The S3 pricing is a mistake

[frame-it]

Advertisement (gone after registration)

2 hours ago, John Smith said:

Just an FYI. Leica USA describes the color depth as 16 bit. See here. What that I means I haven't a clue.

anyways..

https://www.digitalcameraworld.com/news/this-is-what-a-600mb-16-bit-tiff-shot-on-a-fujifilm-gfx100-looks-like

[xiaubauu2009]

10 hours ago, BernardC said:

Myself, no.

According to what I've read, 16-bit mode adds about 0.2 stops of DR in the GFX, which is a lot less than you would expect if the 16 bits were 'filled." That difference is all in the shadow noise, of course, so it may not be visible in real life. There's a reason why 0.3 stops was/is the minimum increment in the DIN and ASA systems.

It should be interesting to see how the S3 tests, compared to the competition. I'm sure all the usual experts can't wait to let their Rawdigger/Matlab scripts loose on the DNGs. We will find-out soon enough.

On a related note, lens contrast and flare suppression have a much bigger influence on shadow detail. That's why you have to consider the imaging chain as a whole: camera, lens, shading, atmospheric conditions, processing, output (hard to display 14 stops of range in the real world), etc.

Personally, the 16bit files from the current Sony MF sensor has a different behaviour compare to the older 14bit 50mp 44x33 sensor they did before. the older 44x33 50mp 14bit vs 100mp 16bit is quite obvious. The tonality is very different.

The current one I think is really a 16bit cause I don:t notice that much difference between the p1 IQ4150mp and the files files from the GFX100, at least in the tonality area. Anyway, just didnt: realise the CMOS on the S007 and S3 is a 14bit unit. 

[xiaubauu2009]

14 hours ago, BernardC said:

It's a 14-bit file in a 16-bit container. The sensor puts-out 14 bits of real information, which is more than most sensors do. It's in a 16-bit container for the same reason that Photoshop gives you the option of working on 16-bit TIFFs, but not 14-bit TIFFs.

Those are class-leading numbers, but people get confused because some brands emphasize the size of the container without mentioning the fact that they don't fill it completely.

no wonder I feel the S007 never has quite the special feel like the files from the S006...

[frame-it]

51 minutes ago, xiaubauu2009 said:

Personally, the 16bit files from the current Sony MF sensor has a different behaviour compare to the older 14bit 50mp 44x33 sensor they did before. the older 44x33 50mp 14bit vs 100mp 16bit is quite obvious. The tonality is very different.

The current one I think is really a 16bit cause I don:t notice that much difference between the p1 IQ4150mp and the files files from the GFX100, at least in the tonality area. Anyway, just didnt: realise the CMOS on the S007 and S3 is a 14bit unit. 

https://www.framos.com/us/imx461aqr-c-21053

"The IMX461AQR-C is a diagonal 55 mm (Type 3.4) CMOS active pixel type image sensor with a square pixel array and 102 M effective pixels. This IC incorporates maximum 36 dB PGA circuit and 16-bit A/D converter. 16-bit digital output makes it possible to readout the signals of 102 M effective pixels at high-speed of 2.7 frame/s in still picture mode. In addition, vertical subsampling binning and 3-horizontal pixel weighted binning realize high-speed 12-bit digital output for shooting moving pictures."

[xiaubauu2009]

2 hours ago, frame-it said:

https://www.framos.com/us/imx461aqr-c-21053

"The IMX461AQR-C is a diagonal 55 mm (Type 3.4) CMOS active pixel type image sensor with a square pixel array and 102 M effective pixels. This IC incorporates maximum 36 dB PGA circuit and 16-bit A/D converter. 16-bit digital output makes it possible to readout the signals of 102 M effective pixels at high-speed of 2.7 frame/s in still picture mode. In addition, vertical subsampling binning and 3-horizontal pixel weighted binning realize high-speed 12-bit digital output for shooting moving pictures."

Thanks, the high speed mode of the GFX100 turns the RAW back down to 14bit to get that fps burst.

Edited March 10, 2020 by xiaubauu2009

[albireo_double]

Here is an owner of an S007 and 7 S/CS lenses. Prepared to pay Leica EUR10k and surrender my S007. Anything more than that doesn't make sense to me. I can afford the full price but do not consider the added value worth as much, in the present competitive landscape. My S007 is still working fine and I hope it will continue to do so for many more years. Don't care about 14 vs. 16 bit, the S007 files are very flexible to work with, great dynamic range. I don't believe there will be any AF improvement, based on what I've been reading here (beta testers, original announcement). Not a big deal for me, I can work around the S007's AF limitations (it is no worse than Phase One XF, in my experience...had a lot of puzzling back focus problems in practical use with the Phase, despite a Lens Align target being always on the spot). Bought a Q2 yesterday, love everything about it and plan outdoor model shoots with that camera only, paired with a Profoto B10 on a tiny Manfroto Nanopole stand (to experience the "unbearable lightness of photographing"). Ok, not exactly relevant here, but what I am trying to say is that, for my purposes, the S3 does not necessarily compete with the X1DII or the Fujis but also with a host of other different cameras that are more suitable or enjoyable for my particular purpose (hmm, thinking of it, I should use my Pentax 67II more often). 

