70-200 or 70-300 for Pentax

Alfred Molon wrote:
In article , Ray Fischer
Why not go straight to a 25MP camera?

Not available yet,

Wrong. Has been available for quite a while.
See e.g. http://www.hasselbladusa.com/promotions/h3dii.aspx (39MP) or
http://www.mamiya.com/mamiya-645zd.html (ok, that's only 22MP).

besides will likely be way above my budget.

That on the other hand is quite possible considering the price tag for
those cameras.

And please don't tell me I don't need the MP. Alamy is requesting
submissions with at least 16.7MP and 14MP is closer to that than 10 or
12MP.

How does that correlate to
quote
We reject images for):
[...]
6. Digital camera not suitable for Alamy (pro level 6 mega pixels)
/quote

No, they ask for images of at least 17MB, not 17MP.

Nope. They ask for 48 MByte images, which translates into 16.7 MP.

According to which formula? Doesn't that also depend a lot on the
compression (lossy as well as lossless) and motive? Very uniform photos
(it's all blue sky) compress better than photos with many different
contrasting objects (masses of twigs).

jue
jue

Alfred Molon wrote:
Thanks for your reply, but that still doesn't answer my question ahout
the 70-300 lens.

Sigma and Tamron have such lenses, but performance is less than stellar.
Sony seems to have a good 70-300 lens, which makes me wonder if I should
get a A350 instead.

If you are willing to switch brands anyway then Nikon makes a nice
70-300 VR lens which for the money is an excellent choice.

jue

David J Taylor wrote:
Alfred Molon wrote:
In article , David J
Taylor says...

If you had bought Nikon, there's a very well respected 70-300mm VR
lens:


http://www.europe-nikon.com/product/.../overview.html
I haven't bought anything yet, but am not considering a Nikon for the
moment because they don't have a 14MP model and I don't want to change
the camera after a year or two just because the MP count has
increased. With a 14MP model I can skip one generation and upgrade
when 16-18MP cameras hit the market.

And please don't tell me I don't need the MP. Alamy is requesting
submissions with at least 16.7MP and 14MP is closer to that than 10 or
12MP.

More MP equals more noise. Are you really sure you want that?

Others have already mentioned that 6MP is enough to Alamy.

More MP doesn't have to equal more noise. While shot noise might seem
to be more viewed at 100% pixel view, noise on a "whole image" basis can
be less if overall sensor efficiency isn't reduced.
It could be unwise to pre-judge performance of future ~15mp aps-c
sensors based on the poor usable dynamic range of the Pentax K20d and
Sony a350 sensors. They are the first examples of dslr makers launching
very high MP cameras with worse sensor performance than earlier lower MP
models. OTOH, if you can live with low DR - limiting ability to PP the
raw images even at base iso - and have lenses that are "up to" the
demands of high MP sensors, they might be okay for some people's uses.
Unfortunately IMO, the "reason" to have high MP (capturing detail in
landscape shooting)is negated by the fact that it's also in landscape
shooting where the dynamic range limitation has potential to hurt image
quality the most. Other makers have consistently improved overall
sensor performance despite increasing pixel counts. I expect that they
will continue to do so - or face some serious criticism.

frederick wrote:
[]
More MP doesn't have to equal more noise. While shot noise might seem
to be more viewed at 100% pixel view, noise on a "whole image" basis
can be less if overall sensor efficiency isn't reduced.

Well it changes the noise spectrum at least, and hence the character of
the nosie. I don't know whether the eye/brain has a single response to
the changed noise spectrum, or whether that's something which depends on
the observer - some people preferring a slightly noisier but crisper
"grain" structure. It may even be subject-dependant. But I agree in
general.

It could be unwise to pre-judge performance of future ~15mp aps-c
sensors based on the poor usable dynamic range of the Pentax K20d and
Sony a350 sensors. They are the first examples of dslr makers
launching very high MP cameras with worse sensor performance than
earlier lower MP models. OTOH, if you can live with low DR - limiting
ability to PP the raw images even at base iso - and have lenses that
are "up to" the demands of high MP sensors, they might be okay for
some people's uses. Unfortunately IMO, the "reason" to have high MP
(capturing detail in landscape shooting)is negated by the fact that
it's also in landscape shooting where the dynamic range limitation
has potential to hurt image quality the most. Other makers have
consistently improved overall sensor performance despite increasing
pixel counts. I expect that they will continue to do so - or face
some serious criticism.

