Go Back   Steve's Digicams Forums > Digicam Help > General Discussion

Reply
 
Thread Tools Search this Thread
Old Mar 24, 2004, 5:24 PM   #11
Senior Member
 
Join Date: Dec 2003
Posts: 383
Default Re: Size vs Resolution

We certainly have an interesting dialog going on here that I believe clearly demonstrates the current confusion about resolution. It doesn't help when printer manufacturers or software companies interchange dpi and ppi. Here is a past thread that contains outside links and may help clarify the resolution issue then again since it has been pretty thoroughly rehashed here it may not clarify anything.

http://www.stevesforums.com/phpBB2/v...asc&highlight=

Quote:
Originally Posted by ImKayd1
All my Oly's automatically download at a resolution of 72 do I need to change the resolution when I'm going to print?

I think the answer to your question is that you do not need to change the resolution when you print.

The most simple way to verify this, is to print the same picture twice, one at the default setting of 72dpi and the other at not more than 300dpi. Ensure that you constrain the actual size dimensions to be the same and just compare your two results.

I will venture to guess that this 72 dpi information is being displayed by default. Either the camera is unable to pass along the actual resolution or the software is unable to determine the real resolution so it is by default listing 72 dpi.
PKchopper is offline   Reply With Quote
Old Mar 24, 2004, 6:07 PM   #12
Moderator
 
Join Date: Jun 2002
Posts: 1,139
Default

Quote:
So your suggestion to ImKayd1 to match the printer DPI with the PPI resolution 1:1 assumes he prints with “Durst Lambda, LightJet and similar continuous tone processes”? It is my understanding that DPI in that sense doesn’t properly refer to continuous tone printing anyway.
No, it doesn't assume anything about his particular print because he didn't state which process was to be used, but the vast majority of printers such as LightJet, Durst Lambda, etc., call for 300 ppi and in some cases such as Fuji Frontier 400 ppi as imput density. What the printer does with this is not material to the discussion, it's what the "requirement" of the print driver is and yes, all continuous tone print devices still use dpi ratings, even dye sublimation printers.

The actual print density called for in Epson print drivers depends, as I said earlier on the print media. Primarily the differences have to do with the amount of ink deposited by the jets. Slick media can handle more ink than porous media so to avoid running smears, Epson uses less ink and fewer drops per pixel with matt finish, less still with standard inkjet paper and the most with glossy film or gloss photo paper. The actual max for highest quality injket prints on all current Epson injkets is 720 ppi input. It's in the print driver specs if you care to look it up.

No one said interpolation was "magic," and yes, interpolation does indeed often "add" data which was not in the original file. Pixels are represented in files as numbers representing colors. Interpolation algorithms look at adjacent pixel numeric values and either create same or "intermediate" values depending on the values found in these adjacent pixels. The "intermediate" values are values which create pixels different from those captured by the
sensor and indeed are "estimates" of what probably would have been represented had the sensor had additional photosites at those positions. Sometimes (most often) these values are very good indeed and represent "reality". On the occasions where extreme demarcation exists such as an image of an American flag with contrasting color stripes, the interpolation algorithm "may" possibly create an intermediate value which is incorrect, but a good algorithm usually has provisions for determining how far apart values must be to use equal rather than intermediate values.

If your imput pixel density is too low (such as 120 ppi) no interpolation algorithm can correctly assess the proper values and make a suitable enlargement, but the interpolation process is still used by the print driver to attempt to achieve the quality you stipulate when you choose the print media, etc. The differences between 180 to 250 and 300 ppi input are less than the diffrences between 120 and 180 simply because you have dropped below the threshold of what can be properly estimated. It's essentially GIGO - Garbage in, Garbage out.

The degree of excellence in interpolated values for enlargement purposes is dependent primarily on two factors which are inter-related: angle of view and capture resolution. The smaller the angle of view on the original capture, the fewer pixels are needed by the interpolation engine to render a satisfactory print. The larger the angle of view on the frame (detailed landscapes, etc.) the more pixel density on the capture is necessary to properly define boundaries of fine detail in the image.

Lin
Lin Evans is offline   Reply With Quote
Old Mar 24, 2004, 6:53 PM   #13
Member
 
Join Date: Feb 2004
Posts: 48
Default

This is off-topic, but for a while I've randomly been getting login prompts on these forums. I believe I've traced it to AMG's avatar. AMG, I'd like to ask if there's anyway to fix this because I'm never comfortable clicking strange prompts.
Effen is offline   Reply With Quote
Old Mar 24, 2004, 10:40 PM   #14
Senior Member
 
BillDrew's Avatar
 
Join Date: Jun 2002
Location: Hay River Township, WI
Posts: 2,512
Default

Quote:
Originally Posted by Lin Evans
... The smaller the angle of view on the original capture, the fewer pixels are needed by the interpolation engine to render a satisfactory print. The larger the angle of view on the frame (detailed landscapes, etc.) the more pixel density on the capture is necessary to properly define boundaries of fine detail in the image.
I don't understand that point. As an example, if a shot is made of a brick wall with a telephoto would require less resolution than one shot with a short lens while standing closer. (I used a brick wall as an example to avoid any issue of depth perspective since the subject is two dimensional).

What am I missing here?
BillDrew is offline   Reply With Quote
Old Mar 24, 2004, 11:10 PM   #15
Senior Member
 
Join Date: Dec 2003
Posts: 383
Default

Just a guess Bill but I would think your example illustrates the point exactly assuming your short lens distorts the edges (such as a fisheye). In you telephoto shot of a brick and mortar wall (assuming you are perpendicular to it) all of the brick and mortar lines will essentially be perfectly square.

However with the short lens you will get lines on the periphery that are no longer square but appear curved.

