Jump to content

Comparison of Leica Technical Data Interesting


davidmknoble

Recommended Posts

Advertisement (gone after registration)

So, first, I am excited about the new S3.  I don't know that I will trade up from the S007 just yet, but I am excited about the possibilities.  However, I have posted two screen shots below and wonder what others think.  Leica lists the S007 in the product sheet as having dynamic range up to 15 stops, and color resolution of 16 Bit per pixel.  The S3 literature from Leica lists the dynamic range at up to 15 stops and color resolution as 14 Bit per pixel.  I am not saying I have the equipment or the printing capabilities to recognize the difference, but I do find it interesting...

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

Link to post
Share on other sites

3 minutes ago, davidmknoble said:

So, first, I am excited about the new S3.  I don't know that I will trade up from the S007 just yet, but I am excited about the possibilities.  However, I have posted two screen shots below and wonder what others think.  Leica lists the S007 in the product sheet as having dynamic range up to 15 stops, and color resolution of 16 Bit per pixel.  The S3 literature from Leica lists the dynamic range at up to 15 stops and color resolution as 14 Bit per pixel.  I am not saying I have the equipment or the printing capabilities to recognize the difference, but I do find it interesting...

This is interesting. I’m looking to move to the S3, but I’d like to know what accounts for the difference.

Link to post
Share on other sites

Just now, John Smith said:

This is interesting. I’m looking to move to the S3, but I’d like to know what accounts for the difference.

Not really sure that it isn't just difference in several years between publishing and the S007 may just be 14 bit, which is really the max I've ever seen.  I'm not putting too much stock in this, but found it interesting...

 

Link to post
Share on other sites

3 minutes ago, skimmel said:

Red Dot has it listed as 16-bit. Leica is very good at confusing!

 

https://www.reddotforum.com/content/2020/03/leica-s3-64mp-medium-format-camera-now-shipping/

Thanks David F. for clarifying!

Edited by davidmknoble
Clarified with another post
Link to post
Share on other sites

Advertisement (gone after registration)

49 minutes ago, dfarkas said:

I can shed a little light on this. 

Both cameras capture a 14-bit image, stored in a 16-bit DNG container. Both are capable of 15 stops of DR, more than any other still camera. 

What? That means it is the same as Hasselblad X1D that is not true 16 bits machine at hardware level. This is the first time I heard that S006/007 capture image at 14 bit and then dithering to 16 bit. 

Link to post
Share on other sites

I noticed that too, and anticipated that this would make for an interesting discussion on this board, which is a good thing 🙂

I followed a very long 14/16 bit discussion over at LL a few years ago, and the general consensus seemed to be that the last two bits are there for show, but I'm sure it's an endless discussion and if people see a difference, who am I to say they're wrong? But the way I understood it at the time, technically, there is not much necessity for 16-bit *

Leica must have known in advance this discussion would come up, and I would like to hear their explanation. PR-wise they would perhaps have been better off if they just kept it at 16-bit in the specifications.

* Can of worms

Edited by peterv
typo
Link to post
Share on other sites

11 minutes ago, peterv said:

I noticed that too, and anticipated that this would make for an interesting discussion on this board, which is a good thing 🙂

I followed a very long 14/16 bit discussion over at LL a few years ago, and the general consensus seemed to be that the last two bits are there for show, but I'm sure it's an endless discussion and if people see a difference, who am I to say they're wrong? But the way I understood it at the time, technically, there is not much necessity for 16-bit *

Leica must have known in advance this discussion would come up, and I would like to hear their explanation. PR-wise they would perhaps have been better off if they just kept it at 16-bit in the specifications.

* Can of worms

You're right @peterv, and I like @dfarkas explanation - which makes sense.  So, really, Leica would be better off putting that in their description "14 Bit image, 16 Bit DNG."  

I am sure there are those that will nit-pick this.  What I like is the comparison to other systems.  Nikon, for example, publishes a 12 bit image and 14 bit DNG.  So whether or not we could make use of the extra room - it is there for the future, which is more important to me just in terms of thinking ahead.  The ProPhoto color space is great for the future,  but you do not see much output in that color space.  To me these are equivalent, and when the S20 is out (LOL) we may be able to use S007 and S3 files in terms of color depth.  

Clearly, Leica has worked hard to maintain future use (M Mount lenses) and the fact that the S lenses were "over-built" for the first S body.  Just makes for good investment (and taking up space on the forum hard drive :)

 

  • Like 1
Link to post
Share on other sites

39 minutes ago, ZHNL said:

What? That means it is the same as Hasselblad X1D that is not true 16 bits machine at hardware level. This is the first time I heard that S006/007 capture image at 14 bit and then dithering to 16 bit. 