Indeed there may be scope for sensor improvement - there are arguments
about the possible practical advantages of front- versus back-illumination
right now.

For landscapes I would suggest using the easily available panorama tools -
even something like Auto Stitch will probably make a better final image
from two or more shots than a high MP sensor and expensive lens. At
least, if that's an acceptable way of working for you.

Cheers,
David

In article 1214171065.46220@ftpsrv1, frederick says...

More MP doesn't have to equal more noise. While shot noise might seem
to be more viewed at 100% pixel view, noise on a "whole image" basis can
be less if overall sensor efficiency isn't reduced.
It could be unwise to pre-judge performance of future ~15mp aps-c
sensors based on the poor usable dynamic range of the Pentax K20d and
Sony a350 sensors. They are the first examples of dslr makers launching
very high MP cameras with worse sensor performance than earlier lower MP
models. OTOH, if you can live with low DR - limiting ability to PP the
raw images even at base iso - and have lenses that are "up to" the
demands of high MP sensors, they might be okay for some people's uses.
Unfortunately IMO, the "reason" to have high MP (capturing detail in
landscape shooting)is negated by the fact that it's also in landscape
shooting where the dynamic range limitation has potential to hurt image
quality the most. Other makers have consistently improved overall
sensor performance despite increasing pixel counts. I expect that they
will continue to do so - or face some serious criticism.

I have no information about the dynamic range of the K20D sensor, and in
fact there is no such information in published reviews, so I wonder how
you reach the conclusion that the K20D has less dynamic range.

But the sensor of the A350 has 8.6 stops of dynamic range. This is more
than enough dynamic range even for landscape shots:
http://www.dpreview.com/reviews/sonydslra350/page20.asp

As a comparison, the Canon 400D, which certainly isn't a poor performer,
despite having only 10MP and consequently larger pixels, has just 8.4
stops of dynamic range:
http://www.dpreview.com/reviews/canoneos400d/page19.asp
--

Alfred Molon
------------------------------
Olympus 50X0, 8080, E3X0, E4X0, E5X0 and E3 forum at
http://tech.groups.yahoo.com/group/MyOlympus/
http://myolympus.org/ photo sharing site

Alfred Molon wrote:

I have no information about the dynamic range of the K20D sensor, and in
fact there is no such information in published reviews, so I wonder how
you reach the conclusion that the K20D has less dynamic range.

But the sensor of the A350 has 8.6 stops of dynamic range. This is more
than enough dynamic range even for landscape shots:
http://www.dpreview.com/reviews/sonydslra350/page20.asp

As a comparison, the Canon 400D, which certainly isn't a poor performer,
despite having only 10MP and consequently larger pixels, has just 8.4
stops of dynamic range:
http://www.dpreview.com/reviews/canoneos400d/page19.asp

The step wedge test used by DPReview is close to totally worthless,
except for looking at default tone curves for jpeg.
Theres some information in the link below about the noisy performance of
the K20d sensor, and impact on usable dynamic range, with an observation
that /usable/ DR is 1.5 - 2 stops less than Canon 40d / D300.
From what I've seen, I don't expect that the a350 sensor is better.
DPReview review of the a350 showing the Canon 450d and K20d give a
better indication of how far behind they are.


http://daystarvisions.com/Docs/Rvws/K20D/pg3.html

In article 1214213673.961314@ftpsrv1, frederick says...
Alfred Molon wrote:

I have no information about the dynamic range of the K20D sensor, and in
fact there is no such information in published reviews, so I wonder how
you reach the conclusion that the K20D has less dynamic range.

But the sensor of the A350 has 8.6 stops of dynamic range. This is more
than enough dynamic range even for landscape shots:
http://www.dpreview.com/reviews/sonydslra350/page20.asp

As a comparison, the Canon 400D, which certainly isn't a poor performer,
despite having only 10MP and consequently larger pixels, has just 8.4
stops of dynamic range:
http://www.dpreview.com/reviews/canoneos400d/page19.asp

The step wedge test used by DPReview is close to totally worthless,
except for looking at default tone curves for jpeg.
Theres some information in the link below about the noisy performance of
the K20d sensor, and impact on usable dynamic range, with an observation
that /usable/ DR is 1.5 - 2 stops less than Canon 40d / D300.
From what I've seen, I don't expect that the a350 sensor is better.
DPReview review of the a350 showing the Canon 450d and K20d give a
better indication of how far behind they are.


http://daystarvisions.com/Docs/Rvws/K20D/pg3.html

Once again, the dpreview tests clearly show 8.6 stops of dynamic range
for the A350 and 8.4 for the Canon 400D.

