Has anyone else with a 50D noticed.......

[ohyva]

that when images are viewed at 100%,the full size raw file is not as
sharp as Sraw1 @ 100%,and Sraw2 @ 100% is sharper than both.???
WTF?

sRAW2 packs per pixel green and either red or blue component. Thus less interpolation needed meaning closer to ideal sharpness.

I have no exact info what sRAW1 does, but scale-down always tends to increase the per pixel sharpness.

From a RAW file you may get around 85% of the linear sharpness of the ideal camera (having RGP pixels and ideal AA filter) - so with a scaledown by linear factor 80% or less you can get quite close to what an ideal camera would produce. This of course depends on the quality of scale-down algo you use. So e.g. 21MP image scaled down to 12MP image is generally much sharper than a 12MP bayer sensor based camera can produce.

I ran this test many times with my 300 2.8IS,so I'm not imagining
anything.
Does this have something to do with my monitor not being able to
resolve all the data or ?
With my old 30D,100% always showed the most detail,now the 15mp
images of the 50 look softer at 100% but when reduced to 50% in
DPP,they look superb.

The lens you have used is a sharp one, but may still have hard times to draw sharp enough image to 50D sensor to get it's full potential. Remember we are now talking of what the contrast of the lens at c. 80 LP/mm ( 160 lines/mm).

Why would a 7mp raw file look sharper than a 15mp raw file @ 100% ?
I will buy you something nice if you can explain this.
Tanks.

 

[RedFox88]

if the full sized images may be outresolving the pixel count of the
Monitor screen ? (ie: my 50D is 4000+ x 3000+ pix whereas the monitor
is around 1600 x 1200 at native resolution). Driing the monitor at
"Non-native" resolution degrades the image so it could be that its
the same effect?

A pixel is a pixel is a pixel.

 

[Brian Wadie]

A pixel is a pixel is a pixel.

except a pixel from the CMOS doesn't map onto a pixel on the monitor screen thus on the screen a pixel will represent an interpolation of several pixels from the cmos.

You are not looking at the original pixels - just the representation of them on another screen

(Which is why if you look at the same image on two different LCD monitors with different resolution you will judge the image sharpness to be different)

 

[Buitre]

I´m considering a 50D( don´t have it yet).

This "issue" regarding sharpness is quite interesting. Bob Atkins has an interesting take on this:

Quote fom Bob Atkins website:

"The fact is that the higher resolution of the 50D will result in higher resolution, sharper, images than those from the 40D whatever lens you use. Doesn't matter if it's the pretty average "all plastic kit zoom" shot wide open or a super sharp lens like the EF 135/2.0L shot at f8."
http://bobatkins.com/photography/digital/canon_eos_50D_resolution_confusion.html

 

[Blue II]

I mean,100% is 100% right?

Not really. A 30D's 8MP image printed at 100% will be smaller than a 50D's 15MP image printed at 100%.

--
Steve

 

[RedFox88]

A pixel is a pixel is a pixel.

except a pixel from the CMOS doesn't map onto a pixel on the monitor
screen thus on the screen a pixel will represent an interpolation of
several pixels from the cmos.

You are not looking at the original pixels - just the representation
of them on another screen

If you need to believe that to help yourself sleep at night, go ahead.

 

[Brian Wadie]

just think about how you are observing the pixels from the camera - you are looking through the medium of an LCD screen at the outpiut from another device - the measurements must therefore be modified by that screen, the tool you are using to make your judgements

(I made my living as a Scientist and was one of the earliest users of digital imaging in the UK doing this sort of thing (1974) and am only to aware of the formation of digital artefacts via measurement methods)

Sleep Well

 

[Brian D. Schneider]

yep,you're a knuckle head as well.
I didn't say my images were soft,I said the smaller ones look
sharper,the larger softER.
so what?-so why have 15mp if they can't be as sharp as an 8mp image
when they are both viewed at 100%?
that's what.
knucklehead.

--You wont make many friends this way ****.
Brian Schneider

 

[Dave In LA]

--
‹(•¿•)› Dave in NW Louisiana
http://critterscapes.com

 

[Dave In LA]

I posted another frame different this this one recently and had sharpened it too much.

This has minimal pp and is at 100%....the dof is shallow so take into consideration.

