Leica M8 sensor issue? Need your help.

[julian79]

Advertisement (gone after registration)

Hello, I hope you are doing well.

I just got my Leica M8 with a great condition on the outside. 
I have noticed a few things on the sensor due to how old the camera is there is some dust which is not an issue for me.

The thing I am most concerned, is the two vertical lines I get which look like a sensor issue. It is always in the same place, one of the lines is almost in the middle of the frame which makes it a bit annoying if am honest. It does show up on pictures mainly when I have to underexpose which is almost always. 

My question is does anyone have had this issue with the sensor? Is this something that will get worse overtime? 
Even though I love this little camera I can still return it so would you? 

Attaching a photo that shows one of the sensor lines. 

Thank you!

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 May 12 by julian79

[romanus53]

looks like "dead pixel" to me, had this on my M8 years ago. Solution: map it out - either sending it to Leica to make this in software inside (normal for digital) or in post-processing where you can create a corretion-profile in Capture1 (or others but I'm used to this)  which you apply when convertig from dng to jpg.

[Advertisement]

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

[julian79]

Hi and thank you! 
I am used to having dead pixels but randomly scattered on the sensor not so much in a straight line from top to bottom. Should I be worried that this could get worse overtime? 

[chris_livsey]

This is a column defect on the sensor and is common on CCD sensors. When made these sensors usually  have these defects and some "hot" pixels these are mapped out in firmware when the camera is built. Over time cosmic rays can cause further defects and the fix is another re-map there is a limit to how many defects can be mapped out, sometimes it can be done remotely, my P45 CCD back developed this and by submitting a DNG file I received a bespoke firmware upload that worked but I understand that is not always possible. And I don't know if Leica offer that.

It is not a line that is faulty but single pixel that is stuck on maximum but in a CCD sensor  the voltages recorded by each pixel are transferred along the column so that pixel corrupts all the readings that are transferred through it so the line appears. As you have found this is usually at higher ISO or when you push the file in post, again the CCD sensor is "fixed" iso and all other values are achieved by amplification which reveals the fault so processing and high ISO are the same.

I must admit I am puzzled by "mainly when I have to underexpose which is almost always" you should try particularly with a CCD not to underexpose as the files do not take kindly to that regardless of faults.

Leica will re-map, still I assume, it used be around £300-400 but will include a full service and repair guarantee. There is software, Pixel Fix, that can be run on each RAW file to remove it but it is reported as "not always" and needs to be run each time on each file, I have no experience but mention it should you be interested in trying.

If you still have return rights I would do so as this is a repair cost you should not have to stand on a purchase but perhaps if you "have to" underexpose a CCD sensor is not the best choice for a happy life, it's the penalty for getting the CCD colours I'm afraid.

[julian79]

2 hours ago, chris_livsey said:

This is a column defect on the sensor and is common on CCD sensors. When made these sensors usually  have these defects and some "hot" pixels these are mapped out in firmware when the camera is built. Over time cosmic rays can cause further defects and the fix is another re-map there is a limit to how many defects can be mapped out, sometimes it can be done remotely, my P45 CCD back developed this and by submitting a DNG file I received a bespoke firmware upload that worked but I understand that is not always possible. And I don't know if Leica offer that.

It is not a line that is faulty but single pixel that is stuck on maximum but in a CCD sensor  the voltages recorded by each pixel are transferred along the column so that pixel corrupts all the readings that are transferred through it so the line appears. As you have found this is usually at higher ISO or when you push the file in post, again the CCD sensor is "fixed" iso and all other values are achieved by amplification which reveals the fault so processing and high ISO are the same.

I must admit I am puzzled by "mainly when I have to underexpose which is almost always" you should try particularly with a CCD not to underexpose as the files do not take kindly to that regardless of faults.

Leica will re-map, still I assume, it used be around £300-400 but will include a full service and repair guarantee. There is software, Pixel Fix, that can be run on each RAW file to remove it but it is reported as "not always" and needs to be run each time on each file, I have no experience but mention it should you be interested in trying.

If you still have return rights I would do so as this is a repair cost you should not have to stand on a purchase but perhaps if you "have to" underexpose a CCD sensor is not the best choice for a happy life, it's the penalty for getting the CCD colours I'm afraid.

