Jump to content

The reality of the 8-bit Leica


Recommended Posts

Advertisement (gone after registration)

After several tests read in Spain on 8-bit sensor Leica, I understand that the 8-bit are actually tablets, actually 12 bits or more within 8. Over the years there have been evidence in this regard, but I do not have the truth about it.

 

8 bitx tablets really are?

 

What's true in the 8-bit compressed Leica?.

 

Looking for answers in the form and not encounter.

 

The 8-bit can be changed by discounting to 12 in firmware?

 

Regards.

Link to post
Share on other sites

  • Replies 103
  • Created
  • Last Reply

Top Posters In This Topic

Top Posters In This Topic

Posted Images

actually, it's 14bits, but they discard 2 of them.

What you say is documented in a very interesting article in LFI, 2/2007.

If you are still interested i can send a scanned image of it thru PMs, I don't know if its proper to publish it here...

There is a reason why M8 produces such cool files, and why you shoot underexposed. The sensor is optimized for blacks

Link to post
Share on other sites

Sorry, but the post isn't at all clear to me... do you refer to M8 sensor , I suppose ? The question of 8-12 bit is rather well known (and deeply examined somewhere in the forum, so as in the quoted LFI article); simply said : the M8 (Kodak) sensor records the pixel values in 12 bit format, then firmware applies a "linearization alghoritm" that maps the values into a 8 bit space, through a "conversion curve" which determines, effectively, the tonal range of the RAW-DNG file, which so is a 8 bit-per-pixel file. In tech terms, it is questionable if this operation can be defined as "compression"... of course a 12 bit file would be bigger than a 8 bit file, but usually they call "compression" an operation whose goal is reducing the dimension of a certain data file...in this case the goal is different, because the linearization alghoritm is necessary to obtain the basic RAW file... the reason for it's a 8 bit file is a design choice.

Link to post
Share on other sites

actually, it's 14bits, but they discard 2 of them.

What you say is documented in a very interesting article in LFI, 2/2007.

If you are still interested i can send a scanned image of it thru PMs, I don't know if its proper to publish it here...

There is a reason why M8 produces such cool files, and why you shoot underexposed. The sensor is optimized for blacks

 

 

Impossible look 2/2007 for me.

 

I mean DNG files 8-bit color depth.

Link to post
Share on other sites

Advertisement (gone after registration)

Although the purists in us say that the full 12 bit samples should be made available (in 16 bit data "packets" for processing convenience), the fact remains that half of the 4096 numeric values each pixel can take on are reserved for the brightest f-stop of dynamic range which many photographers are careful not to over-exploit for fear of clipping and losing highlight detail. "Exposing to the Right" is all to do with making use of as many of the available quantisation levels as possible to increase dynamic range.

 

The same is true in 16 bit Digital Audio where the top 32768 of the available 65536 quanitsation levels are reserved for the loudest 3dB of sound, at which level the ears are incapable of discerning the finest detail anyway.

 

It makes sense therefore (if there are other benefits such as memory space, file write times, battery life) to compress the data into 8 bits and use a non-linear mapping technique which preserves the dark f-stops where there's a shortage of quantisation levels and steal from the bright f-stops where there's an embarrassment of riches.

Link to post
Share on other sites

Whatever anybody says, and I know all the arguments -and I think the M8's files are excellent, I still attribute the slightly better quality of DMR files to the true 16 bits they are ( well, ok, 14;))

Link to post
Share on other sites

Jaap, I agree it would certainly be interesting to be able to place the M8 into a 16 bit mode to see what improvement is available. Whether it's different use to which the DMR and M8 are put or this 16 bit question, I've often felt the DMR images have that extra bite, an extraordinary 3-dimensionality.

 

Leica's earliest firmware supposedly wrote 16 bit files and although they said they made a comparison and could not see any difference, these were presumably the same people who didn't notice that black fabrics were turning magenta and green leaves yellow...

Link to post
Share on other sites

Leica's earliest firmware supposedly wrote 16 bit files and although they said they made a comparison and could not see any difference, these were presumably the same people who didn't notice that black fabrics were turning magenta and green leaves yellow...

Maybe at the time the M8 was developped, the processing of 20 Mbit instead of 10 Mbit files meant such an increase in costs, that given the slight increase in quality it was simply a cost cut rationalisation to go to 8 bits compressed.

 

It would indeed be very nice to compare an uncompressed 14 bit or even 16 bit shot, to a picture compressed to 8 bits.

If somebody can give me a nice raw picture stored in 14 or 16 bit from a well known camera, I can most likely do the compressing to 8 bits and show it to the forum.

