Monochrom M246 Reviews

[Ted Lemon]

Advertisement (gone after registration)

Ditto.  When Leica said they will stand behind the sensor, I immediately made the decision to buy an M-E to go along with my early M Monochrom.  I love using both, and will continue to do so for many years.  If one of them should have sensor issues I have no problem sending it off and using many other Leica cameras while waiting for the repair.  I don't expect that to happen though, as my guess is that sensor failure is far from universal. My M Monochrom is already two years and about 15,000 clicks old.  It has traveled to three continents and been used in all sorts of weather without issue.  I just sent it to Leica before the warranty ended to have it checked out, and while the sensor was dirty, there were no corrosion issues.  I got it back in four weeks, clean and beautiful like the day it was new.

Don't count those chickens. I did the same as you, ME (which was a Leica upgrade exchange/replacement for my coffee stained M8) and MM. The MM (well travelled and used) which is three years old has the corrosion issue, this 6 months from being given a clean bill of health after a service at Leica. I will send it off for replacement but it leaves me with no confidence that this will not happen again. Thats the problem for me  - trust in the camera - its gone.

[jaapv]

We are in the M typ 246 thread here, please do not drag problems with other types into it. Although most if not all of the regular members can differentiate between the models, this might not be the case for outsiders that are not as familiar with all Leica ins and outs, leading to confusion and unfounded rumours. There are plenty of threads where we can discuss the corrosion issues of the M9 family freely.

[Advertisement]

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

[Lenshacker]

Thorsten has provided some High-ISO DNG files for download.

 

I would like to see some low-ISO DNG files for download made available, as were available for the M Monochrom.

 

It would also be very helpful to know the mode in which the camera was operated.

 

The High-ISO DNG files show cross-hatch pattern noise. It would be interesting to know why, and I wonder if using Liveview makes a difference. Liveview requires that the CMOS sensor be operated in a continuous fashion, which will cause temperature to increase. This causes a substantial increase in noise with most detectors. The cross-hatch noise is what I found to be very odd.

 

The M246 produces 12-bit images; the M Monochrom produces 14-bit images. The latter demonstrates Gaussian Noise (random) but does not show signs of fixed pattern noise.

Edited May 4, 2015 by Lenshacker

[menos I M6]

Thorsten has provided some High-ISO DNG files for download.

 

I would like to see some low-ISO DNG files for download made available, as were available for the M Monochrom.

 

It would also be very helpful to know the mode in which the camera was operated.

 

The High-ISO DNG files show cross-hatch pattern noise. It would be interesting to know why, and I wonder if using Liveview makes a difference. Liveview requires that the CMOS sensor be operated in a continuous fashion, which will cause temperature to increase. This causes a substantial increase in noise with most detectors. The cross-hatch noise is what I found to be very odd.

 

The M246 produces 12-bit images; the M Monochrom produces 14-bit images. The latter demonstrates Gaussian Noise (random) but does not show signs of fixed pattern noise.

Actually Brian, the original M Monochrom (which I use since it has been available) shows significant pattern noise when shot at ISO 3200 and above in certain circumstances when developing the DNG files with Adobe RAW converters.

 

To me I never found a reasonable solution for this other than trying to prevent ISO sensitivities above ISO3200 and exclusively use them when I absolutely must.

 

Curiously the very same noise patterns can be found when using very high acute, high resolution scans from silver based B&W films (importing uncompressed tiff scans into Adobe Lightroom).

 

The very only means I have found with Adobe Lightroom counteracting the occurrence of this pattern noise is significantly cranking up the noise reduction and reducing the noise reduction detail preservation in Lightroom (something I usually do not do, as I rather like the noise the MM produces).

 

I would be quite interested in this noise pattern behavior of the new Mono and latest Lightroom revision v6.0.

On a side note I find it rather curious that Leica went for reducing the bit depth on the new model rather then finally boosting bit depth to 16bit (something I am waiting for quite some time for digital 35mm camera makers to implement).

[Lenshacker]

Early on I did some tests with the M Monochrom (delivered to me 12/2012)  using various SD cards. I was able to get very clean images using Sandisk 8GByte 4x memory cards; and was able to introduce Banding into the M Monochrom images using Sandisk Extreme and PNY 10x cards. I bought a lot of the Sandisk 4x 8GByte cards and that is all I use. I have compared ISO 5000 images between the Df (known for High-ISO) and the M Monochrom- about equal quality.

 

My ISO 5,000 and ISO 10,000 images are very clean.