This would make a much larger sharp print than my 30D(that I replaced with 1D MkIIN)

just sharing for the conversation



--
‹(•¿•)› Dave in NW Louisiana
http://critterscapes.com

 

[TimeTraveler67]

yep,you're a knuckle head as well.
I didn't say my images were soft,I said the smaller ones look
sharper,the larger softER.
so what?-so why have 15mp if they can't be as sharp as an 8mp image
when they are both viewed at 100%?
that's what.
knucklehead.

--You wont make many friends this way ****.
Brian Schneider

I'm sorry I offended you. I was hoping to get an answer with a technical fling to it which regretfully,you could not accomplish. There are some brilliant people on here who can answer almost any tecnical question without delay and they are all to be respected for their abilities.
Unfortunately,somehow you got mixed in with these people and answered as well.
No worries. Format your card,roll a fatty,hop on your scooter and disappear.
God bless.

 

[RedFox88]

(I made my living as a Scientist and was one of the earliest users of
digital imaging in the UK doing this sort of thing (1974) and am
only to aware of the formation of digital artefacts via measurement
methods)

Then why aren't you designing your own sensors and electronics, einstein?

 

[Brian D. Schneider]

yep,you're a knuckle head as well.
I didn't say my images were soft,I said the smaller ones look
sharper,the larger softER.
so what?-so why have 15mp if they can't be as sharp as an 8mp image
when they are both viewed at 100%?
that's what.
knucklehead.

--You wont make many friends this way ****.
Brian Schneider

I'm sorry I offended you. I was hoping to get an answer with a
technical fling to it which regretfully,you could not accomplish.
There are some brilliant people on here who can answer almost any
tecnical question without delay and they are all to be respected for
their abilities.
Unfortunately,somehow you got mixed in with these people and answered
as well.
No worries. Format your card,roll a fatty,hop on your scooter and
disappear.
God bless.

--Dude you started off calling mailman88 a knucklhead, if you want answers I suggest you be a little nicer to someone who is trying to help.
p.s. I cant roll one, I'm down to seeds & stems again. Got any?
Brian Schneider

 

[ohyva]

if you printed images from both-at 100%
resolution the 30d image would be sharper-right? The printer is going
to print what I see. If I have to downsize the 50's image to gain
sharpness,why bother?

You don't print at 100%. Printing is done at the native resolution
of the printer. Many of us prefer to do the resampling ourselves, in
order to optimize print quality. Others simply send the file to the
printer as-is, and let the printer sort it out. Whenever the input
resolution doesn't match the hardware resolution, the printer's
software will do the resampling. Either way, the file has been
resampled before printing.

I guess this depends on what printing technology you use. With some real RGP printing technology (sublimation printer???) I guess your statement is accurate. But with the most common inkjet perhaps not. Typicallt their HW resolution is in order of 1440dpi, but AFAIK this means the spacing of the microdrops of ink and at this rtesolution you may not get the full gamut out. So some rasterization with color blending in the paper is required.

I actually do not know any good source telling what is the real resolution you can aim to get from an inkjet print. So if anybody can provide me a good link/links, I'd appreciate.

So as a lazy man, I typically just feed the full size image to the printer driver and get what I get.

For small prints, images from both the 30D and 50D are resampled
downward, and this tends to mask any IQ differences between the two
cameras. The resampling algorithm may even make more difference here
than anything else.

For large prints, images from both the 30D and 50D are resampled
upward. Because larger files from the 50D need a smaller increase
than those from the 30D, large prints from the 50D tend to look
better than those from the 30D.

I guess you are right. Don't know the limits but I guess A3 size already show some difference. A4 may be on the edge where the 8MP is still enough (although you cannot any more get the full 300dpi which I consider as the requirement of "good quality" - and for 40D vs 50D I guess you have to go bigger than A4).

It's as simple as that.

 

[ScratchDisk]

Hi,
Brian I agree with you that there is not 1:1 mapping.

On top of that most imaging software differs in how well it renders at different magnifications (below 100%). Even PS is not immune from this. I can't find it just now, but I'm sure I read a PS readme file that explained this. It means that printed output is always going to look sharper and by a degree that varies with screen magnification (assuming magnifications match). I would be interested to read more on this, do you know of any good sources? Also I think it does contribute to the 'new cam moaning' currently aimed at the 50D.
Cheers, Donald

 

[Phil Hill]

But with the most common inkjet perhaps not.
Typicallt their HW resolution is in order of 1440dpi

I should have been more clear about my use of the term "hardware resolution." I was speaking of the printer's resolution in pixels (as it relates to the image), not dots. DPI and PPI are not interchangeable in this situation.