An interpolation algorithm can easily fill in the missing color/pixel on the telephoto shot because it can easily predict side-by-side values. However on the short lens shot where the periphery lines are curved it has to work harder and there will be a greater possibility of error (the example is a bit simple, but obviously that was your intent).
PKchopper is offline   Reply With Quote
Old Mar 24, 2004, 11:58 PM   #16
Senior Member
 
BillDrew's Avatar
 
Join Date: Jun 2002
Location: Hay River Township, WI
Posts: 2,512
Default

Lens distortion wasn't my point, and I don't think it was Lin's either. Compare using the best 50mm prime lens for the short lens vs a 200mm prime for the long lens (at four times the distance) to get rid of that issue. They should produce the same image of a two dimensional subject so I don't understand why one of them would need less resolution.
BillDrew is offline   Reply With Quote
Old Mar 25, 2004, 12:07 AM   #17
Moderator
 
Join Date: Jun 2002
Posts: 1,139
Default

Quote:
I don't understand that point. As an example, if a shot is made of a brick wall with a telephoto would require less resolution than one shot with a short lens while standing closer. (I used a brick wall as an example to avoid any issue of depth perspective since the subject is two dimensional).
Hi Bill,

It's the amount of actual real estate within the frame which determines resolution requirements. With wide angle we simply have more geography than with telephoto assuming all other factors are equivalent.

Obviously it's theoretically possible to move closer with wide angle to the point that frame content might be identical to a telephoto frame taken from a great distance, but that's not really the point. The more detail which is present within the captured frame, the more resolution is required to properly define the two dimensional boundaries of this detail,

Following is a really simplistic explanation - not designed for you because you already know this, but for those who have less understanding of resolution and interpolation.

Let's take a really simply example for demonstration purposes only. Let's assume our sensor has only 20 total pixels and we want to shoot a frame with a particular lens in which the subjects are uni-colored disc's. First let's use a narrow field of view telephoto which fills the frame with a single disc. Let's say we define the periphery of the disk with ten of our twenty pixels and use the other ten to sample the uniform color. Then we send this information to the interpolation engine which finds uniform color and a simple circle defined by these 20 pixels. The enlargement then resembles the original both in color and shape.

Now lets use a wide angle lens which reveals 20 of these colored disc's within the frame. We only have 20 pixels which lets us define position for 20 objects, but leaves nothing left to depict color. This represents the "marker" pixels concept. When we send this information to the interpolation engine the enlargement in no way resembles what we may have seen through the viewfinder because we simply had insufficient resolution to define either shape or color.

Obviously an extreme and simplistic example, but this is what happens less severely on a micro scale when we use low resolution captures on a wide angle of view frame. Our brains are perfectly capable of making some sense of the print whenever we view it from either a great distance or when it's printed very small. It's like comparing a halftone newspaper print of a group of people with a high resolution photograph. Though we may easily recognize the face of an individual we know in the newspaper halftone, if we subject the image to intense magnification we can't even recognize features like eyes or mouths apart from the Gestalt of the background and surrounding details. What we have instead of true detail are simply position marker pixels which when combined with all other pixels in the subject allow our brains to make some sense. If we take the high resolution print and use identical magnification, an isolated eye, nose or mouth still looks exactly like we expect it to.

The interpolation algorithm examines minute detail much as our magnifying glass. With low resolution captures of wide angle frames, insufficient pixels are frequently "trying" to describe details and end up being simply position marker pixels. When we view a small print or an enlargement at great distance, we can make sense of it because of our human ability to synthesize the illogical components into a whole as long as they are properly associated with the other adjacent position marking pixels. But the interpolation process faithfully enlarges these marker pixels and forces us to see them for what they really are and reveals the deception which we can no longer deal with.

Best regards,

Lin
Lin Evans is offline   Reply With Quote
Old Mar 25, 2004, 12:32 AM   #18
Moderator
 
Join Date: Jun 2002
Posts: 1,139
Default

A little more on this concerning angle of view. If we move a wide angle lens close enough to the subjet to equal a telephoto view covering the identical frame, we actually have equivalence in the angle of view in regards to the subject. That is in respect to the entire horizon the "angle" represented by the frames are equivalent.

Perhaps it would be more correct to use the term "Field of View" rather than "Angle of View" to avoid confusion and to describe the actual "geography" or "real estate" within the frame.

Best regards,

Lin
Lin Evans is offline   Reply With Quote
Old Mar 25, 2004, 12:49 AM   #19
Senior Member
 
Join Date: Dec 2003
Posts: 383
Default

I know there is more information in a greater angle of view shot but, if all things are equal (field of view) between a telephoto shot and a wide angle shot (compromising the exact same spatial dimension and the subject is three dimensional) will there still not be more detail/data in the wide angle shot?

Lin, I would like to get your opinion about sending too much data to the printer. Specifically can you send too much resolution to a printer and if yes, what will the effects be?

Thanks,
PKchopper is offline   Reply With Quote
Old Mar 25, 2004, 3:10 AM   #20
Senior Member
 
Alan T's Avatar
 
Join Date: Jun 2003
Location: Chester, UK
Posts: 2,980
Default Re: Size vs Resolution

Quote:
Originally Posted by ImKayd1
All my Oly's automatically download at a resolution of 72 do I need to change the resolution when I'm going to print?
I suggest you read through the recent thread at http://www.stevesforums.com/phpBB2/v...932&highlight= as well as this one, which has strayed a little.

Or you could just forget about the dpi/ppi setting on your camera, which is an anachronistic historical hangover. It's pixels that count.
Alan T is offline   Reply With Quote
 
Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off



All times are GMT -5. The time now is 9:09 AM.