 

I've uploaded Thorsten's example images and my own sample images with the NR and Sharpening Synced for the DNG files here:

 

https://www.flickr.com/photos/90768661@N02/sets/72157652337398501/

 

The images from the M Monochrom appear to be smoother than the limited sample DNG files provided. (I don't want to post images directly in this thread as it might be confusing)

 

(Side note, I HAVE TO GEEZE) 30+ years ago - worked on an early digital infrared sensor that used 15-bit ADC's and data acquisition system. The original system as delivered produced 9 bits of noise in the 15-bit image. With lots of time, effort, and money: the hardware was revised and noise reduced to under 2 bits.

 

Gaussian noise from the sensor is one thing; but systematic noise can usually be chased down, isolated, and eliminated. Sometimes it's just trying things that should make no difference, like the SD card in my M Monochrom or using Liveview in the M246. My technique: change things at random until it works the way I want it to. (not really, but sometimes it seems that way)

 

And if the M246 DNG files had been left in "I"ntel format and not "M"otorola, I would not have to revise my Fortran code to process them. Never know what somebody else's software will do to your image.

Edited May 4, 2015 by Lenshacker

[Lenshacker]

Dirk-

 

If you believe that the Adobe raw convertor is responsible for banding issues, I can post instructions for bypassing it and importing the raw/unprocessed image data into Photoshop using "Open as RAW". All you need to know is the size of the image header, bits-per-sample (16), byte-order of the pixels (Little Endian for M Monochrom), rows and columns stored in the DNG file. I will post it later. I need to look at a few more parameters in the M246 DNG file, will also try to import it without going through the DNG raw convertor. That should allow inspection of "truly raw" images.

 

Brian

Edited May 4, 2015 by Lenshacker

[jaapv]

Advertisement (gone after registration)

Isn't it simpler to use for instance C1 as raw converter and open the TIFF in Photoshop/LR  or ACR as raw?

[Overgaard]

The original 6400 ISO and 12500 ISO sample DNGs can be downloaded from this page:

 

http://www.overgaard.dk/leica-M-Monochrom-Type-246-Digital-Rangefinder-Camera-black-and-white-sensor-page-26-The-Leica-M-246-CMOS-Monochrom.html

 

[mjh]

And if the M246 DNG files had been left in "I"ntel format and not "M"otorola, I would not have to revise my Fortran code to process them. Never know what somebody else's software will do to your image.

According to the DNG specification: “If BitsPerSample is not equal to 8 or 16 or 32, then the bits must be packed into bytes using the TIFF default FillOrder of 1 (big-endian), even if the TIFF file itself uses little-endian byte order.” Looks like Leica didn’t really have a choice with regard to byte order.

[Lenshacker]

What's interesting- when you save the DNG as a DNG using LR6, it puts it into Intel format.

 

The original intent for Big-Endian or Little-Endian was for the Tiff writer to do whichever was most convenient. However- for packing pixel values that do not fall on byte boundaries, a bit order must be specified. Probably a matter of interpretation, but you will find Intel formatted TIFF with packed data. Word and Longword values (more exactly the various equivalent Tiff data types) can still be either big or small endian.  It's easy to unpack the pixels once you know the order, a simple loop with shift operations. Using the values in the header fields requires word-oriented and longword-oriented reversals. I miss the Intel I960 architecture. "Endianness" was an attribute applied to a memory region in the program. You specified which memory regions within the same code had which attribute. So the block of memory used for the TIFF buffer would do all this in hardware. Unfortunately this did not carry over to the newer Intel architectures, would have made life very easy right now. It's not hard, just tedious and getting an offset value incorrect is "bad". I suspect some raw processors that have limited DNG implementations might have problems with this format. Photoshop Raw cannot be used to unpack the 3-nibble data into a 16-bit value. I'm looking at it in HEX and some interesting stuff.... All you really need to unpack a monochrome image is the colums*rows, sample size, and start of the image. It is all stored in one big strip.

Edited May 4, 2015 by Lenshacker

[Lenshacker]

You can tell this is early firmware, Each TAG entry is allowed 12-bytes, not all have to be used. In the M8, M9, M Monochrom- the unused bytes are set to zero. I just tried to look up what Bits Per Pixel being set to '00 0C 88 08' meant. It means the '88 08' is unused junk.

 

Will be interesting to see how this changes and improves.

Edited May 4, 2015 by Lenshacker

[dfarkas]

In case anyone is interested, I've updated my MM246 article with lots of 100% crops at all different ISOs. Just look in the image captions for the links.

 

Leica M Monochrom (Typ 246) Review

Edited May 4, 2015 by dfarkas

[Lenshacker]

The original 6400 ISO and 12500 ISO sample DNGs can be downloaded from this page:

 

http://www.overgaard.dk/leica-M-Monochrom-Type-246-Digital-Rangefinder-Camera-black-and-white-sensor-page-26-The-Leica-M-246-CMOS-Monochrom.html

 

 

Thorsten- these files look like they were saved using Lightroom as opposed to straight out of the camera? 

