Jump to content

What about a BW-M8?


eckart

Recommended Posts

Advertisement (gone after registration)

Explain please how you are increasing the resolution of a 10.3mp sensor simply by removing the Bayer filter array... Each pixel of the M8’s 10.3mp is represented by full color data points in the final image file. When you print your image, you are printing the full 10.3mp stretched across your print media, unless you crop or interpolate up to a higher pixel count. Removing the filter array will only produce a 10.3mp monochrome sensor with the same resolution of 3916 x 2634 pixels for DNG (10.31 MP), or 3936 x 2630 pixels (10.35 MP) for JPG. The only way to increase resolution of a given sensor is to artificially increase the number of pixels by interpolation. Where are you finding the extra 10.3mp to effectively double the resolution?

To re-iterate, the number of pixels doesn’t increase – the effective resolution does.

 

The point is that an X megapixel sensor with Bayer filters cannot deliver X megapixels of resolution. Typically, the luminance channel is derived mainly from the 50 percent of green-sensitive pixels while the red- and blue-sensitive pixels are mainly used for adding color, and much less for increasing effective resolution. Since you are missing two thirds of the full set of RGB data for each pixel, you need to interpolate the missing data – in other words, you need to guess. You can go for a safe bet, i.e. conservatively assume no steeper change of luminance or color than the available data-points suggest, but then you will have to accept a lower effective resolution – 1/4 of the megapixel count if you are very conservative, but usually about 1/2. Alternatively, you can go for maximum resolution, but that requires rather bold guesses that will often be wrong, resulting in color moiré. Of course, if you are using an AA filter, you don’t even have that choice. Anyway, you cannot achieve an effective resolution equivalent to the number of sensor pixels unless you also accept moiré. For the M8 I had measured an effective resolution of 1840 lines per image height, corresponding to a 5.05 megapixel image. Note that this was measured with a b&w chart; the resolution of color (especially red&blue) patterns would be even lower.

 

So, the point of a b&w sensor isn’t that it would somehow, magically, increase the number of pixels, but that it produces images worthy of its pixel count – something a color sensor with Bayer filters cannot.

 

Perhaps you are suggesting an increase in dynamic range resulting from the filter array removal?

Not at all.

Link to post
Share on other sites

  • Replies 74
  • Created
  • Last Reply

Top Posters In This Topic

Top Posters In This Topic

Posted Images

Second, a lot of M-photographers in the world do not want colours, they prefer a perfect B&W.

 

You state this as fact - I don't accept it as fact unless you have some figures to back it up. Even the oldsters here reminisce about the beauty of their Kodachromes shot with an M3 (or even an M7).

 

The following photographers shoot lots of Leica M color:

 

Nan Goldin - Wikipedia, the free encyclopedia

WILLIAM EGGLESTON

Susan Meiselas

Unpublished American Color

David Alan Harvey

Stuart Franklin Photography

Alex Webb // Rebecca Norris Webb // Photographs

 

------

 

Similarly, I do not accept as FACT that a monochrome M8, still 10.3 megapixels, would have more than a tiny (2-5%) resolution advantage over the current M8. Below is a 400% crop from an M8 image (135 Tele-Elmar). The lens and M8 Bayer sensor are resolving chair wires 1 pixel wide. How can that level of resolution be improved upon - evidence, please.

 

-------

 

Finally, note that ANY rectangular pattern of pixels - including monochrome pixels - may show moire patterns. Not color fringing, but if you photograph a screen door or ventilation grid or other regular repeating texture that comes close to the pixel pitch, it will creat moire artifacts, even on a monochrome sensor.

 

The most common tools used to demonstrate moire patterns are, in fact, often monochrome patterns. Color is not required: File:MoireGrid.png - Wikipedia, the free encyclopedia

 

So - I dispute the premise that most Leica M photographers prefer B&W. I present evidence. Prove me wrong.

 

I dispute the premise that a monochrome sensor will provide a significant increase in detail and resolution. I present evidence. Prove me wrong.

 

I dispute the premise that a monochrome sensor will be free from moire effects. I present evidence. Prove me wrong.

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

So - I dispute the premise that most Leica M photographers prefer B&W. I present evidence. Prove me wrong.

 

I dispute the premise that a monochrome sensor will provide a significant increase in detail and resolution. I present evidence. Prove me wrong.

 

I dispute the premise that a monochrome sensor will be free from moire effects. I present evidence. Prove me wrong.

