Jump to content

Highlight rolloff too abrupt?


Recommended Posts

Posted (edited)

Advertisement (gone after registration)

I thought that it must be my poor exposure / post processing skills that led to ugly holes in cloud ☁️ highlights, and the like, with the SL2. But I see that @jonoslack

https://www.slack.co.uk/leica-m10-r.html

https://www.slack.co.uk/m10-highlights.html

has reported a similar sounding phenomenon and there have been rumours of adding a highlight priority exposure mode in later firmware, presumably because others have complained. 

I have no idea whether this is a characteristic of the sensor, the camera’s image processing circuitry (in an attempt to maximise apparent dynamic range) or Adobe Lightroom’s tone curve is a poor fit for the sensor characteristics. Anyway, it’s a bit of a pain to have to underexpose images by 2/3 stop to keep cloud highlights, which don’t show a significant tonal discontinuity in real life, from generating cutout- like holes. (Yes, I use the blinkies and Natural rendering.)

What does the panel recommend as a remedy?

Edited by jrp
Link to post
Share on other sites

Expose without overexposing, most Leica Camera benefit from under exposing highlights 

I had my M10-P and M10M set to -1 stop all the times. It is good to lean how to best use you camera and sensor.

The M10-R is a bit better in highlights, I am exposing normal, in some cases I fail it down.

 

Link to post
Share on other sites

I do try, but I have no desire to underexpose for the sake of preserving specular or unimportant highlights. Trouble is, the difference between the overexposed highlights and their neighbours seems cartoonishly hard. I don’t remember this phenomenon with other cameras that I have used.

Link to post
Share on other sites
5 hours ago, jrp said:

I thought that it must be my poor exposure / post processing skills that led to ugly holes in cloud ☁️ highlights, and the like, with the SL2. But I see that @jonoslack

https://www.slack.co.uk/leica-m10-r.html

https://www.slack.co.uk/m10-highlights.html

has reported a similar sounding phenomenon and there have been rumours of adding a highlight priority exposure mode in later firmware, presumably because others have complained. 

I have no idea whether this is a characteristic of the sensor, the camera’s image processing circuitry (in an attempt to maximise apparent dynamic range) or Adobe Lightroom’s tone curve is a poor fit for the sensor characteristics. Anyway, it’s a bit of a pain to have to underexpose images by 2/3 stop to keep cloud highlights, which don’t show a significant tonal discontinuity in real life, from generating cutout- like holes. (Yes, I use the blinkies and Natural rendering.)

What does the panel recommend as a remedy?

Well, I don't think it's an issue with the SL2, and I don't think it is with the M10-R either, the M10 certainly did (does) have a problem at 100 ISO, but it's mostly sorted by shooting at 200 ISO. 

Shooting in high contrast situations you really do need to make sure you don't overexpose the highlights! But that's the case with any camera, and as you can easily recover shadow detail in lightroom I don't see that it's much of a problem

All the best

Link to post
Share on other sites

Advertisement (gone after registration)

I have found that at base ISO of 50 on the SL2 you will lose considerable highlight details but you do gain on shadow detail/noise over ISO 100 and above. 

Sensor does not have quite the same shadow detail recovery I have gotten used to from Sony. 
 

Most times I will bracket the exposure to three shots. 
 

Paul 

Link to post
Share on other sites

Digital Cinema cameras typically have their "native" ISO between ISO 400 to ISO 800, delivering the most extensive dynamic range. Why should that be different with still cameras? On the SL2-S, I figure that at ISO 800, I get the most out of the sensor without introducing distracting noise. I would never expect the camera to deliver great dynamic range at ISO 100, let alone ISO 50. The sensor will receive too much light at such low ISO, making it hard to compensate since its native sensitivity is about 3-4 stops higher. As Jono pointed out, there's tons of information buried in the shadows, which can easily be lifted to the desired level. The same can be said about the highlights. That way, one takes advantage of the vast dynamic range the sensor offers. 