 

Hans

Link to post
Share on other sites

Maybe at the time the M8 was developped, the processing of 20 Mbit instead of 10 Mbit files meant such an increase in costs, that given the slight increase in quality it was simply a cost cut rationalisation to go to 8 bits compressed.

 

It would indeed be very nice to compare an uncompressed 14 bit or even 16 bit shot, to a picture compressed to 8 bits.

If somebody can give me a nice raw picture stored in 14 or 16 bit from a well known camera, I can most likely do the compressing to 8 bits and show it to the forum.

 

Hans

 

No I don't think so. It was done because Leica had already chosen a battery and they had already chosen NOT to support SDHC cards and they had already chosen the processor speed and buffer memory size and speed. The battery drained faster, you could only fit 90-100 shots on the largest size card the camera could use, in DNG only, and only 45-60 in DNG+Jpg and it took forever to write these 16bit files to a card and twice as long as forever to write DNG+Jpg.

Link to post
Share on other sites

actually, it's 14bits, but they discard 2 of them.

What you say is documented in a very interesting article in LFI, 2/2007.

If you are still interested i can send a scanned image of it thru PMs, I don't know if its proper to publish it here...

There is a reason why M8 produces such cool files, and why you shoot underexposed. The sensor is optimized for blacks

 

Please could you read the following:

 

KammaGamma Articles Solving the Leica M8 DNG riddle

 

and let us know how it compares with the LFI article.

 

Regards

Per

Link to post
Share on other sites

Maybe at the time the M8 was developped, the processing of 20 Mbit instead of 10 Mbit files meant such an increase in costs, that given the slight increase in quality it was simply a cost cut rationalisation to go to 8 bits compressed.

 

It would indeed be very nice to compare an uncompressed 14 bit or even 16 bit shot, to a picture compressed to 8 bits.

If somebody can give me a nice raw picture stored in 14 or 16 bit from a well known camera, I can most likely do the compressing to 8 bits and show it to the forum.

 

Hans

 

If you PM your e-mail I can do so, but it will only work if you use Leica's compressing algorithm

Link to post
Share on other sites

Leica's earliest firmware supposedly wrote 16 bit files and although they said they made a comparison and could not see any difference, these were presumably the same people who didn't notice that black fabrics were turning magenta and green leaves yellow...

 

ehhh:D

I believe the same.

Anyway, could have been this reduction of bit from due to the recording speed of the M8.

i.e.

Did they reduce the file size (limiting the bitrate) just to speed up the embarassing writing operation of the M8?

Dunno why, but this seems to me the most probable cause.:rolleyes:

Link to post
Share on other sites

No I don't think so. It was done because Leica had already chosen a battery and they had already chosen NOT to support SDHC cards and they had already chosen the processor speed and buffer memory size and speed. The battery drained faster, you could only fit 90-100 shots on the largest size card the camera could use, in DNG only, and only 45-60 in DNG+Jpg and it took forever to write these 16bit files to a card and twice as long as forever to write DNG+Jpg.

 

Could anyone explain why Adobe's DNG converter can reduce the size of the M8's DNG files quite considerably without loss of information and why Leica does not do this in the first place?

Link to post
Share on other sites

Warrior, if you are asking if the 8-bit files of the M8 are actually bigger than 8 bits, yes they are. The raw image is a 12-bit image that is compressed to 8 bits as it is stored. In the article referred to above, this is discussed. When the raw images are converted by PS or C1, the detail is wonderful.

 

The bits at the upper end are compressed more than the ones at the lower end, and this may account for the unusual shadow detail in the M8. I certainly see much greater shadow detail in the images from my M8 than I ever saw iwith the same lenses on my M6 or M4.

 

If you look around this forum, you will see an emormous number of superb images. Even looking at the jpeg's on the Forum, you can see the detail in the images.

 

If you are testing an M8, you are probably very pleased.

Link to post
Share on other sites

Could anyone explain why Adobe's DNG converter can reduce the size of the M8's DNG files quite considerably without loss of information and why Leica does not do this in the first place?

 

The Adobe converter 'zips' them up losslessly. There's no reason why Leica couldn't adopt this way of writing the DNG files, other than the fact that it would take longer than the existing 8 bit conversion, which I assume uses a simple look up table to determine the output values.

Link to post
Share on other sites

Could anyone explain why Adobe's DNG converter can reduce the size of the M8's DNG files quite considerably without loss of information and why Leica does not do this in the first place?

 

 

Maybe the M8's computing power is not sufficient?

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...