Edited March 10, 2020 by albireo_double

[albireo_double]

Advertisement (gone after registration)

5 hours ago, xiaubauu2009 said:

no wonder I feel the S007 never has quite the special feel like the files from the S006...

Not sure that is the case in the first place (depends on processing) and that the S006 is 16 bit or rather 14 bits in a 16 bit container (I don't know, do you?). 

[PeterGA]

The next S from Leica will be mirrorless  I think. I also don't believe that Leica will 'walk away' from the S system. I think Phase One is more at danger than Leica S.the definition of PhaseOne's market and audience is being set by competition not by Phase One themselves. I can see a market for S being not too dissimilar to the coming market for Phase One - in S's case it won't be getting much smaller than it is right now.

 

[xiaubauu2009]

2 hours ago, albireo_double said:

Not sure that is the case in the first place (depends on processing) and that the S006 is 16 bit or rather 14 bits in a 16 bit container (I don't know, do you?). 

I think it is 16bit, all kodak sensor is 16bit.

This 14bit in a 16bit container is the same as the X1d.

[BernardC]

8 hours ago, frame-it said:

This IC incorporates maximum 36 dB PGA circuit and 16-bit A/D converter. 16-bit digital output

That's a great example of what I was talking about. Sony tells you that the sensor has a 16-bit A/D converter, and produces 16-bit files. They conveniently omit to mention if the files contain a full 16 bits of information, but they really emphasize the fact that the container is 16 bits wide. In other words, their system works exactly the same way as Leica's does, but they obfuscate their description so that people will understand something that they did not write (because it would not be true).

The information in the last two bits is essentially random noise. It's there, but it doesn't correspond to anything in your original subject.

Again, the reason why the A/D converter is 16 bits, not 15 or 14 bits, is the same reason that I am typing this on a 64-bit computer, not a 61 bit computer. Digital computers scale in powers of 2.

This 2-bit controversy will disappear when people have the S3 in hand, and test it against the competition. That's when we can compare actual results.

[Jeff S]

1 hour ago, xiaubauu2009 said:

I think it is 16bit, all kodak sensor is 16bit.

This 14bit in a 16bit container is the same as the X1d.

Leica claimed 16 bIt for the M9.  Sandy found it was really 14.

Jeff

[xiaubauu2009]

27 minutes ago, BernardC said:

That's a great example of what I was talking about. Sony tells you that the sensor has a 16-bit A/D converter, and produces 16-bit files. They conveniently omit to mention if the files contain a full 16 bits of information, but they really emphasize the fact that the container is 16 bits wide. In other words, their system works exactly the same way as Leica's does, but they obfuscate their description so that people will understand something that they did not write (because it would not be true).

The information in the last two bits is essentially random noise. It's there, but it doesn't correspond to anything in your original subject.

Again, the reason why the A/D converter is 16 bits, not 15 or 14 bits, is the same reason that I am typing this on a 64-bit computer, not a 61 bit computer. Digital computers scale in powers of 2.

This 2-bit controversy will disappear when people have the S3 in hand, and test it against the competition. That's when we can compare actual results.

so in this case, you are saying that the P1 150 is also essentially a 14bit info in a 16bit container? I thought the container is the file... DNG/fff/RAF/whatever, it can be a 16bit structure, but if the information taken by the sensor is 14 bit, the other 2 bit is like just basically noise/no info at all? 

I would have thought those P1 digital back is 16bit, the GFX100 16bit is tested pretty extensively by one Jim Kasson guy, I think.... but then again, it can have a lot of info in the raw, but if their color profile is fuji... hahahah you get crazy pink color or extreme desaturate grey photos.

[frame-it]

1 hour ago, BernardC said:

That's a great example of what I was talking about. Sony tells you that the sensor has a 16-bit A/D converter, and produces 16-bit files. They conveniently omit to mention if the files contain a full 16 bits of information, but they really emphasize the fact that the container is 16 bits wide. In other words, their system works exactly the same way as Leica's does, but they obfuscate their description so that people will understand something that they did not write (because it would not be true).

The information in the last two bits is essentially random noise. It's there, but it doesn't correspond to anything in your original subject.

Again, the reason why the A/D converter is 16 bits, not 15 or 14 bits, is the same reason that I am typing this on a 64-bit computer, not a 61 bit computer. Digital computers scale in powers of 2.