An inkjet printer, of course, uses multiple "dots" of ink for each pixel. Higher DPI counts are indicative of greater precision (and smaller drops of ink), leading to a visual impression of continuous color. But that's not the same thing as resolution.

An inkjet printer that can lay down ink at 1440 DPI does not resolve 1440 PPI. Typical inkjet printer resolutions tend to be around 300 - 360 PPI, and a photographic image normally gets resampled to that resolution, one way or another.

 

[Brian Wadie]

I was reminding myself of a few sources this morning, I hope they may be of interest:

http://www.normankoren.com/Tutorials/MTF.html (Norman is a bit of a Guru)

http://www.imaging-resource.com/TIPS/TESTS/TESTS.HTM (got a bit about pixel mapping and apparent sharpness in there somewhere)

http://www.clarkvision.com/imagedetail/scandetail.html#digi2camres2 ( a bit ancient but good grounding)

http://www.dpreview.com/learn/?/key=resolution (the dpreview method)

http://www.zeiss.com/C12567A8003B58B9/Contents-Frame/48D8F331DF48EE72C1256CEF002B0240 (only speed read this one but its fun as it relates to film)

http://www.oznet.ksu.edu/prev/5-YP2005-2009/AlaviExtrusion02Apr04.pdf (for nostalgia - this is the sort of work I was doing using digital imaging techniques on in the '70s and - my god - they are still doing it now! )

 

[jonrobertp]

more acceptance of submitted images for said ( OP ) reason. Easier apparently to sell a smaller sharp image than a bigger one with some problems.

 

[ohyva]

But with the most common inkjet perhaps not.
Typicallt their HW resolution is in order of 1440dpi

I should have been more clear about my use of the term "hardware
resolution." I was speaking of the printer's resolution in pixels
(as it relates to the image), not dots. DPI and PPI are not
interchangeable in this situation.

With inkjet printers I do not think there is no such thing as "full gamut HW resolution" - or probably there is, but I have newer seen such specified or told in the specification. I guess this depends at least to some extend in which point we are in the gamut.

An inkjet printer, of course, uses multiple "dots" of ink for each
pixel. Higher DPI counts are indicative of greater precision (and
smaller drops of ink), leading to a visual impression of continuous
color. But that's not the same thing as resolution.

I think the equation of how "sharp" the printer is is a lot more complex starting the physical step size, the ink droplet size, the paper quality and the ink characteristics and how much those actually spread inside the paper.

And resolution as a term is generally quite confusing. Sometimes it refers to the physical parameters like printer mechanical step size (e.g. 1440 dpi) or the sensor pixel amount, and sometimes its used to tell the measured resolving power i.e. what's the real lines/mm you can print or what the real size of details you camera can resolve.

An inkjet printer that can lay down ink at 1440 DPI does not resolve
1440 PPI.

Actually some "high resolution" papers claim they can go at least to 720 dpi - but those are not good for photos IMHO.

Typical inkjet printer resolutions tend to be around 300 -
360 PPI, and a photographic image normally gets resampled to that
resolution, one way or another.

That the info I'd like to get clarified i.e. what the print resolution my printer or the best of the printers can resolve, how much that may depends on the color/shade we talk i.e. what the print resolution I should use when preparing my photo for printing. I have not yet found any real deep going study of this anywhere.

So I either do not resampling if if I do, then 300 dpi. All low reso image I send via email are resampled to about 2MP i.e. 1200*1800 pixels and 300 dpi (4"x6" print).

 

[NetMage]

That the info I'd like to get clarified i.e. what the print
resolution my printer or the best of the printers can resolve, how
much that may depends on the color/shade we talk i.e. what the print
resolution I should use when preparing my photo for printing. I have
not yet found any real deep going study of this anywhere.

I would suggest checking out QImage and some of the posts about it. QImage shows my Canon i9900 at 600x600 dpi.

--
EOS 50D, 20D, 10D, 630, A-1, SD1000
-- Please remove the Quote option!