Me <> wrote:
> On 25/09/2012 9:22 a.m., otter wrote:
>> On Sep 24, 2:28 pm, nos...@nospam.com (Paul Ciszek) wrote:
>>> If you had to pick out one prime lens, what focal length would you
>>> choose? Please mention your sensor type if not full frame.
>> Me, personally? The only prime I have is the Samyang 14mm. I have
>> Canon full-frame. I'd say my future prime purchases would probably
>> also be at the wide, or ultra-wide end. The Canon zooms aren't really
>> stellar in that range. I'm seriously looking at some options, like
>> either the 17 or 24 TS-E, or 21mm Zeiss, etc. I tend to shoot a lot
>> of landscapes.
>> I already have a very sharp zoom to cover the 70-200 range, and
>> looking at another to cover 24-70. I may be selling my 24-105 soon.
>> That said, I am also on the lookout for some fast primes for low light
>> shooting. I've heard that the 50 f/1.2L is a hard lens to master,
>> I'd probably start with the 50 f/1.4. I know people like the 135L,
>> the 85L, and the 35L. Maybe when I get a bonus at work (fat chance of
>> that, these days).
>> The 400mm f/5.6 is also on my wish list, but I have other priorities
>> at the moment.
> Potential buyers of uber-fast prime lenses in the standard focal length
> ranges, if they're after the extra stop or half stop for low light,
> should read the following article before deciding whether it's worth it:
> http://www.dxomark.com/index.php/Pub...s/F-stop-blues
Now, they're not saying a single word about how they got that
data, and down to 1/10 EV exact, when the camera supposedly
counteracts by increasing the ISO.
So how do they get the data? From noise profiling? To 1/10 EV?
I disbelieve!
> So, there may be a case for "larger pixels" (or BSI sensors) after all.
Nope, not for pictorial photography. As long as the pixels are
not too small (and thus lose too much space (to store electrons,
not to capure photons -> microlenses) to circuity & co), you can
always downsample and get the same result. Reading noise is really
low these days, especially with Canon. (And the sensor size being
given, you get the same amount of light onto the sensor -> i.e.
identical photon noise at identical photo size. It's just that
you can magnify more with smaller pixels without becoming blocky.
> This is not to be confused with F-stop vs T-stop light transmission
> through the lens, but presumably the filter stack and sensor microlens
> array's inability to re-focus light arriving at an oblique angle, down
> to the photo diodes.
This is the system T-stop, the shallow focus of the f/stop is
still the same, isn't it?
> At the pixel pitch of a Canon 7d or a Nikon D800, your camera turns an
> f1.2 lens into an f1.8 lens, and worse, your camera maker very sneakily
> tricks you into thinking you're still getting "value for money" by
> secretly tweaking the sensitivity.
So what do they do to an f/1.8 lens? It's also affected ...
it's not the f/1.8 lens you take it for.
And it's well known that additional faster stops cost rapidly
increasingly more.
> So why don't the camera makers tell us this?
Maybe because --- simplest explanation --- dxomark got it's
wires crossed and there's no such effect.
-Wolfgang
Similar Threads
