Resolution measurements - TZ30 (ZS20) - Many images

Started Apr 3, 2013 | Discussions thread
J C Brown
Senior MemberPosts: 1,423
Like?
Re: Resolution measurements - TZ30 (ZS20) - For Ianperegian
In reply to Ianperegian, Apr 9, 2013

Ianperegian wrote:

Hi Jimmy,

Thanks for a very interesting thread, and thanks also for making your resolution charts freely available for people to use for their own testing.  i think your charts are very ingenious, and they provide a much more reliable way of visually assessing resolution than do the B&W "narrowing lines" charts that are often used on the review sites, as well as allowing different colours to be assessed at the same time.

I noticed kkardster's comment about the lower resolution of the yellow Es.  I wondered whether changing the background colour would have any effect, since I understand that in the demosaicing of a Bayer matrix the colour value of a pixel includes interpolation from the surrounding pixels.

I thought to try blue as the background since as you said, yellow is made up of red and green, and pure blue lacks both red and green.  I prepared a copy of your earlier Es chart with a blue background using PSE with the paint bucket tool (obviously the blue Es disappear).  Printing this chart at the correct size (75mm for the black line) with my Epson printer gave good resolution of the smallest Es, as shown in a closeup shot I took with my ZS20:

I attached the blue background chart print onto the chart board I'd previously prepared for testing my FZ38 and ZS20, as shown here .

I tested using my ZS20 at 51mm FL with the tripod adjusted to the correct distance so that the black line on the chart covered 400 pixels. Below is the resulting full sized image:

This is a 100% crop of the blue background chart beside the white background chart:

As can be seen, the blue background has had a marked influence on the Es, with the yellow Es being much better resolved, while black Es (and text) are poorly resolved.  It seems to indicate that yellow, cyan and probably other colours too, give better resolution when they are surrounded by particular other colours.  Also, I think the intensity of the background colour and the exposure would have some effect as well.

A conclusion could be that when taking shots of canaries, choose a blue sky background rather than a white clouds background!

Ian

-- hide signature --

Hi Ian,

Thank you for your very kind remarks about my test charts and the reliability of measurements derived from them. Thanks also for the time and trouble you’ve taken to investigate kkardster’s suggestion about the low contrast of yellow on white by changing the background colour to blue prior to printing.

The central crop from the image of your full test chart demonstrates clearly that the colour of the background has a significant effect on the resolution of Es of different colours.

To illustrate the effect more clearly I’ve taken small crops from the area around the 1.5 pixel Es on both white and blue backgrounds, pasted them side by side into a single image and used the Print Screen option to make the individual pixels clearly visible in the following image.

Comparison of appearance and resolution of coloured Es printed on white and blue background

As may be seen from the composite image above the appearance of the green, cyan and yellow Es on the blue background is very similar to that of the red, blue, black and magenta Es on the white background. In addition the line thickness of the smallest recognisable Es of each colour is about 1.5 pixels which is consistent with the 1.5 pixel value derived from the graphical simulation described in my report on the resolution of the FZ50.

Your conclusions about the influence of background colour on the resolution of Es of different colours are also consistent with my experience of using bracketed exposures while trying to find the optimum exposure during some of my earlier tests.

These tests revealed that the difference in the shades of grey which resulted from a +/- 0.33 EV difference in exposure was sufficient to make a noticeable difference to the ability to clearly recognise Es of different colours and that for assessing resolution a light grey background was preferable to a white one.

Thanks again for your interest and for your interesting contribution.

Jimmy

-- hide signature --

J C Brown

Reply   Reply with quote   Complain
Post (hide subjects)Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark post MMy threads
Color scheme? Blue / Yellow