Controversy about Q3 lens

[robgo2]

Advertisement (gone after registration)

The IQ from my Q2 is truly excellent. Once you get over the fact that the lens’s inherent distortion is computationally corrected, you realize that it is a match for almost any other 28mm lens. All that matters is the final result. Plus the Q lens has image stabilization, which counts for something in terms of IQ. 

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[Dr. G]

I always hear that the 28mm lens is actually wider - a 24mm or 26mm.  If it were a 24mm 1.4 and Leica was only using the 28mm FOV center portion of the lens to eliminate vignetting and any edge distortion would that be equivalent to a 28mm 1.7?

[adan]

3 hours ago, Dr. G said:

I always hear that the 28mm lens is actually wider - a 24mm or 26mm.  If it were a 24mm 1.4 and Leica was only using the 28mm FOV center portion of the lens to eliminate vignetting and any edge distortion would that be equivalent to a 28mm 1.7?

What you "hear" is apparently from uninformed folks with unsophisticated optical knowledge. They make the "newbie mistake" of thinking tha numerical focal length is all about a field of view - it isn't.

The Q lens is a 28mm, period , end of story. It has a somewhat wider "native" angle of view only because of built-in fisheye-type distortion.

It is the same as 15mm lenses (or any other focal length, for that matter) - some are 15mm fisheyes that take in 180° of view corner to corner, but have to bend and curve straight lines to do so. And some are 15mm "rectilinear" lenses ("recti" = "straight or correct;" "linear" = "lines") that don't show the same curvature and take in ~110 degrees of view corner to corner. Nevertheless both are 15mm lenses, pure and simple.

https://www.vrphotography.com/data/pages/techtutorials/technotes/panolenschoice.html

Leica simply uses digital corrections to straighten out the mild fisheye "curved lines" the Q 28mm lens produces natively, into a rectilinear image  - rather than using additional glass elements to do it (as would have been done for film lenses).

As to aperture, the aperture value f/1.4 simply defines how much light the lens transmits. Once the light has passed through the lens, that is how much light was transmitted, regardless of whether the image is cropped or straightened afterwards.

[SrMi]

3 minutes ago, adan said:

What you "hear" is apparently from uninformed folks with unsophisticated optical knowledge. They make the "newbie mistake" of thinking tha numerical focal length is all about a field of view - it isn't.

The Q lens is a 28mm, period , end of story. It has a somewhat wider "native" angle of view only because of built-in fisheye-type distortion.

It is the same as 15mm lenses (or any other focal length, for that matter) - some are 15mm fisheyes that take in 180° of view corner to corner, but have to bend and curve straight lines to do so. And some are 15mm "rectilinear" lenses ("recti" = "straight or correct;" "linear" = "lines") that don't show the same curvature and take in ~110 degrees of view corner to corner. Nevertheless both are 15mm lenses, pure and simple.

https://www.vrphotography.com/data/pages/techtutorials/technotes/panolenschoice.html

Leica simply uses digital corrections to straighten out the mild fisheye "curved lines" the Q 28mm lens produces natively, into a rectilinear image  - rather than using additional glass elements to do it (as would have been done for film lenses).

As to aperture, the aperture value f/1.4 simply defines how much light the lens transmits. Once the light has passed through the lens, that is how much light was transmitted, regardless of whether the image is cropped or straightened afterwards.

What’s the focal length of the Leica Q2 lens? (Jim Kasson)

So, there’s your answer: 26.2 mm. After software distortion correction, the lens won’t appear to as wide and before, but it turns out that, with Adobe Lightroom and the default distortion correction, the effective focal length of the corrected lens is about the same as the uncorrected lens in the middle of the field: about 26 mm.

[adan]

Some very questionable assumptions in Kasson's post, such as:

"I measured the width of the image: 8376 pixels. Figuring that 8376 pixels is 36 mm, I calculated the width of the garage door."