 

With such early files from the camera, difficult to know what parameters are set by the camera and which ones by Lightroom. So far- the strangest one is the location of the Thumbnail image being at offset '1000'x and the main image at '1e00' (m2468257.dng). The problem is: the thumbnail image takes up more space than is given in the file. So LR, if it was used to save the file, might be "clobbering it". I reserve the right to be wrong! The 12-bit data does start at '1e00'x.

 

DNG files straight from the M Monochrom are things of beauty. In Hex. They are just so well organized and easy to process. And enough space is left for the thumbnail and main image.

Edited May 5, 2015 by Lenshacker

[menos I M6]

Dirk-

 

If you believe that the Adobe raw convertor is responsible for banding issues, I can post instructions for bypassing it and importing the raw/unprocessed image data into Photoshop using "Open as RAW". All you need to know is the size of the image header, bits-per-sample (16), byte-order of the pixels (Little Endian for M Monochrom), rows and columns stored in the DNG file. I will post it later. I need to look at a few more parameters in the M246 DNG file, will also try to import it without going through the DNG raw convertor. That should allow inspection of "truly raw" images.

 

Brian

Thankds Brian, the pattern noise I saw with very high ISO samples in my original M Mono were strange in that absolutely nothing could be found on the internet regarding it's cause or potential solutions.

The only thing back then I could find was very similar looking pattern noise, produced by Adobe produced raw converters (Lightroom and ACR) with specific Canon raw files and it was widely acknowledged that the Adobe converters in the Canon raw files did cause the pattern noise back then.

 

I have come over this noise curiously not in a regular fashion (so it is hard to reproduce) but it came up irregularly and only visible with files of ISO 3200 (starting) and higher.

Specifically bad it would show when in the Adobe raw converters any rotation to the image would have been applied (like rotating the image just a tiny bit to correct a tilted horizon). Then the underlying pattern noise would be immediately visible even in the small previews within the library.

 

I am on a few week trip and might follow up showing a sample once back. I would be interested to find a way to view these specific Mono files in a non Adobe raw converter (not for regular use but just for checking these specific files).

[skimmel]

In case anyone is interested, I've updated my MM246 article with lots of 100% crops at all different ISOs. Just look in the image captions for the links.

 

Leica M Monochrom (Typ 246) Review

Fantastic, thank you so much for adding these.  Pretty remarkable I must say!  Did you see any of the pattern noise that Thorsten's images had?

[asiafish]

Don't count those chickens. I did the same as you, ME (which was a Leica upgrade exchange/replacement for my coffee stained M8) and MM. The MM (well travelled and used) which is three years old has the corrosion issue, this 6 months from being given a clean bill of health after a service at Leica. I will send it off for replacement but it leaves me with no confidence that this will not happen again. Thats the problem for me  - trust in the camera - its gone.

 

My sensor very well might corrode, and if it does I'll send it to Leica for sensor replacement.

[thompsonkirk]

I'm catching up on my reading here because I'm wondering (as are others) about upgrading from M9M to M246.  I'm attracted by Live View for WA landscapes, but my main concern now its whether the base ISO tonal characteristics of the CMOS sensor compare favorably with the older CCD.  I'm further concerned about the 14/12 bit issue.  

 

In reading this thread I came across one post, No. 34 by KM-25, with an example that seemed to cloud the question of what one might be looking for in Monochrom photography with either body.   It was a photograph of a woman cooked with blackened shadows and fried highlights, in a film-style that was popular many years ago. 

 

In continuing the discussion here, I suggest we focus not on how the Monochrom cameras might compare to older film styles, but what they're capable of in their own right.  IMO this is a tonal subtlety that has nothing to do with 35mm Tri-X, and might look something like this – or better, with 246 CMOS? 

 

Texting chef by thompsonkirk, on Flickr

(M9M, 5cm LTM Nikkor)

 

With this in mind, I'm anxious to see more images shot at ISO 320, so we can get a better idea of how the tonal renditions of the M9M and M246 sensors compare.  

 

Kirk

 

PS, many thx to Jono, David, Thorsen, Sean, and Gregory for excellent reviews – in asking for more info/examples about tonal rendering, I didn't mean to criticize their reviews.

Edited May 5, 2015 by thompsonkirk

[dkCambridgeshire]

There's a case for having a separate M Monochrom sub-forum - by itself - to discuss both the original MM and the M246.

 

dunk 

[jonoslack]

Would i be helpful if I put up a link to 4 low ISO DNG files?

[Advertisement]

Hallo,

Look here: Leica M

[Lenshacker]

Would i be helpful if I put up a link to 4 low ISO DNG files?

Please, I would very much like to see these.

 

As there is question about the Lightroom and other Raw convertors, If it is possible to post straight-from-camera DNG files- that would be optimal.