[RioRico]

Impossible. Software cannot simply pull more of an image out of thin air.

Oh, I pull stuff out of thin air (or some portion of my anatomy) all the time. Such stuff is usually called 'artifacts'. Why, I can turn a 912x1216px pic into a 60x90cm print, no problem! Sure, it looks rather Chuck-Close-ish. Is that so bad?

[/fun]

No, PP can't add details from the subject field to the captured frame. We can PP to play various tricks to make the captured details more visible. And we can PP to add details that WEREN'T in the original subject field. If we want the resultant image to be as accurate as possible a rendition of the subject field, such additions are bad. If we just want to make a striking image, then we do WHATEVER IT TAKES.

[drougge]

The example images are from an E-PL2. And all they really show is that the default jpeg settings throw the borders away. The raw should contain more border information than you are seeing in the corrected picture, since correction does throw some data away. I suppose it's possible the lens manages to not project anything on those parts, but it seems unlikely.

This thread would probably benefit from not using the word "detail" when talking about image area. So anyone who didn't notice this happened should probably re-read some posts.

[Laurentiu Cristofor]

However, this really makes no sense at all because it would mean that all uncorrected Pentax RAW files are cropped....

I have a load of RAWs with my 16-50 I will check these for this phenomenon.

You should be able to reproduce the same kind of difference by taking shots at 16mm and then processing the RAW file once without distortion correction and once with it.

The example images are from an E-PL2.

You will see the same thing from a Pentax camera. I just happened to have these examples handy.

A couple of years ago, when I was trying out several RAW converters for processing Pentax RAW files, I noticed that each RAW converter was producing images of different size than the PDCU software - they were all larger.

Impossible. Software cannot simply pull more of an image out of thin air.

Yes, it's ludicrous. Except the software pulls it out of the RAW file. There is no magic here. It's just that RAW files contain more data than we see in a straight JPEG conversion.

[drougge]

You will see the same thing from a Pentax camera. I just happened to have these examples handy.

A couple of years ago, when I was trying out several RAW converters for processing Pentax RAW files, I noticed that each RAW converter was producing images of different size than the PDCU software - they were all larger.

And the K-5 produces photos of different shape depending on which resolution you select. But my point still stands just fine as far as I can see: Software that doesn't artificially restrict you will always get more edge-content when you do not correct distortion.

If the software everyone uses is somehow broken, that's a separate issue. I've certainly had problems with RAWs from new cameras getting too much included, leaving calibration data at the edges, so having the opposite problem doesn't seem so unlikely. But it's still unrelated to what would be possible with proper software.

[ElJamoquio]

Yes, it's ludicrous. Except the software pulls it out of the RAW file. There is no magic here. It's just that RAW files contain more data than we see in a straight JPEG conversion.

He was referring to a condition (and quoting a post) where the light was not even on the sensor, or at least the recorded parts of the sensor.

[Alizarine]

While I have yet to believe in LC's statements about "more data" in the RAW files that the camera crops out during distortion correction, I do remember that viewing DNG files in Picasa say that the RAW photos from my K-x indicates the dimensions to be "4309x2848" instead of the usual "4288x2848", but when I open them in Photoshop they're 4288x2848. I'll upload screencaps later when I get home to verify.

[Asha]

But what do you expect from inexpensive lenses? The entire point of this focus on sharpness at the detriment of distortion is to produce lenses that are either more compact or more affordable or both. You can always pay more to get lenses without distortion if that is what you want to use.

For the record, it is possible to achieve a compact lens design with great image quality and low distortion by evenly distributing optical "power" on both sides of the aperture stop. The result is a lens which is relatively short in overall physical length, but the rear element could end up very close to the focal plane. There are some very old patents that describe this kind of design, and they have great performance WITHOUT the use of aspheric elements!

[Laurentiu Cristofor]

Do Pentax RAW files contain extra information?

I did a quick test with a K-7 image and:

PDCU gives me: 4672 x 3104
dcraw gives me: 4684 x 3122 - and there is extra information in that difference
dxomark describes the K-7 sensor as having 4736 x 3136

So extra information is there - the question is how much there is and how to get at it.

For a reference, the E-PL2 images come at 4032 x 3024 and dxomark describes the camera as 4096 x 3084, but that extra detail I pulled seems to indicate a latitude of around 100 pixels on the horizontal, which implied that the actual captured resolution might well be 4100+, which is more than raw converters may pull out.

Can we get at this information? The Olympus software obviously can, but PDCU does not seem to make use of it - I tried some transformations and I always ended up with a cropped result. But just because PDCU doesn't do such trick now doesn't mean that a future version or a third party software cannot exploit such information.

[Mistral75]

Another one: 18-70mm f/3.6-7.0 with only 2.17mm back focus!

Pentax 18-70mm F3.5-7.1 APS-C??????????????????????So-net???

- 6 lenses in 6 groups, of which 3 aspheric lenses
- total length: 7cm at 18 mm and 10cm at 70mm but apparently 4.3cm protruding into the body (register space) thus only circa 2.5cm (around one inch) "apparent" (in front of the mount)
- rear focus ("G3" group consisting in a single lens)

18mm, focus on infinity (if I am not mistaken)



70mm, closest focus (same remark as above)

[parsons]

ouch, f7.0 at 70mm. that is pretty dark. makes the 77 look like a floodlight..

nice that it is only 2.5cm long though! about the size of the 21 and 70mm ltds i guess?

[RonHendriks1966]

70mm/f7 is way to slow. That lens is useless for AF with K-01 in an not bright litt interiour.

I like the idea, but when it works out like this I can't recommend it to anyone.

What is the main problem that it is so slow? Does anyone know that? Would it be better when it doesn't start at such a wide-angle? Would 20-70mm be better in lightgattering?

[parsons]

70mm/f7 is way to slow. That lens is useless for AF with K-01 in an not bright litt interiour.

I like the idea, but when it works out like this I can't recommend it to anyone.

What is the main problem that it is so slow? Does anyone know that? Would it be better when it doesn't start at such a wide-angle? Would 20-70mm be better in lightgattering?

would the focus assist lamp be sufficient?

[ElJamoquio]

Sure, it would help. Generally I turn off the focus assist light - because people find picture taking even MORE annoying with the light.

[philbaum]

What is the main problem that it is so slow? Does anyone know that? Would it be better when it doesn't start at such a wide-angle? Would 20-70mm be better in lightgattering?

As Mistral quoted, 4.3cm of the retracted 7" lens length has to fit within the body (and within the K mount) of the K01. Probably doesn't provide enough diameter for a faster telescopic lens.

Of course, K01 owners can always choose to buy a conventional lens that will sit outside the body. With the peaking function, any number of smallish zoom manual lenses can fit nicely to the outside of the camera, or even the Pentax 17-70 f4 sdm.

[RioRico]

ouch, f7.0 at 70mm. that is pretty dark.

I'll just bet that at 55mm it's f/5.6. So it's the old kit lens, stretched a little, still obeying the same damn laws of optics. If a zoom stop was placed at 55mm, it's essentially the DA18-55/3.5-5.6 but with its gutz inside the mirrorbox space.