Manual Focus Assist (MFA)

Pocket Lint

Senior Member
Messages
2,778
Solutions
2
Reaction score
2,771
Can anyone tell me how well the Manual Focus Assist feature works.

Watched a review that it wont hit focus perfectly, and you will end up zooming in liveview anyway, which, for me, seems like a useless feature.

I'm back and forth deciding to buy this body or not. I want to try FF mirrorless out. If the price is right come Boxing Day, I might pull the trigger.
 
There are three focus assist tools: Peaking, Fokus Guide, Magnification.

Peaking works a lot better than on my a7M3. Wide open I can hit focus w/o magnification in many situations. It is by far not as bothering (strong) as on the a7M3 where even at the "low" setting it may turn your whole image yellow, red or white if not in magnification mode.

The Focus Guide is one of the major features this camera offers over my Sony bodies. It is very accurate w/o need to magnify the image. In order to work, the lens needs to communicate with the bodey. So unfortunately it doesn't work with adapted lenses like my Minolta MC/MDs.

Magnification can be configured to come on by just hitting one button. The EVF is so good that also in magnification there is no shimmering or anything that distracts from focussing. The EVF is another compelling argument for this camera.
 
Can anyone tell me how well the Manual Focus Assist feature works.

Watched a review that it wont hit focus perfectly, and you will end up zooming in liveview anyway, which, for me, seems like a useless feature.
I've seen that review, or one stating the same, and it doesn't make much sense to me.

Let me elaborate: he is complaining that he needs to go in 10x zoom anyways after having focused on a target with the focus assist tool. He might be right, but...

The focus assist thing isn't a magical wizardry item, it's a way to visualize the DPAF information in a way to help focusing manually. It's a good interface, I believe, and quite intuitive.

But it's one of two cases: either it has the very same information, and the same degree of accuracy as autofocus (and we are talking mostly one-shot, hand picked single spot af, camera on tripod, which is *extremely* good in my experience on the R), or it has a different, more accurate source of data and then we would be in our rights to discuss why Canon hasn't use that more accurate available data for autofocus itself.

What the af assist widget can do is just help you focus correctly in the spot you set as focusing point. Can't be magic, or interpret you sense of aesthetic, or even outperform autofocus.
 
Can anyone tell me how well the Manual Focus Assist feature works.

Watched a review that it wont hit focus perfectly, and you will end up zooming in liveview anyway, which, for me, seems like a useless feature.
I've seen that review, or one stating the same, and it doesn't make much sense to me.

Let me elaborate: he is complaining that he needs to go in 10x zoom anyways after having focused on a target with the focus assist tool. He might be right, but...

The focus assist thing isn't a magical wizardry item, it's a way to visualize the DPAF information in a way to help focusing manually. It's a good interface, I believe, and quite intuitive.

But it's one of two cases: either it has the very same information, and the same degree of accuracy as autofocus (and we are talking mostly one-shot, hand picked single spot af, camera on tripod, which is *extremely* good in my experience on the R), or it has a different, more accurate source of data and then we would be in our rights to discuss why Canon hasn't use that more accurate available data for autofocus itself.

What the af assist widget can do is just help you focus correctly in the spot you set as focusing point. Can't be magic, or interpret you sense of aesthetic, or even outperform autofocus.
And it also depends on the subject. As a photographer you need to choose the manual focus assist feature that suits the situation best, magnification, peaking or guide. For example

If i am trying to focus on an orchid (you know that center of the flower region that has colored speckles). Now you want those certain flower features to be sharp. At certain distances that AF guide box can cover not only that part of the flower I want to get in focus but other sections as well (which are not in the same FL plane). As such you need to be careful with respect to what exactly the AF system is focusing on. I haven't looked into how much of that box is used... the whole box and the best focus? The center? a weighted approach? etc.

Now if you focus on a wall that is perpendicular to use viewing angle using the guide should work flawlessly since the whole wall all in the same focal plane.

Anyway, I think you get the point... Note though I haven't had an issue using it focusing on an orchid. The bigger problem was any small forward or backward movement on my part being an issue.
 
Can anyone tell me how well the Manual Focus Assist feature works.

Watched a review that it wont hit focus perfectly, and you will end up zooming in liveview anyway, which, for me, seems like a useless feature.
I've seen that review, or one stating the same, and it doesn't make much sense to me.

Let me elaborate: he is complaining that he needs to go in 10x zoom anyways after having focused on a target with the focus assist tool. He might be right, but...

The focus assist thing isn't a magical wizardry item, it's a way to visualize the DPAF information in a way to help focusing manually. It's a good interface, I believe, and quite intuitive.

But it's one of two cases: either it has the very same information, and the same degree of accuracy as autofocus (and we are talking mostly one-shot, hand picked single spot af, camera on tripod, which is *extremely* good in my experience on the R), or it has a different, more accurate source of data and then we would be in our rights to discuss why Canon hasn't use that more accurate available data for autofocus itself.

