Multiple Questions on Use of R-D1

[barjohn]

Richc,

Thanks, that is a clear explanation. Is this what makes the small P&S images look so sharp? Their COC's around .00x so when blown up on my 23" monitor, even at 100% they look very sharp. On the other hand, it doesn't explain why many of the images I see posted here in the galleries appear to be very soft focused since the images are typically pretty small when I look at them on my monitor (an Aple Cinema 23").

I am posting some images I took with a D-Lux 3 and the same basic image taken with the R-D1 so you can see what I mean. (the D-Lux 3 was taken with 28mm equivalent at 16:9 ratio) All images were as from camera scaled down to approx. 800x600.

The last two pictures are low light shots. The first was taken in a bar that was very dimly lit and is a handheld shot at f2 on a Minolta Rokkor 40/2. The Mission Inn shot was taken the same night at f2. Both at ISO 1600.

 

[jvr]

Richc,

Thanks, that is a clear explanation. Is this what makes the small P&S images look so sharp? Their COC's around .00x so when blown up on my 23" monitor, even at 100% they look very sharp. On the other hand, it doesn't explain why many of the images I see posted here in the galleries appear to be very soft focused since the images are typically pretty small when I look at them on my monitor (an Aple Cinema 23").

I am posting some images I took with a D-Lux 3 and the same basic image taken with the R-D1 so you can see what I mean. (the D-Lux 3 was taken with 28mm equivalent at 16:9 ratio) All images were as from camera scaled down to approx. 800x600.

The last two pictures are low light shots. The first was taken in a bar that was very dimly lit and is a handheld shot at f2 on a Minolta Rokkor 40/2. The Mission Inn shot was taken the same night at f2. Both at ISO 1600.

P&S images appear sharp because:

1) their small sized sensor implies the use of very short focal lengths (ex: 5.9 mm on my GR-Digital) to give you the same FOV from a bigger sensor/film. Even taking into account that the CoC is (much) smaller, the global effect is bigger DOF (see a good DOF calculator that takes into account sensor/film size)

2) sharpening by software is applied in VERY heavy doses in P&S, typically with some sophisticated algorithms to reduce noise. You can acheive the same effect using Unsharp Mask (or other sharpening filter) in Photoshop on RD-1 (or film) files.

3) a good AF, together with more DOF from 1) gives you sharper images. P&S are becoming better in this regard and several photos you see here from RFs are not particularly well-focused (even your 2 last photos seem misfocused: b&w one is globally fuzzy, colored one has distant lights in focus (center) so nearer lights seem "soft"=. Focusing an RF in low-light is not very easy, especially with fast lenses wide-open. Moreover, some RF lenses are just not well collimated, so focus is a bit off.

4) P&S tend to have, nowadays, anti-shake and ISO-auto (not no mention auto-flash...) that reduce camera shake. You won't get that from a RD1, even less from a film RF.

5) ...

So, if it looks like I'm saying that it's easier to get focused, sharp photos (at least on a screen or a 10x15 cm print) with a good, modern P&S than with a digital RF with a fast lens wide-open, yes, that's what I am saying...

This does not mean that I prefer a P&S to a RF (or a SLR). But I really can get very good results from even "humble" P&S, such as the old Nikon Coolpix 880 (picture attached). On a screen, it's difficult to tell them apart from my digital SLRs or the RD-1s...

But we like difficult things, that's why we keep using digital RFs...

 

[ampguy]

stunning RD1 photos

stunning RD1 photos

do you recall the lens you used for the table/fruit basket shot? What film settings do you use, or do you use "standard" ? Really great photos!

P&S images appear sharp because:

1) their small sized sensor implies the use of very short focal lengths (ex: 5.9 mm on my GR-Digital) to give you the same FOV from a bigger sensor/film. Even taking into account that the CoC is (much) smaller, the global effect is bigger DOF (see a good DOF calculator that takes into account sensor/film size)

2) sharpening by software is applied in VERY heavy doses in P&S, typically with some sophisticated algorithms to reduce noise. You can acheive the same effect using Unsharp Mask (or other sharpening filter) in Photoshop on RD-1 (or film) files.

3) a good AF, together with more DOF from 1) gives you sharper images. P&S are becoming better in this regard and several photos you see here from RFs are not particularly well-focused (even your 2 last photos seem misfocused: b&w one is globally fuzzy, colored one has distant lights in focus (center) so nearer lights seem "soft"=. Focusing an RF in low-light is not very easy, especially with fast lenses wide-open. Moreover, some RF lenses are just not well collimated, so focus is a bit off.

