Classic M vs. M EV1 - Focus Accuracy & Speed

[SrMi]

Advertisement (gone after registration)

4 hours ago, Tailwagger said:

Indeed, I have previously mentioned many times over the years the possibility of porting the SL/SL2 CDAF implementation given the common processor family.  That said, I recently since heard from someone in the know that CDAF on the M is not possible due to "a lack of hardware support".  Initially I thought that statement was BS and didn't bother to challenge it for an explanation.  But upon reflection, I've concluded that 'a lack of hardware support' referred not to a deficiency with the camera, but rather to the lens family.  Likely, the algorithms are tightly tied to having a precise notion of lens positioning via the electronic connection, something, of course, the M mount does not provide.  OTOH, my supposition might be all BS and what was meant was that currently there's no suitable led to light up in the VF.  Dunno.

If by Nikon you're referring to (D)SLRs, my Pentax 645D has a similar system, my understanding is that its the same system that was in use back in the film days for AF, ie. you have a split prism and a sensor.  I still own several film-based autofocus rangefinders, so certainly the RF piece is not an impediment per se, but again, such a system would require new hardware in the M's VF optical path which one suspects, as this could have been done several decades ago, is a place they are not likely to go. 

I suppose we might hope they could wave a magic AI wand to get something fully in software, but who knows where this feature is on their priority list (personally I think it should be at or near the top).  There might not be enough processor/memory budget in the overall context of they want to accomplish.

While I do not expect any AF to be implemented, I would expect the same software that detects correct focus in CDAF to be used to notify the user of acquired focus in M. But I have no insider knowledge. I believe that phase detect is needed only to report the direction of focus adjustment, as X2D does.

I was referring to Nikon Zf, a camera everybody is touting as having the best manual focus assists. That is partly because subject detection can be combined with manual focusing. However, the primary focus assist in Zf is just a square that turns green when focus is acquired.

I do not expect subject detection or phase detect points to be added to M-EV1, as they are primarily needed in AF systems.

[Tailwagger]

2 hours ago, SrMi said:

I would expect the same software that detects correct focus in CDAF to be used to notify the user of acquired focus in M.

CDAF algorithms rely on knowing the direction of focus travel and the ability to reverse direction to home in on max sharpness.  Clearly the M understands when a focus operation is taking place given the zoom and peeking functions, but if it doesn't understand which direction and how much, the notion of overshoot and return doesn't really fit at least algorithmically.  OTOH, given the camera supports peeking, perhaps something similar to CDAF could be implemented purely in software assuming the user overshoots slightly and arrow pops up to suggest a course reversal.  Regardless, I'm was simply reporting on what I was told, CDMF isnt possible without hardware support, and so I speculated on why that might be the case.   

The Zf has a PDAF sensor (coupled with CDAF), which presumably its MF implementation takes advantage of.  Curious as I don't spend much time outside the Leicaverse these days, are these reports of excellence around use of non-native MF lenses or just AF lenses in MF mode?

[EDIT] In thinking a little further about it, its actually interesting that the Q3 continues to rely on mag/peeking for MF focus assist rather than doing something a bit more sophisticated, given it does have a PD sensor. Particularly so as the manual acknowledges the limitations of peeking. 

Edited December 5 by Tailwagger

[SrMi]

33 minutes ago, Tailwagger said:

CDAF algorithms rely on knowing the direction of focus travel and the ability to reverse direction to home in on max sharpness. 

Unlike PDAF, the CDAF algorithm does not know which direction to move initially, but can determine it based on the contrast difference. That is why it tends to hunt.

The Nikon Zf does not indicate which direction to move the focusing, and people seem to be OK with that. IMO, it works well and is better than focus peaking.

37 minutes ago, Tailwagger said:

The Zf has a PDAF sensor (coupled with CDAF), which presumably its MF implementation takes advantage of.  Curious as I don't spend much time outside the Leicaverse these days, are these reports of excellence around use of non-native MF lenses or just AF lenses in MF mode?

Manual focusing on the Zf is the same for AF and MF lenses, but the lens must communicate its focal length for it to work. I do not see why PDAF is needed for manual focusing with Zf. CDAF should be enough. The camera still sometimes has problems focusing with lenses that do not have an automatic aperture stop-down.

My main point is that I do not believe that PDAF brings anything to focus confirmation via a dot or a square that changes color. The only thing that the camera needs to report is that the scene under the square is in focus, for which contrast-detect should be fine. If that is true, it gives me hope that Leica could implement an improved focus confirmation method without phase-detect pixels "ruining" the beautiful sensor data :).