No valid reason whatsoever to "assume*" the total image width created corresponds exactly to 36mm. Even if one could take off the lens and actually measured the sensor's active area, not all the pixels on a sensor end up in the picture file. Some are used as "dark reference" pixels' or for other data house-keeping; some are only partly used so that the edge pixels of the final picture get complete GRGB Bayer data - the difference between active pixels and effective pixels.

https://landingfield.wordpress.com/2013/02/06/peeping-into-pixel-a-micrograph-of-cmos-sensor/

I have a Leica 135mm lens that is actually 136.7mm in focal length, according to the little yellow numbers next to its infinity symbol. I have a 90mm Elmarit-M that is actually 91.1mm, according to the markings. Who says the Q actually captures an image exactly 24mm x 36mm?

My M10-M does not produce a perfect "Barnack" image ratio of 1:1.5 - the .DNGs are 1:1.512 (5200 pixels x 7864).

And most people who shoot .DNG + .jpg can confirm that the pixels dimensions are different  (e.g. my vanilla M10 - jpegs of 6000 x 4000 pixels, while .DNGs are 5976 x 3984). Which by Mr. Kasson's logic would imply simply switching from .DNG to .jpg would magically move the glass around and change the focal length of the lens.

Finally, any correction for distortion is going to start right at the center of the image, and then continuously increase out to the corners - the idea that cropping for the center (beyond the single pixel exactly in the center) will somehow "dodge" the corrections is faulty.

Combining all those conceptual and measurement "assumptions" could easily produce a 1.8 mm "experimental error" - if not more.

Jim Kasson may well be a qualified engineer - but if he had submitted that particular post as an actual project in engineering school - he would have gotten an "F."

"Junk Science."

_______________

* "ASSUME" = make an A$$ out of U and ME both.

Edited August 11, 2023 by adan

[Viewfinder_vandals]

Am 9.8.2023 um 22:10 schrieb dpitt:

The Summilux 28 M is sold for €1000 more than the Q3.

Do you think Leica attaches exactly the same lens to your Q3, gives you a €1000 discount and then throws in the FF camera + AF for free?

Yes thats what all YouTubers and Reviewers say 🤣🤡

[Viewfinder_vandals]

Advertisement (gone after registration)

vor 2 Stunden schrieb adan:

Some very questionable assumptions in Kasson's post, such as:

"I measured the width of the image: 8376 pixels. Figuring that 8376 pixels is 36 mm, I calculated the width of the garage door."

No valid reason whatsoever to "assume*" the total image width created corresponds exactly to 36mm. Even if one could take off the lens and actually measured the sensor's active area, not all the pixels on a sensor end up in the picture file. Some are used as "dark reference" pixels' or for other data house-keeping; some are only partly used so that the edge pixels of the final picture get complete GRGB Bayer data - the difference between active pixels and effective pixels.

https://landingfield.wordpress.com/2013/02/06/peeping-into-pixel-a-micrograph-of-cmos-sensor/

I have a Leica 135mm lens that is actually 136.7mm in focal length, according to the little yellow numbers next to its infinity symbol. I have a 90mm Elmarit-M that is actually 91.1mm, according to the markings. Who says the Q actually captures an image exactly 24mm x 36mm?

My M10-M does not produce a perfect "Barnack" image ratio of 1:1.5 - the .DNGs are 1:1.512 (5200 pixels x 7864).

And most people who shoot .DNG + .jpg can confirm that the pixels dimensions are different  (e.g. my vanilla M10 - jpegs of 6000 x 4000 pixels, while .DNGs are 5976 x 3984). Which by Mr. Kasson's logic would imply simply switching from .DNG to .jpg would magically move the glass around and change the focal length of the lens.

Finally, any correction for distortion is going to start right at the center of the image, and then continuously increase out to the corners - the idea that cropping for the center (beyond the single pixel exactly in the center) will somehow "dodge" the corrections is faulty.

Combining all those conceptual and measurement "assumptions" could easily produce a 1.8 mm "experimental error" - if not more.

Jim Kasson may well be a qualified engineer - but if he had submitted that particular post as an actual project in engineering school - he would have gotten an "F."

"Junk Science."

_______________

* "ASSUME" = make an A$$ out of U and ME both.