 

Why does everything have to be so logical (and, frankly, so boring)? Let's just say that some of us would like a B&W-only M8. Extra resolution and sensitivity might be nice but I'd like one most of all for the simplicity of not having to deal with colour when I want to shoot in B&W.

Link to post
Share on other sites

Ian, If you read up on your Orwell, (or just study the last US administration - or a lot of bloggers) you'll discover how easy it is to disguise things in vague and imprecise language and pseudo-science - or in theory over empirical evidence. I've just decided not to stand for that anymore.

 

I accept your contruction without hesitation - some Leica photographers would prefer a B&W-only camera. This thread alone supports that simple proposition.

 

Note that I don't dispute the ISO advantage - I've worked enough with color-separation filters to know how much light they eat.

 

-----

 

Tim, true - but than ANY B&W image is free from certain color moire effects - thin lines that barber-pole from red to cyan, for example. They both turn to gray in a monochrome conversion. I'm willing to allow a monochrome sensor that 2-5% for the few color moires that will still produce visible artifacts (as demonstated in practical shooting) in B&W conversions from an RGB M8.

Edited by adan
Link to post
Share on other sites

No doubt Andy, there are a lot M-photographers and as you examples famous ones as well, who loves the colour performance of the M-lenses with film or with the M8 and the hopefully coming future M9.

I do not want to give my M8 away and I love the results.

But, there are great amount of M-photographers who prefer B&W and a lot more as in any other brand.

In the first 20 years of my "M-career" you would have never found another film than a Tri-X in my M3, than M2 or later M4-2. This kind of freaks are not this seldom, are they?

The Mono-M I am thinking about is just a additional body to make extraordinary B&W, neither to be instead of the All-waiting-of-M9 nor to replace my M8.

You ask for a proof of the better performance...how could I? But when you read Michael's post above, can't you follow his thoughts?

Edited by eckart
Link to post
Share on other sites

Advertisement (gone after registration)

Eckart - I'd agree with you that Leica M photographers are probably more likely to happily give up color even in digital photography than the general population of photographers. I wouldn't say "most".

 

I follow Michael's thoughts - I just don't think they are supported by the real-world evidence. It is true that if you could open a RAW file from a Bayer sensor camera directly as a TIFF, with no conversions applied, you would see a B&W image with a strong checkerboard pattern imposed by the effects of subject colors on the RGB pixels. Similarly, if you ran Kodachrome through regular B&W chemicals, you'd get a nasty-looking B&W negative.

 

The processing is key - and RAW processing algorthims separate the luminance and chrominance calculations such that, absent an anti-alias filter, significantly more than the theoretical amount of luminance (B&W) resolution can be recaptured.

 

In certain specific cases, which I estimate to occur in about 2-5% of images, something like a person in a red-and-green-striped shirt, at just the right size in the image for the stripes to react with the Bayer checkerboard, would be less accurately represented, and the stripes would appear as a jumbled texture rather than stripes. And in that small part - of a small number of images - then , yes, the monochrome sensor would just about double the resolution of the Bayer sensor.

 

Over the other 98% of the image, it would give no advantage at all. IMHO

Edited by adan
Link to post
Share on other sites

Similarly, I do not accept as FACT that a monochrome M8, still 10.3 megapixels, would have more than a tiny (2-5%) resolution advantage over the current M8. Below is a 400% crop from an M8 image (135 Tele-Elmar). The lens and M8 Bayer sensor are resolving chair wires 1 pixel wide. How can that level of resolution be improved upon - evidence, please.

 

 

Sorry Adan, but I tend to accept as a FACT that a sensor with no Bayer filter has a resolution advantage : fact is that any pixel stored in M8 RAW file (and in any other usual digicamera) is de facto computed as an interpolation of 4 adiacent pixel; for me, this is an evidence that the resolution with no Bayer filter is indeed higher. Of course, in a REAL image (not a test chart) this can be less or more significant in terms of "perceived resolution", but the fact remains.

Edited by luigi bertolotti
Link to post
Share on other sites

but than ANY B&W image is free from certain color moire effects - thin lines that barber-pole from red to cyan, for example. They both turn to gray in a monochrome conversion.

Any black&white conversion will turn color moiré into black&white moiré, but it will do nothing to restore the image detail that was lost to moiré in the first place. The only way to preserve detail is to prevent moiré; when the damage is done, there is no cure.