[Tailwagger]

9 minutes ago, SrMi said:

The only thing that the camera needs to report is that the scene under the square is in focus, for which contrast-detect should be fine.

The problem though is how does the CD software know when peek sharpness (contrast) is reached without first going past that point and seeing the contrast level diminish? PD sites OTOH split the image to provide a mechanism (ironically) similar to the RF notion.  When the convergence point (which can be pre-calulated) is reached it can be signaled to the user. 

I'm, btw, not arguing for the inclusion of PD sites (or Lidar for that matter). I suspect that many RF users already do a slight overshoot and return anyway, so CD in software might indeed be able to get the job done.  But regardless of the how, I do think we'd both agree that some simple (led light), accurate form of focus confirm needs to be on the table for all the next gen Ms. 

[SrMi]

3 minutes ago, Tailwagger said:

The problem though is how does the CD software know when peek sharpness (contrast) is reached without first going past that point and seeing the contrast level diminish? PD sites OTOH split the image to provide a mechanism (ironically) similar to the RF notion.  When the convergence point (which can be pre-calulated) is reached it can be signaled to the user. 

I'm, btw, not arguing for the inclusion of PD sites (or Lidar for that matter). I suspect that many RF users already do a slight overshoot and return anyway, so CD in software might indeed be able to get the job done.  But regardless of the how, I do think we'd both agree that some simple (led light), accurate form of focus confirm needs to be on the table for all the next gen Ms. 

I do not think that the focus confirmations in other cameras detect peak sharpness but sufficiently strong contrast, similar to focus peaking. If the DOF is large, e.g., with a closed working aperture, then Zf's focus confirmation is activated for a wide focusing distance.

[Tailwagger]

13 minutes ago, Tailwagger said:

But regardless of the how, I do think we'd both agree that some simple (led light), accurate form of focus confirm needs to be on the table for all the next gen Ms. 

I guess I'd add that as my vision is still up to the task, its not a personal requirement, but rather a belief that such an addition would broaden the M customer base not only to folks who no longer feel they can achieve accurate focus though the RF but for those, like myself when first coming to the M all those years ago, who question whether or not they can get reliable results via using manual focus, let alone via the RF. 

Edited December 5 by Tailwagger

[IkarusJohn]

Advertisement (gone after registration)

The manual focus confirmation in the X2D is very good, though I appreciate that it only works with the AF XCD lenses.  Hopefully, Leica won’t spoil the base M camera with something overly complicated.  The little green light on the Nikon film cameras was enough.

[charlesphoto99]

2 hours ago, Tailwagger said:

I guess I'd add that as my vision is still up to the task, its not a personal requirement, but rather a belief that such an addition would broaden the M customer base not only to folks who no longer feel they can achieve accurate focus though the RF but for those, like myself when first coming to the M all those years ago, who question whether or not they can get reliable results via using manual focus, let alone via the RF. 

Thing is, with the M8/9/240/10/11 etc there's never been an easier time to learn to use the traditional rangefinder than now, as one can get immediate feedback as to whether the focus is good or not. Yes, I know not a help if the subject is moving and you miss the moment, but in the film days one would have to wait a day or week to get results back, and no do-overs on the spot if focus was missed (or the rf went out of calibration). I have a whole trip to Japan with slightly soft images (esp in closer - depth of field covered the distance shots) because the focus screen was out of whack on my Rolleiflex. Digital makes all of this so easy (too easy?). 

[Tailwagger]

21 minutes ago, charlesphoto99 said:

Thing is, with the M8/9/240/10/11 etc there's never been an easier time to learn to use the traditional rangefinder than now, as one can get immediate feedback as to whether the focus is good or not. Yes, I know not a help if the subject is moving and you miss the moment, but in the film days one would have to wait a day or week to get results back, and no do-overs on the spot if focus was missed (or the rf went out of calibration). I have a whole trip to Japan with slightly soft images (esp in closer - depth of field covered the distance shots) because the focus screen was out of whack on my Rolleiflex. Digital makes all of this so easy (too easy?). 

110% agree.  But I'm not sure whether that reality overcomes the preconceptions of those who have only ever used AF (and are bombarded daily with gallons of ink spilled on advanced AF systems being the prime motivator for new camera purchase).  To broaden the horizon of the EV (and M) beyond these walls, the influencers have to board the hype train and toot loudly about how easy, reliable and downright cool MF has become.  That little green light, @IkarusJohn referenced, might go some way in that regard. 

