ISO 50

[sanj]

@Neuro:
Neuro:
[/quote]
It's only half the photons if the aperture is set a stop narrower or the shutter speed is set a stop faster.
[/quote]

And not if ISO is reduced by a stop? Does the sensor not reject half the photons when it is 'asked' to be less sensitive?

Dont pounce on me, am learning here...

I can easily see that the amount of light reaching the sensor is a function of shutter/aperture combination. But am confused after that. I thought sensor reacts to the available light based on what ISO is set.
Thx...

 

[rpt]

@Neuro:
Neuro:

It's only half the photons if the aperture is set a stop narrower or the shutter speed is set a stop faster.

And not if ISO is reduced by a stop? Does the sensor not reject half the photons when it is 'asked' to be less sensitive?

Dont pounce on me, am learning here...

I can easily see that the amount of light reaching the sensor is a function of shutter/aperture combination. But am confused after that. I thought sensor reacts to the available light based on what ISO is set.
Thx...

The ISO set determines the analog gain applied to the signal that resulted after photons are captured in the photo-site.

 

[Loswr]

And not if ISO is reduced by a stop? Does the sensor not reject half the photons when it is 'asked' to be less sensitive?

Dont pounce on me, am learning here...

I can easily see that the amount of light reaching the sensor is a function of shutter/aperture combination. But am confused after that. I thought sensor reacts to the available light based on what ISO is set.
Thx...

No one should be pounced on for trying to learn!

The sensor has no control over the number of photons it collects. The maximum capacity is determined by the design of the sensor. Only aperture and shutter open time determine the amount of photons collected during the exposure (up to the max capacity of a photosite, which is when highlight clipping becomes unavoidable). After an exposure, each photosite is 'read out' and at that point, analog gain (amplification) is applied to the signal according to the ISO set. In other words, every exposure has the photons being collected at "base ISO" (which is different for different for different sensor designs, but is usually somewhere in the 60 to 200 range). Following that capture of photons, and conversion to do electrons by the photo site, the signal has been applied (positive or negative, (within the "native" range of ISOs), then the analog signal is converted to a digital signal, and if necessary based on the selected ISO, additional digital gain is applied (for the "expanded" ISOs).

 

[dlleno]

Neuro wrote: Sorry, but you seem to be falling into the same trap and suffering from the same misconceptions as Mikael.

No it is about words, and what we mean.

And where is my earlier post?

HTP is nothing else than while the camera metering for 200iso the gain is smaller (around 100iso gain) and thereby we get a head room. ( I describe this as under exposure) because the camera metering after 200iso)
You get the same effect by under exposing 100iso 1 stop and later correct the raw file in the raw converter
In Jpg the camera lay a smoother contrast curve which make a smoother high light reproduction and also lift little bit in lower levels which make the noise little bit visible.
I hope I have make this message readable
Have a nice day

pardon my intrusion here Mikael, but the words and concepts you are using in the above is not under question. WHat is under question is the posts where you say that the HTP causes half the number of photons to reach the sensor.

 

[rpt]

Neuro wrote: Sorry, but you seem to be falling into the same trap and suffering from the same misconceptions as Mikael.

No it is about words, and what we mean.

And where is my earlier post?

HTP is nothing else than while the camera metering for 200iso the gain is smaller (around 100iso gain) and thereby we get a head room. ( I describe this as under exposure) because the camera metering after 200iso)
You get the same effect by under exposing 100iso 1 stop and later correct the raw file in the raw converter
In Jpg the camera lay a smoother contrast curve which make a smoother high light reproduction and also lift little bit in lower levels which make the noise little bit visible.
I hope I have make this message readable
Have a nice day

pardon my intrusion here Mikael, but the words and concepts you are using in the above is not under question. WHat is under question is the posts where you say that the HTP causes half the number of photons to reach the sensor.

Yes Mikael, from my perspective you need to review you post on the first page where you mention this. You then say that results in half the charged electrons. That is not right. The same number of photons would be collected in the photo site irrespective if you has set HTP or not.

I hope you understand this. Afterwords, after all the photons are collected and a signal is generated, things happen resulting in what you describe in that post as "half the charged electrons"...

If you disagree with what I have written, do let me know what you disagree with.

Thanks.

 

[Loswr]

And where is my earlier post?

The one where you made a derogatory remark directed at me? No doubt it was deleted by a mod.

No it is about words, and what we mean.

HTP is nothing else than while the camera metering for 200iso the gain is smaller (around 100iso gain) and thereby we get a head room. ( I describe this as an under exposure) because the camera is metering after 200iso)
You get the same effect by under exposing 100iso 1 stop and later correct the raw file in the raw converter
In Jpg the camera lay a smoother contrast curve which make a smoother high light reproduction and also lift little bit in lower levels which make the noise little bit more visible.
I hope I have make this message readable so every one can understand

Yes, that's essentially (or maybe exactly) what you've posted previously, and it's clear that you understand the consequences of enabling HTP.

