hours ago, Video Hummus said:

Then use one of the 30 other bitrates the camera offers including some in 8K HEVC! Not that hard. Use the 8K RAW when it benefits you. Use the more compressed formates when storage space is a concern. It's not like this cameras has 3 bitrates options to choose from.

Yes trade offs for each system. I think you underestimate the sophistication of gyro and image processing but with regards to practical applications you are going to be using a stabilizer of some kind anyway. 

You have simply read one post of mine and forgotten the context of this discussion.  Go back and read it again if you want to understand what I was talking about.

7 hours ago, Video Hummus said:

In the ProAV video the stationary handheld OIS + Digital IS versus IBIS and the results for the digital was just as good if not better without weird warping corners which you would have to crop to get rid of anyway.

I'm not saying Digital IS is going to completely replace IBIS. They each have their trade offs and IMHO optical IS + Digital IS seems to offer less artifacts at wider focal lengths than IBIS can in video. Until we have IBIS mechanisms that can be completely lock down then there will be a market for both. IBIS is clearly beneficial but its not going to suddenly make your camera into a gimbal stabilized beast, not yet anyway.

The ProAV used OIS.  OIS is similar to IBIS in that it stabilises during the exposure.  I've only been talking about Digital Stabilisation ONLY in these comparisons.  Here is the video again - you seem to have mis-read the title.

Plenty of people are using lenses without OIS, which is why IBIS is a useful feature.

6 hours ago, herein2020 said:

What I am seeing is worse than even this video shows, if you look closely at handheld C70 footage you will see weird jumping at times that makes the video hard to watch. Mainly when things with a lot of details are in the shot kind of like the details from the first frame and the details from the second frame were aligned too quickly and the process added the jittery feeling that I keep seeing.

IBIS does not do this, it is a much more natural gradual axis correction process whereas digital IS tries to instantly fix what it perceives as problems but to my eyes the fine details are jittery because the speed at which it corrects the image is unnatural. At first I thought it was YouTube compression artifacts but I've never seen that look in any other footage on YouTube. Also, my GoPro's digital IS doesn't do that at all, it looks much more like gimbal footage. 

I really think this jitter could be fixed by a FW upgrade; if they manage to achieve the same look I am getting out of my GoPro then I'd be perfectly fine with no IBIS. Canon's digital IS looks fine when just hand holding or very slowly moving the camera, anything more with lots of detail in the frame or rapid changes in camera direction is when the jitter occurs.

In the test video I did all the motion in the shots was relatively consistent, so the frame blurring was moderate.  If you're walking or doing something with large sudden movements then those frames would be particularly blurred, so maybe that's what you're seeing?

Maybe you're seeing something else.  I remember the A6300 and A6500 had IBIS problems where the IBIS would jitter about and even go haywire and flick up and down at full-speed, almost threatening to tear themselves apart.  I think they fixed it with a firmware update.

The other elephant in the room here is that a lot of people have gotten used to the video look and the look of poor colours.  There also seems to be a huge number of Canon fans who treat Canon colours and looks as the pinnacle, so when you say there's a problem they cannot understand it because Canon is literally the definition (in their minds) of what is good.
If you bring up the image from an Alexa then they just dismiss you because all expensive cameras are in a parallel universe and therefore do not apply.  Never mind that the 'standard' has gotten worse or should show some problems - that's simply not possible, and how dare you to even suggest it.

6 hours ago, ac6000cw said:

You're not comparing IBIS to in-camera digital image stabilisation in your video, which is what is relevant to the R5 versus R5c situation - you are comparing IBIS to post-production DIS.

AFAIK modern in-camera DIS normally uses in-camera motion sensors to control it (as does sensor-shift IBIS), whereas stabilisation in post normally has to derive the motion information just from the video frame content (motion estimation/prediction). One exception to that is Sony with some A7 models that add camera motion sensor data as metadata to the video files, so their software can perform post-production DIS based on real camera movement data.

I agree with you that you can't remove motion blur due to unintentional camera movement in post DIS - that needs OIS/IBIS. But I suspect that lens OIS + good in-camera DIS can get pretty close to what OIS + IBIS can achieve in a lot of situations, if it's properly integrated and implemented i.e. lens and body from same manufacturer.

Personally, based on my experience with Panasonic GX80, G80 and G9, I almost always have the additional EIS enabled because (for my use cases) I think the minimal image quality degradation is outweighed by the increased stability and lower warping artefacts at wide angles.

I have been talking this whole time about comparing digital stabilisation ONLY to using some form of OIS or IBIS.  That's what my test showed.

Plenty of people shoot with lenses that don't have OIS, which is why the lack of IBIS matters.

10 minutes ago, webrunner5 said:

I don't see why you can't have both. There has to be a way to lock the sensor block and not have it move, and just use electronic when you want and vice versa. No doubt they both have their place at times. Trouble is you are always going to loose real estate using electronic stabilization.

