 |
 |
|
 |
|
 |
|  |
|  |
|
 |
|
 |
|  |
|  |
|
 |
"Fredrik Eriksson" <fe79}--at--{yahoo}--dot--{com> wrote:
> On Thu, 06 May 2010 00:24:07 +0200, Darren New <dne### [at] san rr com> wrote:
>>
>> Second, if you're talking about the ISO setting rather than the
>> exposure, that's about how much you pre-bias the electrons in the
>> sensor. Basically, you load up each pixel of the sensor with some
>> electrons, and if light kicks out an electron, you add one to the
>> intensity of the light there. Adding more electrons makes it easier to
>> get kicked out.
>
> No, there is no pre-bias; that is not how the sensors work. The ISO
> setting only affects amplification of the analogue sensor output.
There you go. So if you have the raw sensor output, you can change the ISO
setting after the fact?
Post a reply to this message
|
 |
|  |
|  |
|
 |
|
 |
|  |
|  |
|
 |
Fredrik Eriksson wrote:
> On Thu, 06 May 2010 00:56:35 +0200, Jim Henderson <nos### [at] nospam com>
> wrote:
>>
>> I've actually wondered this myself - raw editing software gives you the
>> option to adjust the exposure; obviously, you can't pull details out
>> doing this that are completely washed out or completely underexposed, but
>> it is possible to bring additional detail out by making changes to the
>> exposure setting (ev) after the photo has been taken.
>>
>> I've done it, so clearly it's possible, I just don't understand the math
>> behind it.
>
> When the camera creates a JPEG file, part of the raw data is discarded and
> the rest is then adjusted to fit in the 0-255 range. The camera picks a
> black-point somewhere near the low end of the (raw) range, and all values
> below that are made black (i.e. 0) in the JPEG. Similarly, a white-point
> is selected, and all values above that are made white (i.e. 255). When you
> edit a RAW file, you have access to the parts that would otherwise have
> been cut off. If the exposure was somewhat less than perfect, there can be
> useful data in those parts.
In other words, the RAW format gets you High Dynamic Range.
Post a reply to this message
|
 |
|  |
|  |
|
 |
|
 |
|  |
|  |
|
 |
Invisible wrote:
> scott wrote:
>> However pros use the raw sensor data from the camera and not a JPEG.
>> This allows them some margin to adjust the exposure later without adding
>> any artifacts to the final JPEG image they create. Because of this it
>> is extremely important not to saturate the sensor (ie 100% white) in any
>> areas, it is impossible to get back detail in areas that are at 100%.
>> If you under-expose it you can scale up the brightness without
>> introducing artifacts (because usually the raw sensor data is higher bit
>> depth than JPEG).
>
> ...in other words, you're not changing the exposure (i.e., the number of
> seconds that the shutter opens) at all, you're simulating it.
In consumer cameras that show the image on the LCD screen all the time,
there is no "shutter opening for N milliseconds". The sensor is getting
light all the time.
Post a reply to this message
|
 |
|  |
|  |
|
 |
|
 |
|  |
|  |
|
 |
On Thu, 06 May 2010 23:09:18 +0200, Nicolas Alvarez
<nic### [at] gmail com> wrote:
> "Fredrik Eriksson" <fe79}--at--{yahoo}--dot--{com> wrote:
>> No, there is no pre-bias; that is not how the sensors work. The ISO
>> setting only affects amplification of the analogue sensor output.
>
> There you go. So if you have the raw sensor output, you can change the
> ISO setting after the fact?
Not quite. The amplification controlled by the ISO setting occurs before
A/D conversion. By applying amplification to the raw data (i.e. after A/D
conversion), you also amplify any noise introduced by the A/D process. The
result is similar, but slightly worse.
--
FE
Post a reply to this message
|
 |
|  |
|  |
|
 |
|
 |
|  |
|  |
|
 |
On Thu, 06 May 2010 23:10:25 +0200, Nicolas Alvarez
<nic### [at] gmail com> wrote:
>
> In other words, the RAW format gets you High Dynamic Range.
Sort of. Higher than what you get from a JPEG anyway. Not what is usually
referred to as HDR though.
--
FE
Post a reply to this message
|
 |
|  |
|  |
|
 |
|
 |
|  |
|  |
|
 |
>> ...in other words, you're not changing the exposure (i.e., the number of
>> seconds that the shutter opens) at all, you're simulating it.
>
> In consumer cameras that show the image on the LCD screen all the time,
> there is no "shutter opening for N milliseconds". The sensor is getting
> light all the time.
Sure. But when you take an image, it integrates the sensor output for
the duration of X seconds. You can't unintegrate it, or reintegrate over
a different time period, for example. (Unless you got it to shoot a few
trillion RAW images and save them all...)
Post a reply to this message
|
 |
|  |
|  |
|
 |
|
 |
|  |
|  |
|
 |
> My camera (a Canon PowerShot S50) lets me see what the exposure will be
> without taking a picture - so I can adjust it before even shooting. I
> typically set the value to -1/3 because I can use raw tools to lighten it
> up (that idea of not overexposing the image that someone else mentioned).
1. My camera doesn't offer RAW format anyway.
2. Isn't under-exposure equally as bad as over-exposure?
Post a reply to this message
|
 |
|  |
|  |
|
 |
|
 |
|  |
|  |
|
 |
On Fri, 07 May 2010 09:56:50 +0200, Invisible <voi### [at] dev null> wrote:
>
> 1. My camera doesn't offer RAW format anyway.
The same technique can be applied to JPEG as well, but you will run into
posterization much sooner.
> 2. Isn't under-exposure equally as bad as over-exposure?
A slight under-exposure can usually be compensated for in post-processing
without too much quality loss. With over-exposure, some detail is
generally lost.
--
FE
Post a reply to this message
|
 |
|  |
|  |
|
 |
|
 |
|  |
|  |
|
 |
>> 1. My camera doesn't offer RAW format anyway.
>
> The same technique can be applied to JPEG as well, but you will run into
> posterization much sooner.
Indeed, I quickly discovered that it's usually hopeless to try.
>> 2. Isn't under-exposure equally as bad as over-exposure?
>
> A slight under-exposure can usually be compensated for in
> post-processing without too much quality loss. With over-exposure, some
> detail is generally lost.
A slight anything is better than a drastic something. ;-)
If the image is too under-exposed, bits of it will be pure black. Same
problem as if it's over-exposed.
Post a reply to this message
|
 |
|  |
|  |
|
 |
|
 |
|  |
|  |
|
 |
Invisible wrote:
> If the image is too under-exposed, bits of it will be pure black.
Almost never.
--
Darren New, San Diego CA, USA (PST)
Ada - the programming language trying to avoid
you literally shooting yourself in the foot.
Post a reply to this message
|
 |
|  |
|  |
|
 |
|
 |
|  |