Your claim of 1.5 to 2 stop less dynamic range is ridicolous,
considering that the pixel size is only 20% smaller. Where you pull out
a factor of 2 of less dynamic range is a mistery and in any case
contradicts basic physics laws.

Also, have a look at high ISO samples of the K20D. You'll notice that
noise is very well controlled even at ISO 1600.
--

Alfred Molon
------------------------------
Olympus 50X0, 8080, E3X0, E4X0, E5X0 and E3 forum at
http://tech.groups.yahoo.com/group/MyOlympus/
http://myolympus.org/ photo sharing site

In article , David J
Taylor says...

For landscapes I would suggest using the easily available panorama tools -
even something like Auto Stitch will probably make a better final image
from two or more shots than a high MP sensor and expensive lens.

I do in fact lots of panoramas with PTGUI, but it's a lot of work.
--

Alfred Molon
------------------------------
Olympus 50X0, 8080, E3X0, E4X0, E5X0 and E3 forum at
http://tech.groups.yahoo.com/group/MyOlympus/
http://myolympus.org/ photo sharing site

Alfred Molon wrote:
In article 1214213673.961314@ftpsrv1, frederick says...
Alfred Molon wrote:

I have no information about the dynamic range of the K20D sensor, and in
fact there is no such information in published reviews, so I wonder how
you reach the conclusion that the K20D has less dynamic range.

But the sensor of the A350 has 8.6 stops of dynamic range. This is more
than enough dynamic range even for landscape shots:
http://www.dpreview.com/reviews/sonydslra350/page20.asp

As a comparison, the Canon 400D, which certainly isn't a poor performer,
despite having only 10MP and consequently larger pixels, has just 8.4
stops of dynamic range:
http://www.dpreview.com/reviews/canoneos400d/page19.asp
The step wedge test used by DPReview is close to totally worthless,
except for looking at default tone curves for jpeg.
Theres some information in the link below about the noisy performance of
the K20d sensor, and impact on usable dynamic range, with an observation
that /usable/ DR is 1.5 - 2 stops less than Canon 40d / D300.
From what I've seen, I don't expect that the a350 sensor is better.
DPReview review of the a350 showing the Canon 450d and K20d give a
better indication of how far behind they are.


http://daystarvisions.com/Docs/Rvws/K20D/pg3.html

Once again, the dpreview tests clearly show 8.6 stops of dynamic range
for the A350 and 8.4 for the Canon 400D.
They show a tone curve applied to raw data - nothing of real interest if
you post-process. If you don't, then be happy with the K20d.

Your claim of 1.5 to 2 stop less dynamic range is ridicolous,
considering that the pixel size is only 20% smaller. Where you pull out
a factor of 2 of less dynamic range is a mistery and in any case
contradicts basic physics laws.
No it doesn't - and it's explained in the link I posted. To sum up, the
sensor isn't very efficient.

Also, have a look at high ISO samples of the K20D. You'll notice that
noise is very well controlled even at ISO 1600.

Better than looking at the ISO 1600 samples from the K20d, look at the
ISO100 samples here and weep:
http://www.dpreview.com/reviews/sonydslra350/page17.asp
Okay - it *might* be lousy in-camera jpeg conversion, but it still
doesn't look very good.

The claim of 1.5-2 stops wasn't made by /me/ - it was made by the author
of the link I posted. People mentioned in that post who have produced
data on the sensor performance, and it's not very good.
You're active on DPReview's Pentax forums. The subject has been
thrashed to death there.

Alfred Molon wrote:
In article , David J
Taylor says...

For landscapes I would suggest using the easily available panorama
tools - even something like Auto Stitch will probably make a better
final image from two or more shots than a high MP sensor and
expensive lens.

I do in fact lots of panoramas with PTGUI, but it's a lot of work.

Then take a look at the free AutoStitch or its commercial variants.

http://www.cs.ubc.ca/~mbrown/autostitch/autostitch.html

Drop a set of files in a directory, choose your output size and quality,
and press the Go button. Very little work. Perhaps if you want "superb"
results you may need to do more, but for just "very good" these tools are
highly effective.

Cheers,
David