I'm not aware of any camera that has IBIS and a sensor lock feature?  Maybe it's difficult to implement?

In terms of producing cameras with both, I think the film world is far larger than it looks, and by only hanging in forums with mostly videographers and amateurs it can seem like our needs are common but I suspect they're not.  I suspect the vast majority of people would either shoot locked down, or if they needed hand-held stabilisation would just use a gimbal or steadicam.  I am on a number of forums and lists etc that are mostly populated by industry pros and from their perspective a wedding or corporate videographer is about as professional as a 12-year old who streams Minecraft.  Their definition of professional (and therefore who cameras are made for) is people working within the studio system, where shoots are on sets with many people and heaps of equipment, different companies will be handling production and post-production, and they're all essentially specialists in a huge production-line of content.  

The reason I mention this is because when manufacturers want to talk with customers to understand their products, they don't talk to someone who is a solo-shooter, they talk to Roger Deakins, or Walter Murch, or <insert member of the ACS or other professional association here>.  Those people might not have even HEARD of IBIS.

[kye]

26 minutes ago, Davide Roveri said:

This is exactly what I was thinking! I mean.. pretty much every camera that I've ever used equipped with IBIS has a sensor cleaning mode that does just that.. it locks the sensor in a neutral position so you can clean it easily without it moving all over the place. Surely it'll never be as strong as a sensor which is permanently held in place but those magnets are a lot stronger than I thought so I'm wondering why one couldn't shoot with the sensor in that mode.

Of course I'm saying this purely as an end user without any knowledge of the engineering behind it so it's entirely possible there are very good reasons why this can't work 😅

I suspect that engaging those magnets takes a reasonable power draw.  Fine for a few seconds but it might be more than the battery is even capable of outputting.  There are lots of circuits that take huge power for a short burst and so they charge a capacitor prior to the operation and then the capacitor provides the burst.

Even if they didn't do that, it might halve or quarter the battery life of the whole camera.

[Emanuel]

1 hour ago, webrunner5 said:

I don't see why you can't have both. There has to be a way to lock the sensor block and not have it move, and just use electronic when you want and vice versa. No doubt they both have their place at times. Trouble is you are always going to loose real estate using electronic stabilization.

51 minutes ago, kye said:

You cannot un-blur an image.

This seems to be the fundamental thing that you're getting stuck on.  There are billions of dollars waiting for the company that can process an image and restore detail.  Law enforcement would be using it the way it's shown in the movies, but it's just not possible.

Digital stabilisation occurs AFTER the blurry frame has already been acquired.  and if that frame is blurred, there is no way to un-blur it.

It doesn't matter if the digital stabilisation is being done in-camera, in Resolve, in-country, or in-cognito, the only place this is possible is in-fantasyland.

 

1 hour ago, Davide Roveri said:

This is exactly what I was thinking! I mean.. pretty much every camera that I've ever used equipped with IBIS has a sensor cleaning mode that does just that.. it locks the sensor in a neutral position so you can clean it easily without it moving all over the place. Surely it'll never be as strong as a sensor which is permanently held in place but those magnets are a lot stronger than I thought so I'm wondering why one couldn't shoot with the sensor in that mode.

Of course I'm saying this purely as an end user without any knowledge of the engineering behind it so it's entirely possible there are very good reasons why this can't work 😅

I am like you Davide, strictly from an end user perspective with no engineering background other than the film school where we don't discuss the tools, we are trained to make them work and live with their limitations and improvements : )

I imagine there are some other industries looking after something else we don't even need using what we already have.

So far so good. TBH the R5C tests I've seen are promising enough to make it shine within its limits.

No expectations on unblurring anything.

No need for miraculous potion either because the current tools are already fair enough. We have our workarounds too. What we cannot handle, Don, is to reach a price target outside of our budget. Might we buy this R5C if the implementation of the IBIS feature would cost an extra grand, as for instance? Yes. Would we do it from a product already in the $4,500 price range without IBIS or at the cost of its weight then? Probably not.

I believe sometimes it's this kind of decisions the product design team has in mind when they decide to include some feature or not. More than adopting the theory of conspiracy they are there to see our expectations fucked up.

:- )

[herein2020]

1 hour ago, kye said:

 

I'm not aware of any camera that has IBIS and a sensor lock feature?  Maybe it's difficult to implement?

Multiple comments that I have read state that the S1H can lock its IBIS, I don't own one but it seems to keep coming up that the S1H has this feature.

It is funny that Canon states professional film makers wanted IBIS gone yet one of the highest regarded mirrorless cameras for video...the S1H has IBIS. 

My S5's IBIS is awesome and from what I have read the IBIS in the S1 and S1H is even better.

Of course Canon can't just admit that their IBIS implementation just isn't as good as their competition and that it would have cost too much for the R&D to improve it for the R5c, but their endless excuses are quite annoying. 

I am guessing that in 2yrs the R5CII will be out and it will have IBIS along with a long explanation from Canon on why this new R5CII is really for real this time ready for anything.