Link to post
Share on other sites
5 hours ago, hansvons said:

Digital Cinema cameras typically have their "native" ISO between ISO 400 to ISO 800, delivering the most extensive dynamic range. Why should that be different with still cameras? On the SL2-S, I figure that at ISO 800, I get the most out of the sensor without introducing distracting noise. I would never expect the camera to deliver great dynamic range at ISO 100, let alone ISO 50. The sensor will receive too much light at such low ISO, making it hard to compensate since its native sensitivity is about 3-4 stops higher. As Jono pointed out, there's tons of information buried in the shadows, which can easily be lifted to the desired level. The same can be said about the highlights. That way, one takes advantage of the vast dynamic range the sensor offers. 

Senors usually have their maximum DR at base ISO or very close to it.  This is the first time I hear that sensor DR is higher at higher ISO settings

Link to post
Share on other sites



On most (if not all) sensitive surfaces (that was valid for motion picutre film as well), DR reduces AND shifts with higher values from Base sensitivity.
 

On low values (lets say 100 on SL2), most of the available DR will "be found in shadows" to put it simply. At 400 (on SL2 - from my experiences) it is more evenly repartited... and the higher you go paradoxically the more leeway to have in highligh recovery availability.

So, about the low iso 50 controvoersy on the SL2, yes, it does clip highlight faster for the same exposure value than iso 100. But if you preserve your highlights "in camera" by switcing your exposure to the left --- just enought to save those highlights (that is when the clipping feature in stills or zebra in video becomes very useful) --- you do get a bit of extra information recover details from the underexposed part of your image.

If you fancy, the rapid fire HDR works very well for extreme constrasts, even handheld.

Link to post
Share on other sites
2 hours ago, RM70 said:

Senors usually have their maximum DR at base ISO or very close to it.  This is the first time I hear that sensor DR is higher at higher ISO settings

The ARRI ALEXA's CMOS Super-35mm sensor is rated at 2.8K and ISO 800. That sensitivity allows the camera to see a full seven stops of over exposure and another seven stops of underexposure. To take advantage of this, ARRI offers both industry-standard REC709 HD video output as well as the Log-C mode that shows the entire range of the chip's sensitivity, allowing for an extreme range of color correction options in post.https://en.wikipedia.org/wiki/Arri_Alexa

That was ten years ago. Regarding ISO, not much has changed since. The Alexas of today are rated the same. A Red's Monstro sensor is in the same ISO ballpark, so is the Sony Venice's sensor. And from my experience, the Leica SL2-S' pictures look best for me at ISO 800 too. Of course, you can rate these cameras lower, but you will get less DR in the highlights.

What a base-ISO does mean in digital stills photography is an interesting question. I'm pretty sure base-ISO doesn't provide the largest DR or the least noise. I suspect it to be an ISO rating at the lowest meaningful ISO numbers, which the camera can compensate for in average situations. I've learned that stills photographers like to shoot outside at low ISO numbers and want their cameras to cope with ISO 100 as well as with ISO 6400 and more.

Cinematographers, however, select the ISO according to the environment but only within a short range of one or two stops. In studio situations where light is well controlled, a lower ISO can make sense (that'd typically be ISO 400 with the Alexa). In an outside environment like a sunny park where blown highlights and deep shadows are all over the place, the best possible DR is required. Typically, with the Alexa, that'd be ISO 800, which matches my experiences with the SL2-S. Another critical factor is the texture, which is nothing else as a bit of noise the sensor exhibits. Most cinematographers are after texture giving the pictures an organic feel and a less video look. Low ISO numbers defeat texture and make photos look more digital. Conversely, in noise-prone situation such outdoor scenes at night, a lower ISO (e.g. ISO 400) is quite likely the right choice

I always dial noise reduction down to zero. I like to see the sensor's results as pure as possible. YMMV, of course.

 

Link to post
Share on other sites

From what you can deduct from some PhotonsToPhotos analisys for stills and my observation for video.... on the SL2 (not S), it appears we may have some sort of dual base iso going on. 400iso and 3200 for the SL2 seems to be about right.

For the SL2s, and some cross observation with panasonic S1/S5/S1H... it seems indeed to be about 640/800 and 4000/6400...

Link to post
Share on other sites
Posted (edited)
10 minutes ago, Slender said:

From what you can deduct from some PhotonsToPhotos analisys for stills and my observation for video.... on the SL2 (not S), it appears we may have some sort of dual base iso going on. 400iso and 3200 for the SL2 seems to be about right.

For the SL2s, and some cross observation with panasonic S1/S5/S1H... it seems indeed to be about 640/800 and 4000/6400...