Thank you so much for your time and answer! I do underexpose to try and save the highlights as it seems that the metering and the sensitivity of the CCD sensor don't play well in certain situations. It might sound strange but I manly use M8 for black and white as it seems to be great at it, so am not super bothered with the color capabilities it has. 

[farnz]

In the 16 years I've owned M8's I've always deliberately underexposed by 2/3rds of a stop with exposure compensation to preserve the highlights.  It's the same as removing a 'bias voltage' from a circuit so that you have access to real data instead of it being 'blinded' behind the bias voltage.

The M8's shadows contain a lot of data so underexposing the shot and pulling up in post-processing is fine.  (I apologise if this is contrary to what Chris has written but that's how I and others have always found M8 files to be.) 🤔

Pete.

[chris_livsey]

Advertisement (gone after registration)

On 5/13/2023 at 6:55 PM, farnz said:

The M8's shadows contain a lot of data so underexposing the shot and pulling up in post-processing is fine.  (I apologise if this is contrary to what Chris has written but that's how I and others have always found M8 files to be.) 🤔

Pete.

Not contrary at all, It's what changing the iso in camera with CCD does but can be done with more subtlety by selectively working the shadows but doing that has a limit. The noise in the grass example shows that but again more modern processing is better at removing that, remember the processing when the M8 was new was a lot cruder than now available making the files more usable. I'm a little more conservative with a third of a stop "protection" dialled in but overall the dynamic range is limited compared to a "modern" sensor.

As a side note if shooting B/W leave off the UV/IR filter if you are not doing already that produces a noticeably "better" file at the expense, if you suddenly need colour, of accuracy.

 

[CGarrard]

On 5/15/2023 at 12:00 AM, chris_livsey said:

Not contrary at all, It's what changing the iso in camera with CCD does but can be done with more subtlety by selectively working the shadows but doing that has a limit. The noise in the grass example shows that but again more modern processing is better at removing that, remember the processing when the M8 was new was a lot cruder than now available making the files more usable. I'm a little more conservative with a third of a stop "protection" dialled in but overall the dynamic range is limited compared to a "modern" sensor.

As a side note if shooting B/W leave off the UV/IR filter if you are not doing already that produces a noticeably "better" file at the expense, if you suddenly need colour, of accuracy.

 

Agree completely on leaving off the UV/IR filter. Also that is what gives the M8 its unique color.

What a lot of people don't understand, are the benefits of a weaker IR filter. If you have never shot with an IR converted camera before or know much about IR photography, then I can understand  the concern. However, IR light  helps to make for much more attractive mid tones because IR light penetrates cellular surfaces, which includes skin, leaves, vegetation, etc... and reflects back.

And in doing so it creates a bit of a glow (the intensity varies of course on how much light there is), and it does this in shadow areas as well. The overall effect to the image is much more balanced midtones (think of it as fill light almost) than a standard IR protected sensor, and that's one of the reasons why the M8 has such good monochrome output- in fact, it's better than Monochrom cameras imho by far.

Also the unique color is due to the same effect, and one of the reasons why the M8 looks closer to a film than other cameras, even other cameras with CCD sensors that benefit from inherent CCD qualities.

Leica did a poor job defending the M8, it wasn't an oversight the M8's sensor was designed this way, it was intentional in order to produce a look to the images. I particularly like the M8's color. And you can have other cameras converted to get close to the output of the M8 as well, if you work with a shop that does IR conversions such as Spencer's in Utah. 

Edited September 16 by CGarrard

[farnz]

7 hours ago, CGarrard said:

Leica did a poor job defending the M8, it wasn't an oversight the M8's sensor was designed this way, it was intentional in order to produce a look to the images.

I think you'll find that it was a 'positive unintended consequence' rather than by design.  I know Jono Slack quite well who was one of Leica's original field testers who noticed the magenta blacks and yellowish foliage colours owing to excess IR bleed-through and reported it to Leica who were very concerned.  At the time Stefan Daniel admitted that Leica had had to agree to a very thin IR filter cover on the sensor owing to the minimal room allowed by the lens register (the distance between the rear lens element and the sensor or film) of 27.8 mm and the thin filter couldn't stop all the wavelengths in the IR spectrum - above about 720 nm.  Leica's fix was to offer two free UV/IR circular filters to M8 customers. 

