Picture quality...

On Wed, 30 Apr 2014 18:13:54 +0200, android wrote:

In article ,
nospam wrote:

In article , android
wrote:

the raw format is proprietary which means you're stuck with sigma's
(very) slow and crappy software. there are one or two other raw
converters that reverse engineered the format but didn't do a
complete job so the results aren't as good.

You can ding some of the DPs files and use a multitude of converters.
https://www.adobe.com/products/photo...#dngcompatible

those are the old dp series cameras with the 4.6mp sensor, and not even
all of them since the sd15 isn't even included.

none of the recent 15mp 'merrill' sensors are included.

The Merrils ain't that interesting since they do not have true three
color pixels.

sigma started encrypting the data and doesn't offer support to anyone.

Too bad...

The list of supported cameras for dcraw includes:


Sigma SD9
Sigma SD10
Sigma SD14
Sigma SD15
Sigma SD1
Sigma SD1 Merill
Sigma DP1
Sigma DP1 Merill
Sigma DP1S
Sigma DP1X
Sigma DP2
Sigma DP2 Merill
Sigma DP2S
Sigma DP2X

In article ,
ray carter wrote:

On Wed, 30 Apr 2014 18:13:54 +0200, android wrote:

In article ,
nospam wrote:

In article , android
wrote:

the raw format is proprietary which means you're stuck with sigma's
(very) slow and crappy software. there are one or two other raw
converters that reverse engineered the format but didn't do a
complete job so the results aren't as good.

You can ding some of the DPs files and use a multitude of converters.
https://www.adobe.com/products/photo...#dngcompatible

those are the old dp series cameras with the 4.6mp sensor, and not even
all of them since the sd15 isn't even included.

none of the recent 15mp 'merrill' sensors are included.

The Merrils ain't that interesting since they do not have true three
color pixels.

sigma started encrypting the data and doesn't offer support to anyone.

Too bad...

The list of supported cameras for dcraw includes:


Sigma SD9
Sigma SD10
Sigma SD14
Sigma SD15
Sigma SD1
Sigma SD1 Merill
Sigma DP1
Sigma DP1 Merill
Sigma DP1S
Sigma DP1X
Sigma DP2
Sigma DP2 Merill
Sigma DP2S
Sigma DP2X

That would make Lightzone, Darktable and Rawtherapie compatible if you
dare use Indie converters...
--
teleportation kills
http://tinyurl.com/androidphotography

In article ,
ray carter wrote:

On Wed, 30 Apr 2014 18:13:54 +0200, android wrote:

In article ,
nospam wrote:

In article , android
wrote:

the raw format is proprietary which means you're stuck with sigma's
(very) slow and crappy software. there are one or two other raw
converters that reverse engineered the format but didn't do a
complete job so the results aren't as good.

You can ding some of the DPs files and use a multitude of converters.
https://www.adobe.com/products/photo...#dngcompatible

those are the old dp series cameras with the 4.6mp sensor, and not even
all of them since the sd15 isn't even included.

none of the recent 15mp 'merrill' sensors are included.

The Merrils ain't that interesting since they do not have true three
color pixels.

sigma started encrypting the data and doesn't offer support to anyone.

Too bad...

The list of supported cameras for dcraw includes:


Sigma SD9
Sigma SD10
Sigma SD14
Sigma SD15
Sigma SD1
Sigma SD1 Merill
Sigma DP1
Sigma DP1 Merill
Sigma DP1S
Sigma DP1X
Sigma DP2
Sigma DP2 Merill
Sigma DP2S
Sigma DP2X

That would make Lightzone, Darktable and Rawtherapie compatible if you
dare use Indie converters...

Update of the above:
I should have said by dare I meant that opens source software are built
maintained and support are given by volunteers. Therefore you can't
demand nor expect consistency in performance, compatibility or support.
If you're out of luck you might lose massive amount of time on making
things work that could have been spent way better.
--
teleportation kills
http://tinyurl.com/androidphotography

In article , android
wrote:

the raw format is proprietary which means you're stuck with sigma's
(very) slow and crappy software. there are one or two other raw
converters that reverse engineered the format but didn't do a complete
job so the results aren't as good.