What the af assist widget can do is just help you focus correctly in the spot you set as focusing point. Can't be magic, or interpret you sense of aesthetic, or even outperform autofocus.
Another cause might be that Thomas considers the 'green arrow' to be a single point meaning 'perfect focus', not a range meaning 'good enough'.

Let's make up a number that represents how in focus something is, what Canon calls 'phase difference' in their DPAF promos and let's say it's a number between -1000 and +1000. For the different cases:
  • Single point AF, small box: if the average of the absolute value of the focus pixels in that box is less than 5, show focus confirmation box.
  • MF assist arrows: if the average of the absolute value of the focus pixels in that box is less than 20, turn arrows green.
  • 10x zoom might be able to achieve 4.
For non wide open pictures it will be good enough, but for shallow DoF, critical focus it might not be since AF will be 4 times more precise and 10x magnification might even be better than AF.

I suspect Canon uses a wider range for the MF assist than single point AF to make it easier to hit the green arrow. If they make it too precise people will have to fiddle a lot longer to get the green arrow.

If that's the case I would like that exposed as a menu option with a slider to dial in the sensitivity.
 
Can anyone tell me how well the Manual Focus Assist feature works.

Watched a review that it wont hit focus perfectly, and you will end up zooming in liveview anyway, which, for me, seems like a useless feature.
I've seen that review, or one stating the same, and it doesn't make much sense to me.

Let me elaborate: he is complaining that he needs to go in 10x zoom anyways after having focused on a target with the focus assist tool. He might be right, but...

The focus assist thing isn't a magical wizardry item, it's a way to visualize the DPAF information in a way to help focusing manually. It's a good interface, I believe, and quite intuitive.

But it's one of two cases: either it has the very same information, and the same degree of accuracy as autofocus (and we are talking mostly one-shot, hand picked single spot af, camera on tripod, which is *extremely* good in my experience on the R), or it has a different, more accurate source of data and then we would be in our rights to discuss why Canon hasn't use that more accurate available data for autofocus itself.

What the af assist widget can do is just help you focus correctly in the spot you set as focusing point. Can't be magic, or interpret you sense of aesthetic, or even outperform autofocus.
Another cause might be that Thomas considers the 'green arrow' to be a single point meaning 'perfect focus', not a range meaning 'good enough'.

Let's make up a number that represents how in focus something is, what Canon calls 'phase difference' in their DPAF promos and let's say it's a number between -1000 and +1000. For the different cases:
  • Single point AF, small box: if the average of the absolute value of the focus pixels in that box is less than 5, show focus confirmation box.
  • MF assist arrows: if the average of the absolute value of the focus pixels in that box is less than 20, turn arrows green.
  • 10x zoom might be able to achieve 4.
For non wide open pictures it will be good enough, but for shallow DoF, critical focus it might not be since AF will be 4 times more precise and 10x magnification might even be better than AF.

I suspect Canon uses a wider range for the MF assist than single point AF to make it easier to hit the green arrow. If they make it too precise people will have to fiddle a lot longer to get the green arrow.

If that's the case I would like that exposed as a menu option with a slider to dial in the sensitivity.
Good one... Though one must note that for the focus guide you are using the regular AF box and not the small one, which already makes a difference on that averaging. Additionally I found that at macro application shots that AF guide is extremely accurate, but you run into that fiddling you are mentioning already. So you are better off rocking back and forth when you are really close to critical focus. The rest of you assumptions/speculation seems to be spot on in terms of how one should logically see it, though one needs to wonder how much better you can achieve focus than the assists when hand holding vs magnifying. On a tripod, for critical and not rushed work just use magnification as it will always give you the best results and most control.
 
True. The fact the focus assist isn't available with manual lenses might hint that it uses distance/aperture information to calibrate the accuracy. So at a long distance, f/11 might be less accurate (and less fiddly) than at mfd, f/1.4.
 
True. The fact the focus assist isn't available with manual lenses might hint that it uses distance/aperture information to calibrate the accuracy. So at a long distance, f/11 might be less accurate (and less fiddly) than at mfd, f/1.4.
It has been known to work with chipped lenses that offer no distance encoding - So it seems it operates without it.
 
Can anyone tell me how well the Manual Focus Assist feature works.

Watched a review that it wont hit focus perfectly, and you will end up zooming in liveview anyway, which, for me, seems like a useless feature.

I'm back and forth deciding to buy this body or not. I want to try FF mirrorless out. If the price is right come Boxing Day, I might pull the trigger.
 
I only have one manual focus lens right now, the 17TSE. It's kind of hard to miss with that one.

I'm wondering if any of you have tried any of the longer TSE lenses on the R.