But it is about words, and many times you used words describing a 'halving of light', 'shorter exposure', and 'the sensor collects less photons' as part of the HTP process, all of which are false. When that was brought to your attention, you simply restated the same incorrect information. So while you certainly understand the results of HTP, it's not at all clear that you understand the underlying mechanism, which has nothing to do with altering the amount of light hitting the sensor. Looking at the responses above, it's evident that I'm not the only one who is aware that you made incorrect statements about the underlying mechanism by which HTP works (i.e., it does not change exposure per se, but only the gain applied).

Regardless, it's clear that you have not acknowledged that you provided wrong information about that mechanism, and I suspect you're unlikely to do so. That's rather ironic, given that you earlier accused others of being unable to admit their mistakes. But as I previously stated, I'm not at all surprised by that...

 

[dlleno]

Mikael here is the confusion

1. Your latest post is correct, i.e.

HTP is nothing else than while the camera metering for 200iso the gain is smaller (around 100iso gain) and thereby we get a head room

2. But the words you have used previously are quite clear, and quite incorrect, namely about the amount of light hitting the sensor during HTP operation.

mechanism behind HTP In a contrasty sunny day and at base iso 100iso there are no head room and the proximity to clippning is small.
By HTP there will be a under exposure to gain 1 stop of high lights. This is the same as let the sensor be hit by the less amount of photons as compared to 100iso and numbers of created electrons just before readout are therefore less.
and there are now 1 stop more of head room up to clipping.
The sensor collect photons nothing else and some of them will be in to electrons, there are also some gain going on.
To get the effect , lesser photons/less electrons and to get 1 stop of high light head room the signal must be lower than it is at 100iso , this can be done by shorter exposure time, change f-stop, earlier read out. + effect head room - effect more noise in lower levels.
There is no way at 100iso when the sensels are fully charged/near to clipping can get a negative amplification and therby get 1 stop head room .
Since we call 100 iso (near fully charged cell) 100iso and 200 iso 200 iso , the electron charge has been halved at 200iso and the gain has increased.
In HTP the sensor has now been hit by less light/photons = that we called 1 stop under exposure and fewer electrons has been read out which later are adjusted in the raw converter by gain and different curve
This is the same as underexpose 100 iso 1 stop= 1 stop faster shutter speed or 1 more F-stop to get 1 stop more head room and then adjust the raw file and make a own smother curve at the top.

It would be most helpful if you would rationalize the statements you have made that appear to conflict with each other. If your understanding has progressed, and you no longer believe that HTP causes a change in the number of photons hitting the sensor, then please say that. Instead, your English is really quite good and to me you are trying to write as an authority, which invites us to put your statements side by side.

 

[Marsu42]

And where is my earlier post?

The one where you made a derogatory remark directed at me? No doubt it was deleted by a mod.

But it was just getting entertaining, I really liked neuro's "Can you not be insulting, boy?" post :->

So while you certainly understand the results of HTP, it's not at all clear that you understand the underlying mechanism, which has nothing to do with altering the amount of light hitting the sensor.

Ok, to be on topic: I doubt Canon is very keen on people knowing what htp actually does because it sounds like "fix limited sensor dr at no cost": the latest manual on the 6d simply says on p.92: "You can minimize overexposed highlight areas" and that's certainly a good thing, isn't it?

I remember because when I asked what htp is good for I first got flamed for not simply reading the manual, people obviously don't even realize there is something to ask about concerning iso50/htp... http://www.canonrumors.com/forum/index.php?topic=5687.msg108754#msg108754

 

[TheSuede]

Mikael here is the confusion

1. Your latest post is correct, i.e.

HTP is nothing else than while the camera metering for 200iso the gain is smaller (around 100iso gain) and thereby we get a head room

2. But the words you have used previously are quite clear, and quite incorrect, namely about the amount of light hitting the sensor during HTP operation.