Leica designed a permanent fix for the M9 by designing a 'stronger' IR-stopping sensor cover.  The M9's Kodak KAF18500 sensor is simply a larger variant of the M8's KAF10500 and the lens register is unchanged in the M9 of course (and in the M3, M2, M1, M5 ... M240, M10, M11 etc) so the sensor cover Leica designed for the M9 would have worked on the M8 with reduced dimensions.

That the M8's colours are so appealing is no more than dumb luck but that's just fine with me.🙂

Pete.

[CGarrard]

On 9/16/2023 at 12:33 PM, farnz said:

I think you'll find that it was a 'positive unintended consequence' rather than by design.  I know Jono Slack quite well who was one of Leica's original field testers who noticed the magenta blacks and yellowish foliage colours owing to excess IR bleed-through and reported it to Leica who were very concerned.  At the time Stefan Daniel admitted that Leica had had to agree to a very thin IR filter cover on the sensor owing to the minimal room allowed by the lens register (the distance between the rear lens element and the sensor or film) of 27.8 mm and the thin filter couldn't stop all the wavelengths in the IR spectrum - above about 720 nm.  Leica's fix was to offer two free UV/IR circular filters to M8 customers. 

Leica designed a permanent fix for the M9 by designing a 'stronger' IR-stopping sensor cover.  The M9's Kodak KAF18500 sensor is simply a larger variant of the M8's KAF10500 and the lens register is unchanged in the M9 of course (and in the M3, M2, M1, M5 ... M240, M10, M11 etc) so the sensor cover Leica designed for the M9 would have worked on the M8 with reduced dimensions.

That the M8's colours are so appealing is no more than dumb luck but that's just fine with me.🙂

Pete.

Hi Farnz,

Good afternoon to you.

While I would love to accept every part of your reply as simple fact and just move on, I really just have a hard time doing so. Before I go further, I don't doubt you know Jono Slack or that Stefan agreed to the final form of the sensor- or that there is any intentional misinformation in your reply. That said, there was a PR issue to contain and there could be various reasons for the answer Stephan gave to the press about it. But believing that the M8's output is simply "blind luck" is too far of a stretch that, with all due respect,  I just can't accept.

I will elaborate for the audience of this post.

I have a hard time believing that the thinness of the IR filter is due to minimal space for lens register. Afterall, the register point is at the surface of the chip itself, at the pixel level, and the IR filter is just a cover on top of the chip. It doesn't matter what the thickness of the IR filter is, at least for that reason, there are other reasons to consider the thickness of the cover of IR filter glass but lens focus register isn't one of them.

Also the thickness of the sensor had to be considered in the design of the body from the outset, way before it was shown in completed form- the implication in his statement that engineers simply miscalculated for that measurement and made an error, is extremely unlikely, but quite believable on the surface to anyone outside of the industry.

Then you have to consider the prototype phase of the M8 in which it was tested. I'm sure engineers were quite well aware of its color output well before the M8 went into production. For most images you can see how the weak IR filter has a light fill flash effect (because of how brighter than normal IR light penetrates cellular surfaces and reflect back to the camera), especially on shadows. This effectively creates more pleasing mid-tones- the side effect of which are non-clinically accurate colors. A fair tradeoff if you are to emulate film's massive dynamic range in comparison to the sensors of the time.

*The M8 was the first digital M camera since the film era, and I'm quite certain there was a huge emphasis on creating images that closely resembled film as possible. And its images do that better than any digital camera I've used to date (and I believe the key ingredient is the introduction of more IR light into the image, intentionally).*

A more likely story is that Kodak would be to blame here (if one had to blame) for any fumbles with the IR filter on the sensor itself.  They were notorious for issues with their CCD's, not only with Leica but other manufacturers as well. Simultaneously they were the primary manufacturer that were picked by way of contract with Leica.

Then you have to know that engineers tested that sensor well before it even went into a body. And they knew/were quite aware of its dimensions and the output of the sensor itself. I would be looking at Kodak before Leica if I were to blame anyone for a "mishap", it simply makes the most sense if you are going down that route.

However I don't think it was a mishap at all. The corroded sensor of the M9, that my friend was a mishap, and a huge one.