Agreed. Makes perfectly sense. Sensors with smaller pixels deliver higher resolution at a given sensor size but are less sensitive to light. The dual ISO thing, however, remains to me a myth that somewhat seems to be real but needs some more understanding on my side. 

Edited by hansvons
Link to post
Share on other sites
2 hours ago, hansvons said:

Agreed. Makes perfectly sense. Sensors with smaller pixels deliver higher resolution at a given sensor size but are less sensitive to light. The dual ISO thing, however, remains to me a myth that somewhat seems to be real but needs some more understanding on my side. 

What do you mean by "less sensitive to light"? It is typically the sensor size not the pixel size that determines the amount of light collected.

Link to post
Share on other sites
2 hours ago, hansvons said:

The ARRI ALEXA's CMOS Super-35mm sensor is rated at 2.8K and ISO 800. That sensitivity allows the camera to see a full seven stops of over exposure and another seven stops of underexposure. To take advantage of this, ARRI offers both industry-standard REC709 HD video output as well as the Log-C mode that shows the entire range of the chip's sensitivity, allowing for an extreme range of color correction options in post.https://en.wikipedia.org/wiki/Arri_Alexa

That was ten years ago. Regarding ISO, not much has changed since. The Alexas of today are rated the same. A Red's Monstro sensor is in the same ISO ballpark, so is the Sony Venice's sensor. And from my experience, the Leica SL2-S' pictures look best for me at ISO 800 too. Of course, you can rate these cameras lower, but you will get less DR in the highlights.

What a base-ISO does mean in digital stills photography is an interesting question. I'm pretty sure base-ISO doesn't provide the largest DR or the least noise. I suspect it to be an ISO rating at the lowest meaningful ISO numbers, which the camera can compensate for in average situations. I've learned that stills photographers like to shoot outside at low ISO numbers and want their cameras to cope with ISO 100 as well as with ISO 6400 and more.

Cinematographers, however, select the ISO according to the environment but only within a short range of one or two stops. In studio situations where light is well controlled, a lower ISO can make sense (that'd typically be ISO 400 with the Alexa). In an outside environment like a sunny park where blown highlights and deep shadows are all over the place, the best possible DR is required. Typically, with the Alexa, that'd be ISO 800, which matches my experiences with the SL2-S. Another critical factor is the texture, which is nothing else as a bit of noise the sensor exhibits. Most cinematographers are after texture giving the pictures an organic feel and a less video look. Low ISO numbers defeat texture and make photos look more digital. Conversely, in noise-prone situation such outdoor scenes at night, a lower ISO (e.g. ISO 400) is quite likely the right choice

I always dial noise reduction down to zero. I like to see the sensor's results as pure as possible. YMMV, of course.

 

In still photography, base or native ISO means the ISO where the signals is at largest. This also means the ISO where SNR is the largest, which means that the DR is largest.

Link to post
Share on other sites
Posted (edited)
41 minutes ago, SrMi said:

What do you mean by "less sensitive to light"? It is typically the sensor size not the pixel size that determines the amount of light collected.

Yes and no in the sense that pixels are separtated units, each with its own borders and determined surface.... 4.3 square um on SL2, for example, and 6um2 on SL2s. So for a simmilar exposure time, each "larger" pixel logically harvests a bit more photons each, regardless the size of the whole sensor.
So I assume you could safely say that a the 24MP@6microns of the 35mm SL2s sensor are arguably "more sensitive" than a 100MP phase one at 4.63 microns each... Despite the fact that the actual surface of the things technically harvest more light.

And inversely I imagine it is safe to assume a 61MP FF sensor would fare worse on individual pixel level than an older.... lets say 36MP medium format sensor?

With film all grains of the emulsion are kind of like bonded together.... whereas pixel stand very close together, but nonetheless clearly separated in their own individual pixel wells. That's the whole "analogue" curve vs "digital" 0&1 on a physical level I imagine.

Conversely, even if a 47mp image gets comparatively "noisier" at higher iso setting: smaller pixel also mean smaller grain. It gets even better if you oversample your final image from 47mp to 24, for example. You gain some more sharpness and grain becomes even less noticeable.

