Need a opinion on this picture.

Started 3 months ago | Discussions
fferreres Senior Member • Posts: 2,333
Re: Need a opinion on this picture.

Guy Parsons wrote:

lomoapontaechuta wrote:

I heard, read and saw that film has an equivalent resolution of 3 MP.

That figure is a bit low. From my experience with 35mm film and experiments comparing film to early digital cameras, I came to the conclusion that at about 10MP digital equalled film. Olympus later said 10 to 12MP was the target to beat 35mm film.

In practice (usually then printing for camera club competitions) it worked out that at 8MP digital looked better than film due to the cleaner result. But digital files when printing definitely have a hard size limit where they just looked "empty" if printed too large, while film printed too large had all that lovely grain to hide the fact that the resolution wasn't there.

I'm PP your colour shot for several days now to have a benchmark next time I go out.

thank you

Frankly, it's beyond me why anyone would want to mess up a digital image to make it as bad as film.

Regards.... Guy

Quick question, is there any process to transform from pixel (squares) to circles?

Obviously, a sensitive 2D surface which isn’t a grid will register things as circles, without any alignment to horizontal or vertical grids. But in a camera, they would be aligned. However, on paper, and in our retina where the rods and cones are not square but more akin to circles, we could say that film “look” more real in a sense, or more pleasing. The act of being made of circles matches our neuronal functions. Is there a process to scale things in a way that treats squares as circles? For example, each pixel is a circle that don’t overlap but where they don’t we place a circle that the average of the four adjacent ones (maybe not covering all of it just a good portion) which leaves I little still uncovered, and again we average on each axis the new dot with those 2 adjacent in that axis, and use a smaller dot there, etc? Would this look better if printed large?

Guy Parsons
Guy Parsons Forum Pro • Posts: 32,510
Re: Need a opinion on this picture.

fferreres wrote:

Quick question, is there any process to transform from pixel (squares) to circles?

Obviously, a sensitive 2D surface which isn’t a grid will register things as circles, without any alignment to horizontal or vertical grids. But in a camera, they would be aligned. However, on paper, and in our retina where the rods and cones are not square but more akin to circles, we could say that film “look” more real in a sense, or more pleasing. The act of being made of circles matches our neuronal functions. Is there a process to scale things in a way that treats squares as circles? For example, each pixel is a circle that don’t overlap but where they don’t we place a circle that the average of the four adjacent ones (maybe not covering all of it just a good portion) which leaves I little still uncovered, and again we average on each axis the new dot with those 2 adjacent in that axis, and use a smaller dot there, etc? Would this look better if printed large?

I'm really not sure what the problem is that you see.

When printing, the result is not simply pixels on paper but a random splatter of micro dots all over and about the position of that notional pixel. If anything that has made your assumed square pixel into a round one that merges gently into the adjacent pixel.

The thing to remember is that it takes around 300 original camera pixels per inch of print to look sharp and detailed, down to a sensible low limit of around 200 of those pixels per inch to still look reasonably good.

The actual printing process in printer drivers reorganises the pixels (whatever the file being fed to the printer) by interpolation into 600 pixels per inch for Canon and Epson into 720 pixels per inch so that is the density of those new round pixels on the paper. The human eye simply can't resolve that resolution, it all simply looks like smooth printing with no discernible "pixels".

Regards..... Guy

 Guy Parsons's gear list:Guy Parsons's gear list
Sony RX100 VI Olympus PEN E-PL1 Olympus PEN E-P3 Olympus PEN E-PL5 Olympus PEN E-P5 +4 more
Tony10 Regular Member • Posts: 147
Re: Need a opinion on this picture.

The comments here for me just highlight the issues around the film/digital debate, I get the feeling we are supposed to react with awe when shown a photo taken on film but in reality it's irrelevant because it's image content & impact that matters.

PhotoNerd180 New Member • Posts: 1
Re: Need a opinion on this picture.

It's a nice frame. The noise level looks a bit too much though.

fferreres Senior Member • Posts: 2,333
Re: Need a opinion on this picture.

Guy Parsons wrote:

fferreres wrote:

Quick question, is there any process to transform from pixel (squares) to circles?

Obviously, a sensitive 2D surface which isn’t a grid will register things as circles, without any alignment to horizontal or vertical grids. But in a camera, they would be aligned. However, on paper, and in our retina where the rods and cones are not square but more akin to circles, we could say that film “look” more real in a sense, or more pleasing. The act of being made of circles matches our neuronal functions. Is there a process to scale things in a way that treats squares as circles? For example, each pixel is a circle that don’t overlap but where they don’t we place a circle that the average of the four adjacent ones (maybe not covering all of it just a good portion) which leaves I little still uncovered, and again we average on each axis the new dot with those 2 adjacent in that axis, and use a smaller dot there, etc? Would this look better if printed large?