I'm not sure dithering is the correct term here. Leica isn't interpolating the 14-bit data to make it 16-bit. The data from the A/D converters is 14-bit. The container to hold the image information is 16-bit. Imagine a water glass that is 16 cm tall. You fill it with water to the 14 cm mark. The water is your digitized sensor data and the glass is the DNG container. As to why, this method might result in more flexible manipulation in post processing, or it might simply be a matter of convenience.

I actually have covered this information before, specifically when I interviewed the head of R&D at Leica Dr. Volker Zimmer around the S007 launch at Photokina 2014: 

https://www.reddotforum.com/content/2014/11/why-leica-is-staying-at-37-5mp-for-the-s-typ-007/

 

Quote

And with the supercharged Maestro II processor at the heart of the S 007, image quality from the same sensor is taken even farther with new hard-wired noise reduction and image processing algorithms. The 7,500 A/D converters are all humming along with 14-bit precision and the camera can still maintain a record-setting 3.5 fps, three times faster than a Phase One or Hasselblad system. 

 

Apart from this one tidbit, there is actually a lot of really interesting information about Leica's approach to sensors and processing and may offer insight into why they waited so long to move towards higher resolution sensors. 

 

 

  • Like 3
  • Thanks 1
Link to post
Share on other sites

Also interesting (not evaluating, just find interesting), here are the stats from the S2 brochure.  While the sensors were older, and essentially read noise higher (so less dynamic range), the illustration is 12 stops of dynamic range, and 16 bit depth...

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

Edited by davidmknoble
Link to post
Share on other sites

Since we are in a bit-depth discussion:

-Do anyone have a reasonable answer to why the S006-files (=the S-body I know, but I assume the same holds for the rest if the S-bodies) are more/much more malleable than any files from FF-sensors I have experience with (including all digital Leica Ms but M10, Leica SL/SL2, Panasonic S1R, and Nikon D850)?

Surely the S-sensor is somewhat larger, but FF-systems have similar pixel size, so it must be more than the sensor size. 

Link to post
Share on other sites

Helge, I guess the short answer is that it’s the complete imaging pipeline, from lens hood to print/display presentation.

Of course the internal camera electronics and post processing play a large part, and more bits surely will help with getting more malleable files.

  • Thanks 1
Link to post
Share on other sites

1 hour ago, helged said:

Since we are in a bit-depth discussion:

-Do anyone have a reasonable answer to why the S006-files (=the S-body I know, but I assume the same holds for the rest if the S-bodies) are more/much more malleable than any files from FF-sensors I have experience with (including all digital Leica Ms but M10, Leica SL/SL2, Panasonic S1R, and Nikon D850)?

Surely the S-sensor is somewhat larger, but FF-systems have similar pixel size, so it must be more than the sensor size. 

The other item to consider is the read noise of the sensor.  The noise generated during the process of reading the value of a pixel and getting it stored in memory is the read noise.  The higher the read noise, the less shadow detail is available in the sensor.  As technology gets better and read noise is lowered, more shadow detail - more dynamic range - can be realized.  I would guess that the S system has had a better tuned sensor with less read noise earlier than used in the other bodies.  Just a guess on my part.

  • Thanks 1
Link to post
Share on other sites

  • 4 months later...

One source of the confusion is that the older digital cameras used 16 bits to store each pixel output value in their raw files even though the meaningful part of the data was only 14 bits of A/D output from each cell.  The extra bits, probably in the low order end of every value, would be trimmed off by setting the black point during conversion. Newer cameras can shrink the raw file size by using a lossless conversion scheme, which takes some compute power and battery life, before writing the data out.  The M8 outputted every pixel in only 8 bits, but used a lossy conversion, simply taking the square root of each output value as the stored raw file data.  The SL (601) stored every pixel's output in 14bits, and left out the extra two bits completely, storing the results densely.  Each file was thus 7/8 as long as it would have been if they had used 16 bits to store each pixel.  I would guess the SL2 does the same, but havne't checked yet. 

But I still am skeptical that you can get 15 bits of dynamic range out of 14 bit data. That sounds like marketing-speak.  Presumably there is some compression of the actual image information at one or both ends of the scale.  (A toe or a shoulder, in film terms.) So the quality of those last few bits is less, which may or may not matter.

Edited by scott kirkpatrick
  • Like 1
Link to post
Share on other sites

On 7/17/2020 at 2:32 AM, scott kirkpatrick said:

But I still am skeptical that you can get 15 bits of dynamic range out of 14 bit data. That sounds like marketing-speak.  Presumably there is some compression of the actual image information at one or both ends of the scale.  (A toe or a shoulder, in film terms.) So the quality of those last few bits is less, which may or may not matter.

I think that almost all raw formats use a log scale, starting with Kodak's Cineon format from the early 1990s.

Link to post
Share on other sites

  • 4 months later...
On 3/6/2020 at 8:21 AM, dfarkas said:

I can shed a little light on this. 

Both cameras capture a 14-bit image, stored in a 16-bit DNG container. Both are capable of 15 stops of DR, more than any other still camera. 

Does this now also include the M10R?

Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...