Link to post
Share on other sites

To re-iterate, the number of pixels doesn’t increase – the effective resolution does.

 

The point is that an X megapixel sensor with Bayer filters cannot deliver X megapixels of resolution. Typically, the luminance channel is derived mainly from the 50 percent of green-sensitive pixels while the red- and blue-sensitive pixels are mainly used for adding color, and much less for increasing effective resolution. Since you are missing two thirds of the full set of RGB data for each pixel, you need to interpolate the missing data – in other words, you need to guess. You can go for a safe bet, i.e. conservatively assume no steeper change of luminance or color than the available data-points suggest, but then you will have to accept a lower effective resolution – 1/4 of the megapixel count if you are very conservative, but usually about 1/2. Alternatively, you can go for maximum resolution, but that requires rather bold guesses that will often be wrong, resulting in color moiré. Of course, if you are using an AA filter, you don’t even have that choice. Anyway, you cannot achieve an effective resolution equivalent to the number of sensor pixels unless you also accept moiré. For the M8 I had measured an effective resolution of 1840 lines per image height, corresponding to a 5.05 megapixel image. Note that this was measured with a b&w chart; the resolution of color (especially red&blue) patterns would be even lower.

 

So, the point of a b&w sensor isn’t that it would somehow, magically, increase the number of pixels, but that it produces images worthy of its pixel count – something a color sensor with Bayer filters cannot.

 

Michael,

 

I appreciate your tenacity in making your point, but I believe either we have a language barrier, or we do not use the same terms regarding digital imaging. You can theoretically discuss effective resolution, or spatial resolution, but as I see it, there is only one resolution that a digital sensor can physically output. In the case of the M8 it is 10.3mp. When I print a 3916 x 2634 pixel image on any printer at 300ppi, I will get an image 13 inches by 8.78 inches containing 10.3ppi on a piece of paper. That is real practical resolution, not subjective resolution or measurements of individual pixels made in Lab mode.

 

Having said that, I have printed much larger prints from an M8 that in no way resemble a 2.5mp or 5.05mp capture. I am not as interested in resolution charts (color or B&W) as I am in photographs – captured, converted from Raw files, and rendered on paper. I print images far better than the M8's 10.3mp would suggest and compare more than favorably with higher resolution cameras that I also use.

 

My response to you is based on personal and practical experience with digital photography going back to its earliest stages so I know what 2.5 and 5mp images look like. I have used sensors with and without the AA filters you mentioned. Currently all the cameras I use have no AA filter or have removable ones, so I am well aware of the toll an AA filter places on image quality.

 

Moiré does not affect only color images; it is a function of the Nyquist frequency, which also bites B&W images.

 

I will gladly concede that under certain conditions in a lab, with the right kind of glass, etc., you may come up with a barely discernable difference that will not show on a print hanging on a wall. A pixel is a pixel, and for all practical purposes, no one would see a 30%, 50%, 100%, or 200% increase in resolution in an image taken without a Bayer filter compared to an image taken with one using the same sensor.

 

One final point worth repeating, the Raw image converter used will have a far greater impact on image quality and what you are referring to as resolution than most people realize. On-board camera Raw converters are only adequate to produce JPEG thumbnails; this goes for any camera, not just Leica.

 

Perhaps we are not disagreeing, but rather approaching image quality from different perspectives…

Link to post
Share on other sites

Pixels are not pixels, if you take a picture with a cheap camera with a small 14 MP sensor you will have a picture with a size of 14 MP, no matter what the lens have "seen" and the same picture made with the 10,3 MP M8 will show you a lot more details, no matter that the size is smaller.

Simply said, if some colours need to be described more than one pixel, than this pixels are missing to discribe the picture, compared to a monochrome sensor where every pixel just show the intensity of the light.

Link to post
Share on other sites

There was a serious rumor back in the early days of the M8 of a B&W version - serious enough that Popflash was taking order positions. Nothing has come of it and I doubt that it will this late in the M8 product cycle.

 

Not only a rumour,as Brett (bybrett) said in a earlier post, Leica asks him what he is thinking about before the release of the M8.

But now the question is not mono or colour, now it would be colour and mono

Link to post
Share on other sites

Pixels are not pixels, if you take a picture with a cheap camera with a small 14 MP sensor you will have a picture with a size of 14 MP, no matter what the lens have "seen" and the same picture made with the 10,3 MP M8 will show you a lot more details, no matter that the size is smaller.