[Giulio Zanni]

Sorry for my ignorance, I just acquired an MV-1 coming from rangefinder. I tried to read the instructions but I could not see the description of the function that by moving the little cross with the directional pad on the point that you want to enlarge, when you actually enlarge that point and focus you can then zoom out without the need to recompose. I just wandered if on the instructions there was more to it...

Edited December 6 by Giulio Zanni

[lct]

Unsure if this answers your question but did you try to half-press the shutter release?

Edited December 6 by lct

[bcaslis]

3 hours ago, Giulio Zanni said:

Sorry for my ignorance, I just acquired an MV-1 coming from rangefinder. I tried to read the instructions but I could not see the description of the function that by moving the little cross with the directional pad on the point that you want to enlarge, when you actually enlarge that point and focus you can then zoom out without the need to recompose. I just wandered if on the instructions there was more to it...

Yes you can. Set a button to magnify (such as the frame line lever), then when focused, you can press the button again or half press the shutter button.

 

[FlashGordonPhotography]

On 12/5/2025 at 1:28 AM, algrove said:

Gordon

Do you have the EVF set to the B&W profile?

No. But that’s a good idea. I will try it. Thanks.

Gordon

[algrove]

4 hours ago, FlashGordonPhotography said:

No. But that’s a good idea. I will try it. Thanks.

Gordon

I generally like Monochrome HC.

[Derbyshire Man]

On 12/5/2025 at 9:00 PM, charlesphoto99 said:

Thing is, with the M8/9/240/10/11 etc there's never been an easier time to learn to use the traditional rangefinder than now, as one can get immediate feedback as to whether the focus is good or not. Yes, I know not a help if the subject is moving and you miss the moment, but in the film days one would have to wait a day or week to get results back, and no do-overs on the spot if focus was missed (or the rf went out of calibration). I have a whole trip to Japan with slightly soft images (esp in closer - depth of field covered the distance shots) because the focus screen was out of whack on my Rolleiflex. Digital makes all of this so easy (too easy?). 

This is one thing that’s changed now that I use Leica. On trips I take a ‘first aid’ kit - 2mm hex key, plastic screwdriver for the front screw, blower, LED loupe, lens brush, couple of lens swabs and cleaning fluid. Having been caught out before!

[jaapv]

On 12/5/2025 at 5:52 AM, SrMi said:

I do not think anything needs to be added in hardware for a focus confirmation similar to that in Nikons. After all, CDAF can determine the focus without PDAF, LiDAR, etc.

But if does need Focus Pixels on the sensor to be able to show the location of optimal focus.

[SrMi]

31 minutes ago, jaapv said:

But if does need Focus Pixels on the sensor to be able to show the location of optimal focus.

The focus confirmation only shows if the focus under the focusing square has been acquired.

With CDAF, and PDAF on latest Olympus cameras, every pixel is a focus pixel.

[algrove]

I just remember the teaser images Leica Rumors showed consistently of a child's eyeball having a red on it (obviously focus peaking in red). At the time I imagined that meant specific focus points would be  in red. I now assume that was due to a narrow depth of focus lens and nothing more.

But now I have not heard of anyone saying that this feature is useful at present in their EV1. 

 

Or am I missing some posts to a post happy subject?

[charlesphoto99]

11 minutes ago, algrove said:

I just remember the teaser images Leica Rumors showed consistently of a child's eyeball having a red on it (obviously focus peaking in red). At the time I imagined that meant specific focus points would be  in red. I now assume that was due to a narrow depth of focus lens and nothing more.

But now I have not heard of anyone saying that this feature is useful at present in their EV1. 

 

Or am I missing some posts to a post happy subject?

Those images on Leica Rumors were not from Leica, but somebody imagining what an EV1 might be like. I think they fooled a lot of people. 

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[lct]

1 hour ago, algrove said:

I just remember the teaser images Leica Rumors showed consistently of a child's eyeball having a red on it (obviously focus peaking in red). At the time I imagined that meant specific focus points would be  in red. I now assume that was due to a narrow depth of focus lens and nothing more. But now I have not heard of anyone saying that this feature is useful at present in their EV1.  [...]

I don't remember these images but focus peaking is not only useful but indispensable to me, either with magnification for nailing focus or without it for good enough focusing. Matter of practice i guess. I prefer FP in white personally. YMMV.