4) P&S tend to have, nowadays, anti-shake and ISO-auto (not no mention auto-flash...) that reduce camera shake. You won't get that from a RD1, even less from a film RF.

5) ...

So, if it looks like I'm saying that it's easier to get focused, sharp photos (at least on a screen or a 10x15 cm print) with a good, modern P&S than with a digital RF with a fast lens wide-open, yes, that's what I am saying...

This does not mean that I prefer a P&S to a RF (or a SLR). But I really can get very good results from even "humble" P&S, such as the old Nikon Coolpix 880 (picture attached). On a screen, it's difficult to tell them apart from my digital SLRs or the RD-1s...

But we like difficult things, that's why we keep using digital RFs...

 

[jlw]

Richc,

Thanks, that is a clear explanation. Is this what makes the small P&S images look so sharp?

Basically, yes. Small images always look sharper and appear to have greater DOF.

To understand why, you've got to remember that Rich was right: DOF is only an illusion.

The number of DOF calculators on the 'Net, the tendency of people to quote DOF figures down to three or four decimal places, and the impressive sound of terms such as "circle of confusion" sometimes make it seem as if DOF is an inherent, precisely-calculable property of lenses. It isn't. It's just a byproduct of the fact that we view photographs with a slimy, jiggly blob of jelly and tissue called the human eyeball!

Researchers in the human-eyeball field have done a lot of studies over the years of just how well the eye can see, and one consensus value that has arisen is that, for typical subjects under normal viewing conditions, the eye's ability to discriminate is limited to about three minutes of arc. A minute of arc is 1/60 of a degree, so three of them amount to 1/20 of a degree -- a very, very skinny sliver of a circle.

What it means is that if you're looking at a detail in a picture (or in real life) that covers less than 1/20 of a degree within your angle of view, you won't be able to tell whether that detail is or is not sharp.

Quoting this value as an angle neatly takes image magnification and viewing distance out of the equation. If you're viewing at a close distance, 1/20 of a degree covers a very, very tiny spot. If you're at the drive-in movies, 1/20 of a degree is going to be a pretty large patch of screen real estate. Either way, as long as a particular detail fits within the arms of this magic angle, you won't be able to tell whether it's sharp or not. Not being able to tell means you won't care, so an almost-sharp image will be just as good as a sharp one.*

That's why small prints look sharp: Their details aren't enlarged very much, so most of them fall under the minimum size for which your eye can discriminate between sharpness and unsharpness, when viewed from the same distances at which we customarily view larger prints.

The fact that it's what the eye sees that really counts also throws off "mathematical" DOF calculations in other viewing situations. For example, take an image you're viewing on your computer monitor. The monitor image is composed of raster dots formed by a shadow mask (tube monitor) or liquid-crystal elements (LCD monitor) and these dots aren't capable as forming quite as sharp an image as is obtainable on a printed sheet. (If you don't believe me, look at your monitor with a magnifying glass. See?) The eye judges overall sharpness by comparing the sharpest thing it can see to others that aren't quite as sharp, so it's more tolerant of slightly-unsharp details on a monitor... the result, greater apparent depth-of-field.

The same thing happens in printed reproduction, where the color or gray tones of the original image are reproduced by turning them into a grid of fuzzy-edged "halftone dots." In newspapers, these dots are spaced at a fairly coarse 65 to the inch -- loose enough for a healthy eye to see the individual dots at a close distance. The fact that the finest detail can't be any finer than one dot, again, makes the eye more tolerant of details that aren't quite sharp -- so, an image reproduced in a newspaper will seem to have more depth of field than the same image viewed at the same size as an original photo print. (This once caused a bit of embarrassment for me when I was a newspaper photographer, involving a bit of obscene graffitti on a background wall that I thought safely out-of-focus... but I don't want to go into that story right now!)

Effective DOF also can be influenced drastically by specific types of lens aberrations, and the resolving power of the recording and printing media, as well as the subject matter also play some role. (If you're photographing white sheep in a fogbank, there's so little detail that depth-of-field is basically irrelevant.) If nothing else, I hope all this detail has convinced you that depth-of-field is a much more slippery, vague concept than it sometimes appears to beginning photo-technicians!

Just remember that it's all based on the concept of the size of the finest detail that your eye can see in the final reproduction, compared to the size of that same detail in the original scene.

