100% on the LCD

My apologies if this thread did not pass your quality standards.



I just meant a regular print I would have done like an 8 x 10 at 300 ppi say. Thank you (and everybody else) for your input. It is useful to know what the LCD is showing me.

Yes, completely

You can extract the embeded JPG using DCRAW, it's a full resolution ** quality file. Quite handy when you want a quick JPG without having to actually process the file properly.

Of course on the K10 you have in camera processing and Raw+JPG mode but it comes in handy with *ist series cameras sometimes.

This is a very interesting topic. We all know how important it is to set an LCD monitor to its native resolution for maximum quality. I guess the same applies to watching pictures on an LCD screen.

Selecting Actual Size in PSE shows the picture pixel-by-pixel, where it is at its sharpest.

From my calculations this happens somewhere between 6x and 7x in the K100D camera. It would have been nice had Pentax provided a zoom position to obtain the exaxt match.

BTW, there is nothing wrong in providing further zoom (up to 12x in the K100D) as long as one is aware that the picture is oversampled.

Well, this is really only true if the sharpness of the image as resolved by the lens is greater than (or equal to) the resolution of the sensor.

Which is only true occasionally, of course, since lens imperfections, and indeed our own technique, often mean that we don't capture images as sharp as the sensor is capable of recording.

In such cases, sharpness as perceived by the eye (which is the only type that counts, at the end of the day) is often improved by resampling down to a smaller size.

Or standing a bit further away, of course.

Well, maybe "sharpest" was not the correct word here. The sharpness, as you say, is down to lens quality and photographic skills.

However, displaying your image at pixel-by-pixel size provides the most true representation of your picture.

but when you want to have a 1:1 ratio on the camera LCD you might run into another snag.
The display is 320x240 which is a 4:3 ratio and the image is 3:2.
So 1:1 magnifaction in the horizontal pane would be 14,6 times for the K20D.
Vertical it would be 12,9 times for the K20D.
Or am I wrong about this.

Pixel by pixel viewing might be the most accurate way of viewing but then we would fall into the pixel peeping category.
Did you view your photo's at the grain level when shooting film? No you judged the image when it was printed at the desired size... That is what you should do.

Hi golfdiesel,

From where did you get 320x240 pixels? This yields only 76800 pixels, while the specifications state 210000 (some places 215000). I would expect the ratio is more like 536x402 (=215472), but I have not been able to find these figures anywhere.

Given 3008 horizontal pixels (K100D), the ratio is 5.6. So a one-to-one view should be obtained somewhere between 5x and 6x zoom. Unfortunately. the camera does not have a 5.6x zoom position.

As for the image ratio of 3:2 this applies only at 1x zoom, where you can see a black edge on top and bottom of the LCD. This is filled out to 4:3 as soon as you zoom in.

The film equivalent is hardly relevant, as we did not have any other way of viewing our pictures than the printed copy. But yes, I did check sharpness using a good magnifying glass, the equivalent of viewing the digital image at 100% (pixel-by-pixel).

See my post a little earlier.
In short, the number of pixels they mention with camera's are the combined total (all subpixels). This calculation is based on the approx 230k pixels mentioned for the K20D.

Each pixel has three subpixels side by side so you end up with:
960x240 = 230.400 pixels.

See it like the CCD or CMOS in your camera which doesn't have 6, 10 or 14,6 million full colour pixels. 50% of the pixels are green, 25% blue and 25% red. (the primary filter that is, a CCD/CMOS can only detect light levels).
There is a lot of interpolation done to give you 14,6 million full colour pixels.

[edit]
actually we have call them dots, the lcd has 230.000 dots which account for 76800 pixels.
[/edit]

Sorry, but I don't think you are correct here. My camera has a picture format of 3008 x 2000 pixels = 6.016.000 pixels = 6 Megapixels. Each of those pixels contain subpixels.

It is the same as an LCD monitor with, say, 1280x1024 pixels = 1310720 pixels in total. Each of these contain red, green and blue subpixels.

Terry, in a 3008x2008 CCD, there are exactly 3008 x 2008 photo sites. There are not "sub pixels".

A CCD works by having a "Bayer matrix filter" in front - so the 6million photo sites are arranged as squares of 2 greens, 1 red, 1 blue. Consequently you have 3million green pixels, 1.5 million red, and 1.5 million blue.

To get a full 3-color-per-pixel JPG or TIFF image out the camera (or RAW conversion software) has to interpolate the missing colours (that's one reason why the quality of RAW conversion software varies so much - they all use slightly different algorithms for guessing the missing data).

Check out "Bayer filter" on Wikipedia.org
Matt

Thanks for the enlightenment on CCDs, Matt. Now I understand where and how the 3008x2000 pixels originate.

But this certainly does not apply to LCDs, where each specified pixel contains three sub-pixels (red, green and blue). So I still think I'm right in assuming that the LCD monitor of 21 kPixels must have a resolution of about 536x402 pixels, making a 1:1 display at about 5.6x zoom.

Terry,

You are right about pixels being constructed of 3 subpixels. Those subpixels are sometimes called dots.
There is a lot of confusion about these things especially in camera land but the manufacturers give the number of dots and not the number of pixels.
If they would quote the number of pixels then they would most certainly give you the resolution of the screen instead of the number of dots.
They use the word pixels in the wrong way, if they would name them subpixels they would be right.

/Remco

[edit]
I have sent an email to Pentax with the question if they could clarify this once and for all...
[/edit]