mechanism behind HTP In a contrasty sunny day and at base iso 100iso there are no head room and the proximity to clippning is small.
By HTP there will be a under exposure to gain 1 stop of high lights. This is the same as let the sensor be hit by the less amount of photons as compared to 100iso and numbers of created electrons just before readout are therefore less.
and there are now 1 stop more of head room up to clipping.
The sensor collect photons nothing else and some of them will be in to electrons, there are also some gain going on.
To get the effect , lesser photons/less electrons and to get 1 stop of high light head room the signal must be lower than it is at 100iso , this can be done by shorter exposure time, change f-stop, earlier read out. + effect head room - effect more noise in lower levels.
There is no way at 100iso when the sensels are fully charged/near to clipping can get a negative amplification and therby get 1 stop head room .
Since we call 100 iso (near fully charged cell) 100iso and 200 iso 200 iso , the electron charge has been halved at 200iso and the gain has increased.
In HTP the sensor has now been hit by less light/photons = that we called 1 stop under exposure and fewer electrons has been read out which later are adjusted in the raw converter by gain and different curve
This is the same as underexpose 100 iso 1 stop= 1 stop faster shutter speed or 1 more F-stop to get 1 stop more head room and then adjust the raw file and make a own smother curve at the top.

It would be most helpful if you would rationalize the statements you have made that appear to conflict with each other. If your understanding has progressed, and you no longer believe that HTP causes a change in the number of photons hitting the sensor, then please say that. Instead, your English is really quite good and to me you are trying to write as an authority, which invites us to put your statements side by side.

Now does those quote chains get long, damn...
Mikael should handle his own words, but in the text you quote in the second part, the words "ISO" and "100" are sprinkled all over the place. I think the main bulk of the confusion lies here. At ISO100, the exposure that's normally set by the camera in "automatic" mode approaches the physical charge storage limit of the sensor. Hence, the ISO50 fake setting is close to useless in reality for anyone that shoots raw (but not for people that shoot jpg - Canon's jpg-in-camera contrast curve does normally leave quite a lot of image material close to blown white out of the jpg - so there's still room to nudge the image brightness a bit more in post before saving the jpg)

What happens at higher ISOs (than 100) and with HTP is another story.
Since the actual aperture and shutter speed does not change when activating HTP, photometric exposure stays constant - i.e no loss of photons, and also hence no increase in natural Poisson noise. The only loss you get is that you boost the relative levels of electronic noise and read noise per gray level by one stop by first halving conversion ratio and then boosting that in software.
When you halve conversion ratio (ISO) the electronic noise and RN still stays the same

 

[dlleno]

good point -- as neuro suggested its the special case that got promoted to the general case thats confusing. you're right lets let Mikael explain what he really means by "the mechanism of HTP", halving of photons, and "HTP is nothing more than ..." right now, whether due to a lanuage problem, not understanding the difference between an electron and a photon, differences in the behavior of HTP at ISO 100 versus 200 whatever... the explanations are a curious mixure of truth and error without a clear delineation of which is which.

 

[rpt]

(Snip!)

What happens at higher ISOs (than 100) and with HTP is another story.
Since the actual aperture and shutter speed does not change when activating HTP, photometric exposure stays constant - i.e no loss of photons...

(Snip!)

Dang! Silly me, I thought Mikael had found a way to set his Canon camera to HTP and control how the Sun streamed Photons towards his sensor!

You just burst my bubble.

 

[RLPhoto]

This thread just shows how much of us could really use a "True" ISO 50 in DSLR's. I'd like one canon and perhaps they will be the first with the new 1Dxs.

 

[dlleno]

This thread just shows how much of us could really use a "True" ISO 50 in DSLR's. I'd like one canon and perhaps they will be the first with the new 1Dxs.

+1 to that. perhaps the new crop of sensors may have that capability to appear on the big MP body that exists only in the imagination right now

 

[Loswr]

This thread just shows how much of us could really use a "True" ISO 50 in DSLR's. I'd like one canon and perhaps they will be the first with the new 1Dxs.

The problem is that there's no free lunch. A "true" ISO 50 would mean a lower base ISO, meaning that to achieve higher ISOs, even more amplification would be needed - meaning more high ISO noise. Usually, if ISO 100 is not low enough, one stop more is insufficient, at least in terms of shutter speed. The waterfall example posted earlier at ISO 50 and 5 s exposure required f/18 to get there - personally, I'd have preferred to shoot that at f/9 and ISO 100 with a 3-stop ND.

 

[RLPhoto]

This thread just shows how much of us could really use a "True" ISO 50 in DSLR's. I'd like one canon and perhaps they will be the first with the new 1Dxs.

The problem is that there's no free lunch. A "true" ISO 50 would mean a lower base ISO, meaning that to achieve higher ISOs, even more amplification would be needed - meaning more high ISO noise. Usually, if ISO 100 is not low enough, one stop more is insufficient, at least in terms of shutter speed. The waterfall example posted earlier at ISO 50 and 5 s exposure required f/18 to get there - personally, I'd have preferred to shoot that at f/9 and ISO 100 with a 3-stop ND.

The 1Dxs would be a 46mp+, pure studio camera from practical use from 50-800 ISO. IE: the S designation.