Edited by Slender
Link to post
Share on other sites
3 minutes ago, Slender said:

Yes and no in the sense that pixels are separtated units, each with its own borders and determined surface.... 4.3 square um on SL2, for example, and 6um2 on SL2s. So for a simmilar exposure time, each "larger" pixel logically harvests a bit more photons each, regardless the size of the whole sensor.

With film all grains of the emulsion are kind of like bonded together.... whereas pixel stand very close together, but nonetheless clearly separated in their own individual pixel wells. that's the whole "analogue" curve vs "digital" 0&1 on a physical level I imagine.

Conversely, even if a 47mp image gets comparatively "noisier" at higher iso setting: smaller pixel also mean smaller grain. It gets even better if you oversample your final image from 47mp to 24, for example. You gain some more sharpness and grain becomes even less noticeable.

To simplify: small and large pixel cameras (same sensor size and technology) have outputs with very similar noise characteristic when compared at the same output size. 

Link to post
Share on other sites
Just now, SrMi said:

To simplify: small and large pixel cameras (same sensor size and technology) have outputs with very similar noise characteristic when compared at the same output size. 

Yes for grain/sharpness in my experience. DR/details loss with very high iso will not really benefit from oversampling.

Link to post
Share on other sites
3 hours ago, hansvons said:

 The dual ISO thing, however, remains to me a myth that somewhat seems to be real but needs some more understanding on my side. 

I dont know (and likely will never) for sure how Leica processes for this. But from my understanding of this, a dual native iso camera has actually two amplification circuit to process the sensor signal. One rated for a low value (say 100iso) and that will use higher and higher amplification current for the next values (200....400...800...1600...) until the second circuit kicks in for its own "base value" (say 3200). It will then work a bit harder for 6400.... harder still for 12500....etc.

Like a manual gearbox car: you can drive up to 50kph in 2nd gear.... but when you switch to 3rd or 4th the engine works much less hard for the same result.... kind of.

In effect, on a recent camera like sony FX6, it shows that ISO 6400 looks worse than 12800.... because that is when the second circuit kicks into action.

@hansvons to find out on your SL2s, turn to video mode, engage focus peaking to RED and HIGH sensitivity and point at a dark subject and underexpose your image. Scroll through iso values and perhaps you will see a threshold when the random peaking dots highlinthing excess lowlight grain rises before diminishing again. It certainly does on the SL2 going from ASA 1600 (grainy) to ASA3200 (less grainy). It could be "scalling" like photonstophotos seems to indicate for stills... but the results are the same: you are better off shooting the SL2 at 3200 in low light than 1600ASA.
Alternatively if you have a Ninja you can use the "blue channel" tool to analyze grain ---- but its a bit like shining a torch through a lens ---- if you don't know what to expect you may have panic attaks lol---- there is always dust/grain 😅

Link to post
Share on other sites
5 minutes ago, Slender said:

I dont know (and likely will never) for sure how Leica processes for this. But from my understanding of this, a dual native iso camera has actually two amplification circuit to process the sensor signal. One rated for a low value (say 100iso) and that will use higher and higher amplification current for the next values (200....400...800...1600...) until the second circuit kicks in for its own "base value" (say 3200). It will then work a bit harder for 6400.... harder still for 12500....etc.

Like a manual gearbox car: you can drive up to 50kph in 2nd gear.... but when you switch to 3rd or 4th the engine works much less hard for the same result.... kind of.

In effect, on a recent camera like sony FX6, it shows that ISO 6400 looks worse than 12800.... because that is when the second circuit kicks into action. You car get less noisy lol 🤪for the same speed.

@hansvons to find out on your SL2s, turn to video mode, engage focus peaking to RED and HIGH sensitivity and point at a dark subject and underexpose your image. Scroll through iso values and perhaps you will see a threshold when the random peaking dots highlinthing excess lowlight grain rises before diminishing again. It certainly does on the SL2 going from ASA 1600 (grainy) to ASA3200 (less grainy). It could be "scalling" like photonstophotos seems to indicate for stills... but the results are the same: you are better off shooting the SL2 at 3200 in low light than 1600ASA.
Alternatively if you have a Ninja you can use the "blue channel" tool to analyze grain ---- but its a bit like shining a torch through a lens ---- if you don't know what to expect you may have panic attaks lol---- there is always dust/grain 😅

 

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.

Loading...
  • Recently Browsing   0 members

    No registered users viewing this page.

×
×
  • Create New...