On-film (or on-sensor) circles of confusion, numerical apertures, focal lengths, and so forth are only aids to approximating this relationship. They all depend on various assumptions about viewing the final image, and their results may or may not apply to your situation depending on how well your conditions match those assumptions!

*If you want to calculate how big this patch is at various distances, without getting mixed up with trig functions and so forth, here's an easy way to do it:

First let's consider viewing something at close reading distance of, say, 10 inches away from your eye. This means you can picture your eye as being at the center of a circle with a 10-inch radius, and the print is at the edge of the circle.

It's easy to see that the diameter of this circle is 20 inches, and that means its circumference (distance around) is 62.83 inches, because C = pi * D. I'm going to call this value the "viewing circle."

But what we're interested in is not the whole viewing circle, but a sliver of it that amounts to 3 minutes of arc -- a very slender "pie slice" out of the total circle. We know there are 360 degrees in a full circle, and 60 minutes of arc in each degree, so one minute of arc is going to cover 1/(360*60) or 1/21,600 of the circumference. So, three minutes of arc would be 3/21,600, or 1/7200, of the circumference.

Since we know our circumference is 62.83 inches, 1/7200 of that equals 0.0087 of an inch -- a very tiny spot!

(For those so inclined, feel free to reproduce this calculation in metric units. The fractions don't change, since they're angular measurements -- just substitute millimeters for the viewing distance and "viewing circle" diameter.)

Meanwhile, back at the drive-in movie, you might be 200 feet from the screen. That's 2,400 inches, or a viewing circle of 15,080 inches! And 1/7200 of that is 2.1 inches -- which is slightly larger than a golf ball. Think of it: any detail on the screen that's the size of a golf ball or smaller, and you won't need to care whether it's sharp or not!​

 

[RichC]

Richc, Thanks, that is a clear explanation. Is this what makes the small P&S images look so sharp?.

Partly, but jvr list several more important reasons. In most cases, depth of field is deep enough not to be a worry (assuming your subject is in focus!).

There could be many reasons why some photos in the RFF galleries don't look sharp. For example:

• Most members are film users, so the online photos are scans - which may not be done well (whole books have been written on how to scan).

• Photos may be taken on old cameras/lenses going back to before World War II. Not only can they be awkward beasts to use, they may be in need of a service.

• Old lenses have a different look to modern ones - in practice a good old lens and a modern one may actually be of similar sharpness, but the latter looks a sharper because it's more contrasty: shadows are blacker, edges more defined.

• RFF members are useless photographer <joke!>

Regarding your two R-D1 photos, as jvr says, the colour photo is sharp, in the region of the second, more distant archway. The b&w photo should be sharper - I think it's soft because of camera shake.

The R-D1 does produce nice, sharp images, easily the equal of any compact camera. Here are a few examples. The first two are taken with a modern lenses, the last with a 1950s one. Note the detail in the shadows and the wider tonal range in the last photo.

 

[jvr]

Thank you, ampguy!

The roses were taken with my (now sold to get the D80) Nikon D70s and the AF 50/1.8, on a tripod. ISO 200, f8, 1/15. .NEF, converted in Photoshop CS2.

The fruits were taken as a test to my (then) new CV 12/5.6 Ultra Heliar (no tripod but hand supported on the arm rest of a sofa). ISO 200, 1/5 (at least that's what the file says...). Don't remember aperture but should be f8 or f11.

The fruits shot was taken much closer to the subject than it seems, the 12/5.6 coverage is awesome, even on the RD-1s (it's just beyond crazy on my M3! ). If you take the time, you'll see the typical "stretching" near the corners that comes from this kind of lens.

Shot in .ERF and converted in Photoshop CS2, detail set to 50 (my default for Epson Rd-1s files), all others parameters neutral. Vignetting correction set on 25 (the 12/5.6 vignettes a bit, but nothing real bad).

All pictures resized and sharpened for screen (HQ=75) on BreezeBrowser.

 

[jvr]

Partly, but jvr list several more important reasons. In most cases, depth of field is deep enough not to be a worry (assuming your subject is in focus!).

There could be many reasons why some photos in the RFF galleries don't look sharp. For example:

• Most members are film users, so the online photos are scans - which may not be done well (whole books have been written on how to scan).

• Photos may be taken on old cameras/lenses going back to before World War II. Not only can they be awkward beasts to use, they may be in need of a service.

• Old lenses have a different look to modern ones - in practice a good old lens and a modern one may actually be of similar sharpness, but the latter looks a sharper because it's more contrasty: shadows are blacker, edges more defined.