It would return back to its true roots as a studio camera.

 

[R1-7D]

Neuro, I initially joined this Forum because of you. Your knowledge on all this camera equipment blows my mind. This thread is another great educational experience.

 

[rpt]

This thread just shows how much of us could really use a "True" ISO 50 in DSLR's. I'd like one canon and perhaps they will be the first with the new 1Dxs.

The problem is that there's no free lunch. A "true" ISO 50 would mean a lower base ISO, meaning that to achieve higher ISOs, even more amplification would be needed - meaning more high ISO noise. Usually, if ISO 100 is not low enough, one stop more is insufficient, at least in terms of shutter speed. The waterfall example posted earlier at ISO 50 and 5 s exposure required f/18 to get there - personally, I'd have preferred to shoot that at f/9 and ISO 100 with a 3-stop ND.

My two cents (and I am not disclosing the currency) is that to get a "true" ISO 100 if we need 14 bits, for a "true" ISO 50 we will need 15 bits. I am willing to bet that if you wait a bit longer, we will get native ISO 25 (how did I figure that? Well, 16 bits...). Now Einstein said time is relative so I guess the "bit" he was talking about was not a 0 or a 1...

 

[Aglet]

Mikael here is the confusion

1. Your latest post is correct, i.e.

HTP is nothing else than while the camera metering for 200iso the gain is smaller (around 100iso gain) and thereby we get a head room

2. But the words you have used previously are quite clear, and quite incorrect, namely about the amount of light hitting the sensor during HTP operation.

mechanism behind HTP In a contrasty sunny day and at base iso 100iso there are no head room and the proximity to clippning is small.
By HTP there will be a under exposure to gain 1 stop of high lights. This is the same as let the sensor be hit by the less amount of photons as compared to 100iso and numbers of created electrons just before readout are therefore less.
and there are now 1 stop more of head room up to clipping.
The sensor collect photons nothing else and some of them will be in to electrons, there are also some gain going on.
To get the effect , lesser photons/less electrons and to get 1 stop of high light head room the signal must be lower than it is at 100iso , this can be done by shorter exposure time, change f-stop, earlier read out. + effect head room - effect more noise in lower levels.
There is no way at 100iso when the sensels are fully charged/near to clipping can get a negative amplification and therby get 1 stop head room .
Since we call 100 iso (near fully charged cell) 100iso and 200 iso 200 iso , the electron charge has been halved at 200iso and the gain has increased.
In HTP the sensor has now been hit by less light/photons = that we called 1 stop under exposure and fewer electrons has been read out which later are adjusted in the raw converter by gain and different curve
This is the same as underexpose 100 iso 1 stop= 1 stop faster shutter speed or 1 more F-stop to get 1 stop more head room and then adjust the raw file and make a own smother curve at the top.

It would be most helpful if you would rationalize the statements you have made that appear to conflict with each other. If your understanding has progressed, and you no longer believe that HTP causes a change in the number of photons hitting the sensor, then please say that. Instead, your English is really quite good and to me you are trying to write as an authority, which invites us to put your statements side by side.

Maybe it's all the foreign language manuals, schematics and engineers in person that I have to decipher..
.. because despite the imperfect english grammar and syntax, I understood perfectly well what Michael stated in that quote above.
Now it's important to many of us with tech-geek credentials to use the most correct terminology possible to describe something properly, lest we be misunderstood.. or worse. I'm willing to cut people plenty of linguistic slack as they try to describe something in a language as convoluted as english, whether it's their first language or not.

ISO 50 = YES, wish i had it available as a real ISO at times.

BTW, as an extra, i shots some crude tests with my UN-beloved 5D2 last night.
MIDTONE BANDING at iso 100 - it still has it! Real-world photos are where I first found the problem, specific test shots certainly replicate it. I need to do a few more tests just to make sure this is not a glitch of my display calibration curve but I'm pretty sure it isn't as I can accentuate the pattern with a simple unsharp function in PS.
Should i start a new topic with that when I get a chance?

 

[Marsu42]

BTW, as an extra, i shots some crude tests with my UN-beloved 5D2 last night.
MIDTONE BANDING at iso 100 - it still has it! Real-world photos are where I first found the problem, specific test shots certainly replicate it. I need to do a few more tests just to make sure this is not a glitch of my display calibration curve but I'm pretty sure it isn't as I can accentuate the pattern with a simple unsharp function in PS.
Should i start a new topic with that when I get a chance?

Sure - bashing the 5d2 will make the 6d look better, and I'm desperate for that since this is most probably the ff camera I can/want to afford over the 5d3

 

[J.R.]

Should i start a new topic with that when I get a chance?

Yes please ... I'm interested in this

Cheers ... J.R.