Also in the end, Leica chose to produce the M8 in the final form they brought to the public. Not once, but TWICE. The M8.2 used the same sensor with it's weak IR filter as the previous model that came before it. 

Lastly, I am quite aware of the similarities of the Kodak chips for the APS-H and 35mm variants. The second one had the real issue imho (due to its cover glass adhesion often referred to as "corrosion"), and its the sole reason why I don't own an M9 now. I did  have one before that had a chip replaced, I sold it off prior to learning that the replacement chips were also corroding, thankfully. I'll never own an M9 again, there should have been a class action suit taken against Kodak by Leica for this problem.

Anyways, those are my thoughts.

And I'll leave on one last note, where I move the subject to the fact that the M8 (or 8.2 in my case) can be shot all the way up to 2,500 ISO if you have the patience to understand how to expose and post process correctly.

Here is a picture of my M Typ 262 with the new lovely grip I found for it (matches perfectly), serendipitous photo-bomb puts the cherry on top.


 

 

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 September 22 by CGarrard

[farnz]

1 hour ago, CGarrard said:

I have a hard time believing that the thinness of the IR filter is due to minimal space for lens register.

I have no desire to turn this into a blow-by-blow drag-out argument so I gracefully accept what you have written. 

You might have a hard time believing that the space available owing to the minimal lens register is a cause of the thinness of the M8's particularly thin IR filter (whose efficiency was thereby reduced) and I accept that.  The available space was one factor but the other factor that I didn't mention was the proximity to the sensor of the rear element and the exit pupil of rectilinear lenses that would be used with the camera.  You're doubtless aware of Leica's proud history of excellent wide-angle rectilinear lenses but the close proximity of the exit pupil meant that light rays at the edges of the exit pupil will strike the sensor at a highly oblique angle - hence the need for off-set pixels at the edges of the KAF10500 and KAF10800 sensors at Leica's request.  (This was new to Kodak and I understand that a lot of research and development cost went into producing the only sensor with offset pixels to minimise vignetting owing to collecting less radiant energy from the oblique angle.  The fact that Kodak managed to achieve it is kudos to them.)

If a normal (thick) IR filter layer was placed on top of the sensor then the oblique rays would need to travel through more glass than just the thickness of the filter and thereby some attenuation would take place and more vignetting introduced.  I understand that to prevent this vignetting Leica opted for the thinnest IR filter they could use and only discovered during field testing that more IR than they had hoped was leaking through.  I doubt that factory testing would have picked it up because factory technicians are unlikely to be photographers and would be unlikely to notice the IR-caused colour shifts that the field testers (all experienced photographers) would.

I hope this illustrates the reason for the thin IR filter and is sufficient to resolve this side discussion. 🙂

Pete.

[CGarrard]

5 hours ago, farnz said:

I have no desire to turn this into a blow-by-blow drag-out argument so I gracefully accept what you have written. 

 

I wasn't nor am I looking for such an argument. I thought we were having a fine respectful conversation and sharing knowledge.

[farnz]

5 minutes ago, CGarrard said:

I wasn't nor am I looking for such an argument. I thought we were having a fine respectful conversation and sharing knowledge.

Very glad to hear it; sometimes it can be hard to tell online.

Pete.

[CGarrard]

Understandable!

[Borna]

I have that too on my M8 in certain lights. I think it mimics scratched film 🤠

[chris_livsey]

In the further interest of shared knowledge I posted this some time ago after much, probably too much, search engine use, the frequency data is from the manufacturers data sheet, IR we tend to forget is quite a broad range of frequencies and the filters are variable across the spectrum.

The M8 series have an 0.5 mm thick IR filter that is about 50% effective overall (it varies across the range, see below) the M9 a 0.8 mm of the same glass which is about 90% effective. The trade off is the thicker filter means that as different wavelengths of light are affected differently by the IR filter depth so the M8 at pixel level appears sharper, it resolves better, the effect though is relatively slight and the downside is it is more prone to moire. The filter is Kyocera BS7 glass which is relatively less efficient for a given thickness than more recent products and actually thinner glass that was more efficient was in use by eg Nikon even at the time of the M8, around 0.3mm, but their AA filters bulked up the total filter pack thickness. The other issue is the filtration range of the BS7, it is very inefficient blocking IR in the 1000 to 1200 nm range but much better in the 700 to 1000nm range.