You dont have to do such mathematics. Just compare any Other 28mm Lens to the q Lens and you will See That Its wider. Period 🤷🏽

[liggy]

"Controversy"

Give me a break.  It's a great package.  Embrace it or not.

I've got sharper lenses.  I've got faster lenses.  I've got wider lenses.  I've got longer lenses. 

They aren't attached to a Q.  

Still love the overall package regardless of this measurbating.  

Edited August 11, 2023 by liggy

[BJohn]

I got both for quite a period of time, Q2 and the M10-R + Lux 28. I decided to keep the M + Lux 28. I think the Lux 28 has a little more sparkle. But for sure, that is more personal preference or just imagination. What changed my mind is the following: I love the rangefinder and MF. That's the only 'real' reason. A friend wanted the Q2 and I could sell it for a reasonable price, nobody needs both.

The Q2 (and Q3 for sure) is a fabulous package! You can see the additional "space" of the lens with Capture One, so it's indeed a little wider, which I enjoyed while shooting architecture. But the truth is, it is hard to see a real difference when comparing both lenses. I did a lot of comparison shots and it was even hard to see 1.4 / 1.7 differences. The bokeh is a little different and - what I meant a few lines before - the Lux 28 got a little more character or sparkle. But again, you can only see that in very specific shots.

In the end, both are perfect tools! Not mediocre or close to perfect, no - perfect! It depends on your use cases and how one does enjoy photography! If you enjoy your tool the most, you will use it more often which helps you grow!

Edited August 12, 2023 by BJohn

[cboy]

10 hours ago, Viewfinder_vandals said:

Yes thats what all YouTubers and Reviewers say 🤣🤡

Yep. It's the headline argument ad nauesum in favour of the Q series. 'Why get just the lens if you can get the camera for free'. Followed by the 'best Q' to date.

[M11 for me]

vor 11 Stunden schrieb Viewfinder_vandals:

Period 🤷🏽

Period. Full Stop.

[city of bristol]

The Q summilux will decrease in value over time but the M summilux will at least historically will retain its value.

In theory you can use the M lens 25 years from now but the Q will be in the rubbish bin while people use their Q 11 , thats why its more expensive in my view as technically both lens have different strengths and weaker areas.

I own a Q 2 

[dpitt]

17 hours ago, city of bristol said:

The Q summilux will decrease in value over time but the M summilux will at least historically will retain its value.

In theory you can use the M lens 25 years from now but the Q will be in the rubbish bin while people use their Q 11 , thats why its more expensive in my view as technically both lens have different strengths and weaker areas.

I own a Q 2 

If we look at the past to predict the future, the Leica Digilux 2 is a good example. The Digilux 2 was released in 2004 at €1890 and I bought mine for €450 only 5 years later. It sort of kept its value around the €400 mark and around its 18th birthday started to increase in value! Good samples can now fetch around €550. For those who do not know this model. It is probably the best 5MP camera with fixed lens ever made. It has a crop factor of 4, but is the best sensor Leica could combine with a lens at the time. The main value is in its superb Vario Summicron F2.0 28-90mm (equivalent) zoom lens. Unfortunately this lens will die with the camera once it fails. But lots of users are still using this ancient camera today for its great quality.

So my prediction is that the Q series will bottom out at some point and then will start increasing in value until it dies. It will never reach the age of my oldest lenses that I still use daily today (e.g. my 70 year old Summicron 50 collapsible) and it will not be sold at 5 times its release price, like the Summicron 50. But the Q will probably outlive 99% of the digital camera's sold today, and that is not bad for anything digital.

As @BJohn said above. The main reason to prefer any M system with Summilux 28mm over the Q is the RF. And on top of that the combo will (probably) hold its value better on the long run. Also, if you want to keep up with progress, the M system seems the most economical solution. Your current Summilux 28 will give better pictures wth the any M released in the future. I doubt Leica will break backwards compatibility with M lenses anytime soon. I also think this combo of newer M with your current Summilux 28 will even keep up (in IQ) with any newer Q for a very long time. 