Does the focus peeking feature help you to "see" the plane of focus as you tilt the lens?

I've been very interested in one of the longer TSE macro lenses for florals and food, probably the 135. I've read a little about the focusing process, which requires most to carry tables on paper or their phone and calculate the focus/focal plane relationship to get a shot in focus on a particular plane. Could I just forget all of that with focus peeking?
 
Interesting, so it only needs aperture ?
I don’t think it needs anything communicating to the body. I think it ‘forces’ a chipped lens to be present to activate, but those chips don’t tell the camera anything about its current state, other than to fudge being a lens present so these features will work. AF confirm on DSLRs works in the same way.
 
Interesting, so it only needs aperture ?
I don’t think it needs anything communicating to the body. I think it ‘forces’ a chipped lens to be present to activate, but those chips don’t tell the camera anything about its current state, other than to fudge being a lens present so these features will work. AF confirm on DSLRs works in the same way.
Are there any chipped adapters for MD lenses?
 
Interesting, so it only needs aperture ?
I don’t think it needs anything communicating to the body. I think it ‘forces’ a chipped lens to be present to activate, but those chips don’t tell the camera anything about its current state, other than to fudge being a lens present so these features will work. AF confirm on DSLRs works in the same way.
Are there any chipped adapters for MD lenses?
There have been chipped EF adapters for ages and the later ones should be ok with the EOS R and the EF adapter. The problem is that the MD mount has a shorter registration distance than the EF mount, so the current MD to EF adapters have glass in them to maintain infinity focus. How long it will take people to decode the RF mount protocol and produce RF mount adapter chips is anyone's guess.
 
Last edited:
How long it will take people to decode the RF mount protocol and produce RF mount adapter chips is anyone's guess.
I suspect that's not necessary and the RF is just a superset of the EF mount signals and the "basic" adapter is just a passive connector.

Is there any evidence to the contrary?
 
Last edited:
How long it will take people to decode the RF mount protocol and produce RF mount adapter chips is anyone's guess.
I suspect that's not necessary and the RF is just a superset of the EF mount signals and the "basic" adapter is just a passive connector.

Is there any evidence to the contrary?
Well that's your guess. We'll find out with the second or third generation chips. The presence of a chip is only required to tell the camera that there's a lens mounted for the focus assist to work, but that's no use if the chip confuses the camera.

In the meantime I'll be using EF adapters or focus peaking/magnification.
 
Last edited:
How long it will take people to decode the RF mount protocol and produce RF mount adapter chips is anyone's guess.
I suspect that's not necessary and the RF is just a superset of the EF mount signals and the "basic" adapter is just a passive connector.

Is there any evidence to the contrary?
I agree to an extent, I believe this is basically what happened with EF-M. I feel like one of the reasons that RF exists is that EF-M was too open as a result and maybe they are trying to lock down their new major platform. BUT I also agree that the adapters are doing zero translation, so that there can be direct pass through of EF protocols to the body. I partly think this is why EF focus performance is more robust than RF (for now) as it is ported direct from the existing range (70D onwards).

Thing is we have to wait for manufacturers to make a chip / adapters that take advantage of this. Plus there might even be some kind of encryption or who knows what. So while the protocols are direct, there might be a key of sorts involved in accessing them directly.
 
How long it will take people to decode the RF mount protocol and produce RF mount adapter chips is anyone's guess.
I suspect that's not necessary and the RF is just a superset of the EF mount signals and the "basic" adapter is just a passive connector.

Is there any evidence to the contrary?
I agree to an extent, I believe this is basically what happened with EF-M. I feel like one of the reasons that RF exists is that EF-M was too open as a result and maybe they are trying to lock down their new major platform.
Think you are reading into things too much. I doubt that the RF exists due to the EF-M in any form or way. Canon had the chance for the first time in decades to just rethink their FF mount, probably with the EF-M in consideration but not a must have. As they point out in their launch, the mound dimensions were chosen based on robustness. Remember any reports of sony A7 mount issues? So I think it is about foresight regarding these very large heavy high quality lenses they planned on designing and placing on the thing (how heavy is the 28-70 f2??).
BUT I also agree that the adapters are doing zero translation, so that there can be direct pass through of EF protocols to the body. I partly think this is why EF focus performance is more robust than RF (for now) as it is ported direct from the existing range (70D onwards).
R software probably recognizes which type of lens is mounted and uses the appropriate protocols for it.
Thing is we have to wait for manufacturers to make a chip / adapters that take advantage of this. Plus there might even be some kind of encryption or who knows what. So while the protocols are direct, there might be a key of sorts involved in accessing them directly.
 
I just had a chance to briefly try MF Assist with my 24 tilt shift on EOS R and the result are very good. This is especially useful handholding, which is very difficult to do when tilting and/or shifting!
 

Keyboard shortcuts

Back
Top