• RFF members are useless photographer <joke!>

Regarding your two R-D1 photos, as jvr says, the colour photo is sharp, in the region of the second, more distant archway. The b&w photo should be sharper - I think it's soft because of camera shake.

The R-D1 does produce nice, sharp images, easily the equal of any compact camera. Here are a few examples. The first two are taken with a modern lenses, the last with a 1950s one. Note the detail in the shadows and the wider tonal range in the last photo.

Yes, I totaly agree with RichC on the RD1 being able to produce wonderful photos, as sharp as any digital P&S I know. And add to that a much cleaner image, a beutiful tonal range, incredible "volume" sensation, etc, etc, when used with good, calibrated lenses and the correct technique (focus, camera shake, exposure).

I would even go a bit further: IMHO (and I stress the M from My and the H from Humble...), the RD1s can produce photos as good (and in some ways better) than those I get from my Nikon D80. This is even more interesting, as the D80 surprises me everyday with the quality leap that it represents over my D70s, in terms of "cleanliness" of image and, even more important, the way it handles highlights.

Interestingly enough (and here we go again... ) it's a fact that I have more "technical keepers" on the D80 than on the Epson. And it boils down to exposure, focus and camera shake (not megapixels...). Using the D80 with the 18-200 VR has been spoiling me on this. The D80 seems to nail the focus everytime, exposure tends to be bang on and the VR really allows me to go low on speed with surprising results (and since the apertures are small, DOF is bigger). Just because of this, the perceived "sharpness" of the photos I get from the 18-200 is so good that all my beloved Nikon primes stay more and more on the drawer...

Now, don't get me wrong: absolute sharpness is useless, and being technically correct is just a way of seeing better a lousy photo... And on that count, I would say I have a better percentage of "good" pictures on the Epson. But I have to be more concerned about technique: shoot RAW (so that exposure errors are dealt with and I can go up on ISO without grain), be VERY careful of camera shake, confirm focus very well when wide-open.

I bought my Epson less than one year ago and my pictures technical quality with it has been improving, as I get more used to the camera and the way lenses react on it. Which, by the way, is very different from the Leicas I have been using in the last 15 years: focus is more critical, camera shake too, bad lenses show up with a vengeance.

I still don't have, with the Epson, the almost "absolute confidence" I have with my M3 or with the D80 that the shot will be ok. Sometimes it's camera shake, sometimes is focus error, sometimes is bad exposure. Sometimes, it's all together...

But every once in a while I get a picutre that is so good, in every sense, with the Epson, that I always get back to it, instead of the M3 or the D80, unless I want to be absolutely sure I'll get the shot. My best pictures with the Epson have all the qualities I have with the Leica (ok, glass is the same... ) plus a kind of "transparency" and "smoothness" that good digital has and that only a full wet process (no scanning involved) can aspire to. And this is really addicting...

Ok, this has become way off topic!

 

[jvr]

BTW, RichC, congratulations again on your terrific site about the Epson!! And your photos too...

The last one, at least on thumbnail format, looked so much like a photo I took in Cannes some years ago (with the Konica Hexar AF and Sensia) that I can't resist attaching it.

Ok, the Hexar AF is not strictly a rangefinder but it has probably one of the best 35 mm "RF" lens I used...

 

[barjohn]

Gentlemen:

Thank you all for the excellent help and advise. This thread has a lot of informative information for any R-Da new user (like me). Now, if I amy trouble you all a little further, the three shots attached are a little different.

The first two shotw were of my monitor taken with the R-D1 on a tripod. The only difference is the lens used. The first is the CV 40/1.4 Nockton and the second is with the Minolta 40/2. The third picture was taken at a different time, handheld D-Lux3. Note the artifacts in the first two not present in the last. My questions are 1. Is there a way to not get the artifacts? 2. Which lens appears to be the better? and 3. How would you compare them to the D-Lux's Leica lens?

 

[mwooten]

barjohn,
Are you refering to the jpeg artifacts?

Michael

 

[barjohn]

All three are jpegs. The blue cross hatch visible in the first two is not visible in the third.

 

[mwooten]

Do shots taken in epson raw show the same cross hatch? Or is it just when shot in jpeg?

 

[jvr]

barjohn,

How are you producing these jpgs? They have a lot of artifacts! I don't get anything like it on any digital camera I own (and I own a few...).

It's impossible to compare lens on this kind of shot, with this size and these artifacts. Regarding the comparison between the Minolta and the Nokton (and also the Summicron-C 40/2), I got some full pics by a Epson RD-1 from a site and the Nokton was the sharpest and more contrasty lens. If that's what one's after, the Nokton is a "better" lens.