You can ding some of the DPs files and use a multitude of converters.
https://www.adobe.com/products/photo...#dngcompatible

those are the old dp series cameras with the 4.6mp sensor, and not even
all of them since the sd15 isn't even included.

none of the recent 15mp 'merrill' sensors are included.

The Merrils ain't that interesting since they do not have true three
color pixels.

you might be thinking of the new quattro sensor.

the merrill and pre-merrill sensors have three layers per pixel but
they are not rgb devices. they sample three overlapping curves from
which rgb is derived and it's *less* accurate than bayer.

the quattro sensor has a high resolution top layer and two low
resolution bottom layers, which means there's not only the colour
interpolation but now there's spatial interpolation on top of that, the
very thing sigma has been saying was bad.

sigma started encrypting the data and doesn't offer support to anyone.

Too bad...

not really. it's more hassle than it's worth for a very, very tiny
market share.

In article , ray carter
wrote:


The list of supported cameras for dcraw includes:


Sigma SD9
Sigma SD10
Sigma SD14
Sigma SD15
Sigma SD1
Sigma SD1 Merill
Sigma DP1
Sigma DP1 Merill
Sigma DP1S
Sigma DP1X
Sigma DP2
Sigma DP2 Merill
Sigma DP2S
Sigma DP2X

dcraw does a rather poor job with sigma images because there is no
official support. he's guessing, like everyone else.

only sigma knows the secret sauce to get reasonably good results. even
adobe, who worked with sigma for the early sensors, isn't as good as
sigma's own software (but it's not as slow).

it's basically sigma's software or nothing.

In article ,
nospam wrote:

In article , android
wrote:

the raw format is proprietary which means you're stuck with sigma's
(very) slow and crappy software. there are one or two other raw
converters that reverse engineered the format but didn't do a complete
job so the results aren't as good.

You can ding some of the DPs files and use a multitude of converters.
https://www.adobe.com/products/photo...#dngcompatible

those are the old dp series cameras with the 4.6mp sensor, and not even
all of them since the sd15 isn't even included.

none of the recent 15mp 'merrill' sensors are included.

The Merrils ain't that interesting since they do not have true three
color pixels.

you might be thinking of the new quattro sensor.

the merrill and pre-merrill sensors have three layers per pixel but
they are not rgb devices. they sample three overlapping curves from
which rgb is derived and it's *less* accurate than bayer.

Wikipedia: The Foveon X3 sensor creates its RGB color output for each
photosite by combining the outputs of each of the stacked photodiodes at
each of its photosites. http://tinyurl.com/tqlyj

Do you have a better source?

the quattro sensor has a high resolution top layer and two low
resolution bottom layers, which means there's not only the colour
interpolation but now there's spatial interpolation on top of that, the
very thing sigma has been saying was bad.

sigma started encrypting the data and doesn't offer support to anyone.

Too bad...

not really. it's more hassle than it's worth for a very, very tiny
market share.
--
teleportation kills
http://tinyurl.com/androidphotography

In article , android
wrote:

the raw format is proprietary which means you're stuck with sigma's
(very) slow and crappy software. there are one or two other raw
converters that reverse engineered the format but didn't do a
complete
job so the results aren't as good.

You can ding some of the DPs files and use a multitude of converters.
https://www.adobe.com/products/photo...#dngcompatible

those are the old dp series cameras with the 4.6mp sensor, and not even
all of them since the sd15 isn't even included.

none of the recent 15mp 'merrill' sensors are included.

The Merrils ain't that interesting since they do not have true three
color pixels.

you might be thinking of the new quattro sensor.

the merrill and pre-merrill sensors have three layers per pixel but
they are not rgb devices. they sample three overlapping curves from
which rgb is derived and it's *less* accurate than bayer.