Simply said, if some colours need to be described more than one pixel, than this pixels are missing to discribe the picture, compared to a monochrome sensor where every pixel just show the intensity of the light.

 

You've lost me here. If your response is to my post, then you did not read it. Where did I mention "14 MP" or the physical size of a pixel?

 

We could go on about pixel density, and compare how many microns across the photosites are, but unless I missed something, this thread was about the M8, not a cell phone or some P&S camera. Now, when you start comparing P&S photosites to the M8's, you are obviously going to see some large quality disparity.

 

I never said a higher pixel count equates to a better image producing sensor, it only produces higher resolution. If the higher pixel count sensor has photosites the same size as the M8 (which would of course mean a physically larger sensor), then it has the potential for a higher resolution image of equal quality as the M8.

 

We were discussing the photosites (pixels) of the M8 and my point is pixels are the measure of resolution and the addition or removal of a Bayer filter array does not change the number of pixels on a sensor or its resolution (which still remains at 10.3mp with or without a Bayer filter).

Edited by nzav
typo correction (duplicate word)
Link to post
Share on other sites

You've lost me here. If your response is to my post, then you did not read it. Where did I mention "14 MP" or the physical size of a pixel? ......We were discussing the photosites (pixels) of the M8 and my point is pixels are the measure of resolution and the addition or removal of a Bayer filter array does not change the number of pixels on a sensor or its resolution (which still remains at 10.3mp with or without a Bayer filter).

 

 

You did not mention a cheap camera with 14 MP, but I did.

What I tried to declare with my example is, that the size of the file says nothing about the details.

The amount of the pixels, in your example, stay the same, nobody deny that, 10,3 MP stay 10,3 MP, but the possibilities of describing details are changing with the different use of them.

Link to post
Share on other sites

i would love a b&w M lens compatible camera. The best looking photos ive ever taken are still those i did years ago on my Contax G2 which only ever had b&w film in it before i moved to digital - and ive been trying to create that same look and feel ever since! I must say that the colour photos from the M8 that i convert to b&w are the closest ive seen to what my G2 produced - so a b&w sensor would be wonderful.

 

That was a great camera. I went on a lot before selling it hoping for a digital back to come.

And yes, the M8 is the one getting closer to that B&W feeling (sometimes even better).

Link to post
Share on other sites

I dispute the premise that a monochrome sensor will provide a significant increase in detail and resolution. I present evidence. Prove me wrong.

 

I dispute the premise that a monochrome sensor will be free from moire effects. I present evidence. Prove me wrong.

 

Andy,

 

You are entirely correct that a monochrome sensor is not free of sampling artifacts. However, a mono sensor, for the reason Michael discusses in his posts, can have a anti-aliasing filter which is "lighter" by a factor of at least two than a Bayer sensor for the same level of susceptibility to artifacts. So there is the potential for an increase in effective resolution.

 

Sandy

Link to post
Share on other sites

Maybe when Leica announces the M9 (FF), they can offer a sensor upgrade to a true M8-BW so we can be all happy. I would go for it...

Edited by esquire53
Link to post
Share on other sites

You can theoretically discuss effective resolution, or spatial resolution, but as I see it, there is only one resolution that a digital sensor can physically output. In the case of the M8 it is 10.3mp. When I print a 3916 x 2634 pixel image on any printer at 300ppi, I will get an image 13 inches by 8.78 inches containing 10.3ppi on a piece of paper.

Yes, and when you take a photograph with a sub-standard lens, don’t focus correctly, or get motion blur from camera shake, what do these 10.3 million pixels buy you in terms of resolution? What if you trade your M8 for some cheap compact digicam that also happens to be equipped with a 10.3 megapixel sensor? Do you think 10.3 megapixel is 10.3 megapixels, resolution-wise?

 

Effective resolution is the power to resolve fine detail and it is measured in lines or line pairs per image height (or mm), not pixels. That’s where different cameras and lenses differ, even when their megapixel count is the same. If you do not care for effective resolution, what do you care for?

Edited by mjh
Link to post
Share on other sites

There was a serious rumor back in the early days of the M8 of a B&W version

Leica had indeed been considering such an option back then. I suppose they came to the conclusion that the market was too small for a b&w model, but whatever the reason, they eventually decided against it. Not knowing how many units you would have to order from Kodak and at what price, it is difficult to argue for or against the commercial viability of such a product, its technical viability notwithstanding.

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...