I post scaled down version of the photos (with a bit of sharpness). I can mail the intact full versions. Even in the resized smaller versions, it's easy to see the difference in rendering: the Nokton has more constrast, then the Rokkor, then the Summicron-C. That's to be expected: the Nokton is a "current-technology" lens (and a very good at that...), the Rokkor is older but multi-coated and, AFAIK, the Summicron-C is single-coated. Regarding sharpness and detail, again the order is the same, for the same aperture.

Some people complain about OOF rendering (bokeh) of the CV Nokton 40/1.4 and rave about the Rokkor's bokeh. In these pictures, I have to say I prefer (but that's higlhy subjective) the OOF rendering of the Nokton, it seems "quieter".

I don't have the 40/1.4 Nokton but I have the 50/1.5 and 35/1.2 Noktons and I have to tell you they are very, very good lenses.

Having said that, I really like the Rokkor rendering on film (never tested in the Epson)...

Sorry for using photos that are not mine without credit but I don't really remember the site!

 

[jvr]

And the Summicron-C pictures.

 

[barjohn]

Do shots taken in epson raw show the same cross hatch? Or is it just when shot in jpeg?

It is the same in raw.

I rank the Sumicron first, the Nockton second and the Rokkor 3rd though the differences are quite small, just a better tonality, tonal range and feeling from the Sumicron though the Nockton is close behind it and the Rokkor is close behind the Nockton.

 

[mwooten]

If it is just when you shoot photos of the monitor my guess it is from the monitor refresh. If it's on other shots of "static" object, then I would reckon that your camera has a problem.

 

[LCT]

As far as IQ is concerned, lenses can me compared by their resolution, contrast, distorsion, vignetting, color rendition, 3D rendition and OoF rendition ("bokeh").
None of those comparisons can be seriously achieved when you shoot a monitor or a TV screen.

 

[RichC]

BTW, RichC, congratulations again on your terrific site about the Epson!! And your photos too...

Thanx jvr!

If it is just when you shoot photos of the monitor

If it is just the monitor, it looks suspiciously like moire interference to me - like when net curtains sort of shimmer in the light and you see strange patterns.

If that is the case, the pixel grid on your TFT monitor and the light-well grid on your camera sensor are interfering optically with each other. This sort of situation occurs very rarely, but, if it does, the only solution is to clean up the image after the shot in an image-editing program, e.g. Photoshop. More info on digital cameras and this moire effects here: http://www.cambridgeincolour.com/tutorials/sensors.htm

Also, your JPG images are being saved at very low quality as they're full of JPG artifacts. I assume the files straight out of the camera look much cleaner.

By the way, the website on your monitor looks somewhat familiar!

 

[barjohn]

I don't get this on other objects and I agree it looked like moire interference to me too. The thing that was different was the fact that it formed a square grid pattern looking like it was engineering note paper with a transparent image laid on it, not just the rainbow shimmering effect.

As for quality, I took the original shots and scaled them down to either 1100x600 or 800x600 which reduced the number of pixels and quality, I did not try some other method such as cropping to keep original pixel density or some other compression method. I also think I set quality of the JPEG to low to keep the file sizes small. What is the recommended method for posting?

 

[jvr]

I don't get this on other objects and I agree it looked like moire interference to me too. The thing that was different was the fact that it formed a square grid pattern looking like it was engineering note paper with a transparent image laid on it, not just the rainbow shimmering effect.

As for quality, I took the original shots and scaled them down to either 1100x600 or 800x600 which reduced the number of pixels and quality, I did not try some other method such as cropping to keep original pixel density or some other compression method. I also think I set quality of the JPEG to low to keep the file sizes small. What is the recommended method for posting?

For posting I would use something like Photoshop to resize. At the moment, I'm very happy with BreezeBrowser, that does a very good job on resizing and sharpening for screen. They call it HQ - High Quality - sharpening and it looks different from what you get with Unsharp Mask. It looks a bit like sharpening via layer duplication, high pass, hard light in Photoshop but easier... I'm using 90 for quality on BBs scale (100 is limit) and quality isquite good, files have around 120k for a 800x600. If I go to 100, files go to 600k but not much better on screen (if you sharpen them afterwards, that's another matter).

Anyhow, I don't know how it would react on your image, the interaction between the sensor grid and the monitor grid (maybe with some refreshing issues added...) is really weird!