Wikipedia: The Foveon X3 sensor creates its RGB color output for each
photosite by combining the outputs of each of the stacked photodiodes at
each of its photosites. http://tinyurl.com/tqlyj

that doesn't contradict what i said.

it does create rgb from its three layers, and does so with a very
complex transform (which is one reason why it's as slow as it is).

what it *doesn't* do is have red, green and blue layers like the pretty
pictures suggest. all of that is a lie.

also, wikipedia in general isn't authoritative, particularly when the
bulk of that page was written by foveon themselves. it's just more of
the same.

Do you have a better source?

yes. start with the patents.

here's the spectral response of the layers:
http://static.rcgroups.net/forums/at...851638-191-Sta
cked%20imager%20color%20response%20curves.jpg?d=11 49316976

note that there is a ****load of overlap.

In article ,
nospam wrote:

In article , android
wrote:

the raw format is proprietary which means you're stuck with
sigma's
(very) slow and crappy software. there are one or two other raw
converters that reverse engineered the format but didn't do a
complete
job so the results aren't as good.

You can ding some of the DPs files and use a multitude of
converters.
https://www.adobe.com/products/photo...#dngcompatible

those are the old dp series cameras with the 4.6mp sensor, and not
even
all of them since the sd15 isn't even included.

none of the recent 15mp 'merrill' sensors are included.

The Merrils ain't that interesting since they do not have true three
color pixels.

you might be thinking of the new quattro sensor.

the merrill and pre-merrill sensors have three layers per pixel but
they are not rgb devices. they sample three overlapping curves from
which rgb is derived and it's *less* accurate than bayer.

Wikipedia: The Foveon X3 sensor creates its RGB color output for each
photosite by combining the outputs of each of the stacked photodiodes at
each of its photosites. http://tinyurl.com/tqlyj

that doesn't contradict what i said.

it does create rgb from its three layers, and does so with a very
complex transform (which is one reason why it's as slow as it is).

what it *doesn't* do is have red, green and blue layers like the pretty
pictures suggest. all of that is a lie.

also, wikipedia in general isn't authoritative, particularly when the
bulk of that page was written by foveon themselves. it's just more of
the same.

i know. Anyone can chip in. Including wikipedias robots. No no
relation... Thats why I asked for a better source.

Do you have a better source?

yes. start with the patents.

here's the spectral response of the layers:
http://static.rcgroups.net/forums/at...851638-191-Sta
cked%20imager%20color%20response%20curves.jpg?d=11 49316976

note that there is a ****load of overlap.

Watch thou lingo, Sir! That s*** of overlap is simply an effect of the
top layers being filters for the ones that's below. ;-p
--
teleportation kills
http://tinyurl.com/androidphotography

In article , android
wrote:

Do you have a better source?

yes. start with the patents.

here's the spectral response of the layers:
http://static.rcgroups.net/forums/at...851638-191-Sta
cked%20imager%20color%20response%20curves.jpg?d=11 49316976

note that there is a ****load of overlap.

Watch thou lingo, Sir! That s*** of overlap is simply an effect of the
top layers being filters for the ones that's below. ;-p

that's not why. theoretically, red light goes to the bottom, but light
is random and that doesn't always happen, plus silicon is a bad filter
anyway. it's actually rather amazing that they got it to work at all.

and since its response is not the same as a human eye, you get weird
colour casts and metamerism from converting, plus a lot of noise from
the sensor.

interesting idea, but not very good in practice.

In article ,
nospam wrote:

note that there is a ****load of overlap.

Watch thou lingo, Sir! That s*** of overlap is simply an effect of the
top layers being filters for the ones that's below. ;-p

that's not why. theoretically, red light goes to the bottom, but light
is random and that doesn't always happen, plus silicon is a bad filter
anyway. it's actually rather amazing that they got it to work at all.

and since its response is not the same as a human eye, you get weird
colour casts and metamerism from converting, plus a lot of noise from
the sensor.

interesting idea, but not very good in practice.

Well, it doesn't have that physical spread of the color elements of the
photosites that you were grumping about earlier in the tread.

Anyways, the OP seem to have found his camera...
--
teleportation kills
http://tinyurl.com/androidphotography