This 2-bit controversy will disappear when people have the S3 in hand, and test it against the competition. That's when we can compare actual results.

aha i understand now...so basically you mean Sony is lying about the extra 2 bits..makes sense.

so what you explained above applies to Phase one, Hasselblad & Fujifilm, possibly Leica too..

Interesting.

[Jeff S]

8 minutes ago, frame-it said:

aha i understand now...so basically you mean Sony is lying about the extra 2 bits..makes sense.

so what you explained above applies to Phase one, Hasselblad & Fujifilm, possibly Leica too..

Interesting.

Not ‘cut and dried’... I believe it’s called marketing.

Jeff

[davidmknoble]

11 hours ago, xiaubauu2009 said:

no wonder I feel the S007 never has quite the special feel like the files from the S006...

The S007 has an increased dynamic range from the S006 and the color depth is the same for both bodies.

[frame-it]

59 minutes ago, Jeff S said:

Not ‘cut and dried’... I believe it’s called marketing.

Jeff

exactly..but his very definitive reply Was ‘cut and dried’ 😀😁😁

[ZHNL]

There are three case:

16bit ADC native with more than 14 bit DR performance. Sony 100M or phase sensor belong to this group. I thought 007 was from here as well and hope S3 is indeed here. (There might be chance of typo still)

14 bit ADC and digital process to 16 bIt. There is no way such file will have more than 14 stop DR, unless file be down sampled to less pixel for final print target smaller than full resolution sensor. X1D, GFX50S/R belong to this group. S007 and S3 might be here as well and I do hope it is wrong.

16 bit ADC native with less than 14 bit DR performance. Old MFDB and S006 should belong to this group. Most discussion happen here because people question the purpose to using 16bit ADC for analog signal has less than 14 bit performance. (Quantization noise is way less than sensor noise so what is the point to add 2 bit of noise.) however, I personally think this is preferred/expensive approach than 2nd, I believe there will be benefit from this approach. In audio, people can hear signal well below noise floor just one of example. 
 

Either this (S3) was a typo, or Leica did not disclosure info for 007 which should have be done IMO. Or simply 007 was 16 bit native but S3 is 14 bit now. Does this come from under powered processor or just sensor architecture? 
 

if S007 is indeed 14bit ADC, then DSP to 16 bit DNG file, it really should not be considered or documented as 16 bit color depth. It is a market BS or simply lie as X1D claimed. 
 

having said that, I would still say this will not be a purchase factor for me. I don’t buy S3 has nothing to do with this TBF.
 

The difference might be indeed small that is impossible or not easy to demonstrate. However, the hardware implementation will be totally different from cost point of view.  
 

Edit: Fuji didn’t claim they have 16 bits for 50s/50R. So they are not in 2nd group. 

Edited March 10, 2020 by ZHNL

[BernardC]

1 hour ago, frame-it said:

exactly..but his very definitive reply Was ‘cut and dried’ 😀😁😁

No, they did not lie, they omitted some data so that people wouldn't notice that they were talking about the A/D converter and file format, but not about the actual files.

As Jeff very correctly observed, it's called marketing!

I think it's all fair when dealing with a product that (almost) no one will buy on a whim. I am more upset when this happens with something marketed to beginners. As things stand, what they wrote is just as likely to backfire as it is to work.

[Chaemono]

The 16 or 14 bit debate is a bit meaningless for mirrorless MF because the OSPDAF banding may be so strong that it makes any subtle differences in read noise quality inconsequential.  That’s what Jim Kasson concluded in his visual comparisons of Fuji GFX 100 14 and 16 bit raw precision: https://blog.kasson.com/gfx-100/visual-comparisons-of-fuji-gfx-100-14-and-16-bit-raw-precision/

A 16-bit S3 will, therefore, have an edge over the Fuji.
 

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[xiaubauu2009]

5 hours ago, Chaemono said:

The 16 or 14 bit debate is a bit meaningless for mirrorless MF because the OSPDAF banding may be so strong that it makes any subtle differences in read noise quality inconsequential.  That’s what Jim Kasson concluded in his visual comparisons of Fuji GFX 100 14 and 16 bit raw precision: https://blog.kasson.com/gfx-100/visual-comparisons-of-fuji-gfx-100-14-and-16-bit-raw-precision/

A 16-bit S3 will, therefore, have an edge over the Fuji.
 

Ya, if you want the PDAF, you have banding, it:s the archilles heel of the GFX100, but you cannot have both things at the same time, OSPDAF for high accuracy AF and not have banding or less banding with better software.

S3 is a 14bit in CMOS era, again, if this is true, then it explain my feeling of shooting S007 and S006 and the difference I felt from their files.