I'm really not sure what the problem is that you see.

When printing, the result is not simply pixels on paper but a random splatter of micro dots all over and about the position of that notional pixel. If anything that has made your assumed square pixel into a round one that merges gently into the adjacent pixel.

The thing to remember is that it takes around 300 original camera pixels per inch of print to look sharp and detailed, down to a sensible low limit of around 200 of those pixels per inch to still look reasonably good.

The actual printing process in printer drivers reorganises the pixels (whatever the file being fed to the printer) by interpolation into 600 pixels per inch for Canon and Epson into 720 pixels per inch so that is the density of those new round pixels on the paper. The human eye simply can't resolve that resolution, it all simply looks like smooth printing with no discernible "pixels".

Regards..... Guy

Since nature is “round”, the eye cones are “rounded” and the observation from the previous poster, I was wondering if our camera sensors wells are square, and if the film noise could be interpreted (perceptually) as having more resolution. After all it seems that film is building the scene with little circles. The grain, if any is visible at certain magnification, if always circles, never squares.

Guy Parsons
Guy Parsons Forum Pro • Posts: 32,510
Re: Need a opinion on this picture.

fferreres wrote:

Since nature is “round”, the eye cones are “rounded” and the observation from the previous poster, I was wondering if our camera sensors wells are square, and if the film noise could be interpreted (perceptually) as having more resolution. After all it seems that film is building the scene with little circles. The grain, if any is visible at certain magnification, if always circles, never squares.

In ink-jet printing it will always be circular blobs of ink splattered around to give the impression of the content of the pixel that is in that position. It is no longer "square" or "round" but a pseudo random mist of tiny droplets, that then may disperse into the layers of the print medium to deliver a smoother result.

The data that comes from the sensor has no squareness to it, it is simply data about luminance and colour from each plotted pixel position. How it is represented after that can change.

On a monitor the pixels are rectangular (3 RGB rectangular pixels to make up a square pixel on the screen) but you will only see that up really close and only see the camera's pixel representation if you have 100% display, that is simply like looking at a 5 foot wide print with your nose against it, a bit unrealistic don't you think?

I totally fail to understand how the eye's rods and cones can have any interaction with the pixels that the camera produces. The eye is constantly moving and scanning and the brain does 90% of the work by interpreting what the eye gathers into some sensible version of vision.

If you are seeing square pixels in a print then that's extremely bad printing.

Regards.... Guy

 Guy Parsons's gear list:Guy Parsons's gear list
Sony RX100 VI Olympus PEN E-PL1 Olympus PEN E-P3 Olympus PEN E-PL5 Olympus PEN E-P5 +4 more
fferreres Senior Member • Posts: 2,333
Re: Need a opinion on this picture.

Guy Parsons wrote:

fferreres wrote:

Since nature is “round”, the eye cones are “rounded” and the observation from the previous poster, I was wondering if our camera sensors wells are square, and if the film noise could be interpreted (perceptually) as having more resolution. After all it seems that film is building the scene with little circles. The grain, if any is visible at certain magnification, if always circles, never squares.

In ink-jet printing it will always be circular blobs of ink splattered around to give the impression of the content of the pixel that is in that position. It is no longer "square" or "round" but a pseudo random mist of tiny droplets, that then may disperse into the layers of the print medium to deliver a smoother result.

The data that comes from the sensor has no squareness to it, it is simply data about luminance and colour from each plotted pixel position. How it is represented after that can change.

On a monitor the pixels are rectangular (3 RGB rectangular pixels to make up a square pixel on the screen) but you will only see that up really close and only see the camera's pixel representation if you have 100% display, that is simply like looking at a 5 foot wide print with your nose against it, a bit unrealistic don't you think?

I totally fail to understand how the eye's rods and cones can have any interaction with the pixels that the camera produces. The eye is constantly moving and scanning and the brain does 90% of the work by interpreting what the eye gathers into some sensible version of vision.

If you are seeing square pixels in a print then that's extremely bad printing.

Regards.... Guy

If it’s already at a resolution way above I can see, I wouldn’t be talking about it. My question is a bit more technical and curiosity related, so you need to assume that my eye is not outresolved. For example, how does an Inject printer decide the droplets, if these are circles and overlap? Does any algorithm optimize how to lay out things on an Inkjet (nozzle level) vs RGB screen vs other medium? Probably, square things are the least likely to exist in nature too.

Keyboard shortcuts:
FForum MMy threads