Q3: Triple resolution Sensor

[Simone_DF]

Advertisement (gone after registration)

On 6/20/2024 at 11:31 PM, Denebola said:

A good technical explanation of the BSI sensor,  triple resolution, pixel binning, dynamic range and so on are quite well discussed in this video:
 

 

This has been debunked multiple times. There’s no binning on Leica cameras. Just search the forum

[jaapv]

1 hour ago, Simone_DF said:

This has been debunked multiple times. There’s no binning on Leica cameras. Just search the forum

Somehow this guy is quite good but always makes one basic mistake in his videos.

You are correct of course; Leica uses different algorithms during demosaicing. 

[M11 for me]

We have to be aware that in the very beginning of the Q3 (or M11) Leica themselves called it pixel binning. Only later it got clear that Leica actually used a different method. Pixel Binning is in fact a term that refers to one specific method so that it should definitely not be wrongly applied. But as a matter of fact this is only a term and in the video the term pixel binning was not explained and is actually of little or no  importance.

I think that there would be an interesting new video explaining how pixel binning works and what different methods there are to reduce the number of pixels. There are probably such videos. If someone knows a good explanation then it would be welcome to post here. Maybe at the end it does not matter whether Leica uses method A or method B or whatever. Why did Leica people themselves call it Pixel Binning? Just because it was a known term?

[01af]

vor 1 Stunde schrieb M11 for me:

Why did Leica people themselves call it Pixel Binning? Just because it was a known term?

Yes, that would be my guess.

[Guest]

1 hour ago, M11 for me said:

..

Pixel Binning is in fact a term that refers to one specific method so that it should definitely not be wrongly applied. 

...

 

Just for clarification, are you referring to Apple's Patent https://patentimages.storage.googleapis.com/69/1a/cf/cabae58e600e34/US9686485.pdf ?

 

[M11 for me]

What I learned anyway: I set my Q3 to M-DNG and will see if there is a difference.

What I can say already is that the M11 images are easier to work with in Lightroom Classic compared to Q3 set to L-DNG on both cameras.

If I understand things right then the Q3 should have the same DR when set to M-DNG as the M11 has with either L- or M- or S-DNG. The sensors of the 2 cameras are far from being the same (besides AF functions on the sensor of the Q3 which the M11 does not need). 

Edited June 27, 2024 by M11 for me

[M11 for me]

Advertisement (gone after registration)

vor 4 Minuten schrieb Corius:

Just for clarification, are you referring to Apple's Patent https://patentimages.storage.googleapis.com/69/1a/cf/cabae58e600e34/US9686485.pdf ?

 

Thank you. I was not aware that Pixel Binning was an Apple patent dating from 2017.

Edited June 27, 2024 by M11 for me

[SrMi]

3 hours ago, M11 for me said:

Thank you. I was not aware that Pixel Binning was an Apple patent dating from 2017.

There are many patents relating to pixel binning.

[Guest]

26 minutes ago, SrMi said:

There are many patents relating to pixel binning.

Yes, there are, although some relate to video, and a couple of patents were never fully approved, I  agree with you. Having multiple patents that relate to pixel binning would imply that "pixel binning" is not, in itself, a patentable algorithm. i.e. there are multiple ways of implementing pixel binning.

So given that pixel binning, in its broadest definition, means to analyse adjacent pixels and apply an algorithm that reduces the number of pixels recorded then why is Leica's method not "pixel binning" as some people have said? 

 

[SrMi]

42 minutes ago, Corius said:

Yes, there are, although some relate to video, and a couple of patents were never fully approved, I  agree with you. Having multiple patents that relate to pixel binning would imply that "pixel binning" is not, in itself, a patentable algorithm. i.e. there are multiple ways of implementing pixel binning.

So given that pixel binning, in its broadest definition, means to analyse adjacent pixels and apply an algorithm that reduces the number of pixels recorded then why is Leica's method not "pixel binning" as some people have said? 

 

https://en.m.wikipedia.org/wiki/Pixel_binning

[01af]

vor 13 Stunden schrieb M11 for me:

I was not aware that Pixel Binning was an Apple patent dating from 2017.

It isn't.

The referred-to Apple patent is not about pixel binning in general but about a special way of pixel binning.

.

vor 9 Stunden schrieb Corius:

... then why is Leica's method not "pixel binning" as some people have said?

Because pixel binning creates each binned signal from an integral number of pixels. Each original pixel contributes to exactly one binned pixel. So pixel-binning a 60 MP sensor would result in 30 MP, 20 MP, or 15 MP images. But not 36 MP and 18 MP.

[jjroroek]

there is a good lab test showing that DR does indeed increase slightly at lower MP.
 I have tested it myself in practice and, to be honest, I cannot see the difference, although I sometimes 
have the impression that I can see slightly less detail from the shadows than with the m10r.
 but it is not clear whether this is the case and whether this has to do with the DR. So I only use the Q3 at 60 MP,
 my photo archive is not a laboratory

Edited September 10, 2024 by jjroroek

[SrMi]

6 hours ago, jjroroek said:

there is a good lab test showing that DR does indeed increase slightly at lower MP.
 I have tested it myself in practice and, to be honest, I cannot see the difference, although I sometimes 
have the impression that I can see slightly less detail from the shadows than with the m10r.
 but it is not clear whether this is the case and whether this has to do with the DR. So I only use the Q3 at 60 MP,
 my photo archive is not a laboratory

Do you have a reference for that test? The common mistake is checking DR without normalizing the output size, which is wrong.

[jjroroek]

On 9/10/2024 at 5:00 PM, SrMi said:

Do you have a reference for that test? The common mistake is checking DR without normalizing the output size, which is wrong.

 

[SrMi]

2 hours ago, jjroroek said:

 

Thanks, that one was discussed.

I did not watch the video completely, but its conclusions have been rebuked even by Leica (the effect is similar to resizing in PhotoShop).

 

[jjroroek]

5 hours ago, SrMi said:

Thanks, that one was discussed.

I did not watch the video completely, but its conclusions have been rebuked even by Leica (the effect is similar to resizing in PhotoShop).

 

can you please post te link to Leica's rebuke. I wonder why?

[SrMi]

4 hours ago, jjroroek said:

can you please post te link to Leica's rebuke. I wonder why?

Because it is wrong to say that lower resolutions have more DR. While Leica initially claimed that there is more DR, they removed that claim once it became clear to them that DR is measured at the same output size.

Leica’s Jesko von Oyenhausen said in a YT:

• Going to a lower resolution means less noise in the shadows (better DR). The same effect can be achieved with 60MP if you reduce the resolution in post. (43:10)

 

 

[Aram Langhans]

I wonder if a future firmware update would be able to introduce true pixel binning to the Q3 or if there is some hardware component involved.