 

Edited August 13, 2023 by dpitt

[bobtodrick]

Well, I purchased my Q used (less than 3000 frames on it) four years ago when the Q2 came out for $3800 Canadian.

Just checked, going price on eBay right now is about $4000 Canadian.  When you look at used prices for a Canon 5D Mk lV (introduced in 2016) $700 to $1000, I’d say the Q series holds its value quite well.

Will be interesting to see what happens down the road but the Q series, like most Leica’s hold their value fairly well.

[Rdf]

2 hours ago, Steven said:

The answer is no, obviously. Still, don’t forget that with the Q line, leica knows that they will eventually get the lens they originally sold you for a fraction of the price, put it in a new Q and sell it back at full price. So they could have sold it cheaper, if they were not a company that wants to make money. 
That being said, the 28 Summilux M is probably the most spectacular lens I’ve ever owned. 

Am I misunderstanding?  Are you saying that Leica is reselling used Q lenses on new Q cameras?  Thanks. 

[bobtodrick]

28 minutes ago, Rdf said:

Am I misunderstanding?  Are you saying that Leica is reselling used Q lenses on new Q cameras?  Thanks. 

If he is saying this…he is most definitely mistaken!!

[city of bristol]

8 hours ago, dpitt said:

If we look at the past to predict the future, the Leica Digilux 2 is a good example. The Digilux 2 was released in 2004 at €1890 and I bought mine for €450 only 5 years later. It sort of kept its value around the €400 mark and around its 18th birthday started to increase in value! Good samples can now fetch around €550. For those who do not know this model. It is probably the best 5MP camera with fixed lens ever made. It has a crop factor of 4, but is the best sensor Leica could combine with a lens at the time. The main value is in its superb Vario Summicron F2.0 28-90mm (equivalent) zoom lens. Unfortunately this lens will die with the camera once it fails. But lots of users are still using this ancient camera today for its great quality.

So my prediction is that the Q series will bottom out at some point and then will start increasing in value until it dies. It will never reach the age of my oldest lenses that I still use daily today (e.g. my 70 year old Summicron 50 collapsible) and it will not be sold at 5 times its release price, like the Summicron 50. But the Q will probably outlive 99% of the digital camera's sold today, and that is not bad for anything digital.

As @BJohn said above. The main reason to prefer any M system with Summilux 28mm over the Q is the RF. And on top of that the combo will (probably) hold its value better on the long run. Also, if you want to keep up with progress, the M system seems the most economical solution. Your current Summilux 28 will give better pictures wth the any M released in the future. I doubt Leica will break backwards compatibility with M lenses anytime soon. I also think this combo of newer M with your current Summilux 28 will even keep up (in IQ) with any newer Q for a very long time. 

 

For many the main reason not to buy an M is the RF .

[dpitt]

6 hours ago, city of bristol said:

For many the main reason not to buy an M is the RF .

That is ok 😎
I think for most of the M users it was the main reason to buy into the M system. And it probably should be,  because Leica invested a lot of time and effort to make the RF work for them in stead of just doing what all major manufacturers do. Reflex systems (in the past) and EVF systems now where the obvious thing to do. I am glad Leica did not choose the easy route.

Also, the RF system is the main reason why M lenses must be so compact because they would block the RF otherwise . And getting that kind of quality in such a small package is the main selling point for buying an M lens if you do not own an M camera.

i.o.w. to me the RF is the very essence of the M bodies and lenses.

[Advertisement]

Hallo,

Look here: Leica Q

[TDE-Photo]

Am 9.8.2023 um 21:21 schrieb Opaea:

Hello 

There is controversy about the Q lens 28 mm Sumilux witch would be not the same as the original for Leica M for instance. 
The quality should be lower specifically on the edge. 
Has anyone here that ? I haven’t notice that in my use but I have never tested differences between center and edge of my pictures. 

the question is, do you have a problem with the quality of your Q3 lense, if not, why discuss if something else is better or not?

[M11 for me]

vor 10 Stunden schrieb city of bristol:

For many the main reason not to buy an M is the RF .

I think that it is rather the missing Autofocus.