Home
SearchSearch MemberlistMemberlist RegisterRegister ProfileProfile Log in to check your private messagesLog in to check your private messages Log inLog in

r-d1 vs e-p1: opinions sought!!!
View previous topic :: View next topic  


PostPosted: Wed Apr 14, 2010 8:45 pm    Post subject: r-d1 vs e-p1: opinions sought!!! Reply with quote

ive had a hankering to pick up an epson r-d1. what's been holding me back is the 6mp count of the camera, and the fact that i have a 10mp olympus ep1 that can utilize similar lenses.

now i would really like to hear members thoughts on the following question: r-d1 and e-p1 (or panny gf1, same thing), with the same lens, shooting the same scene, at the same time: which will bring better results???

thanks for taking the time to weigh in!
tony


Last edited by rbelyell on Wed Apr 14, 2010 8:55 pm; edited 1 time in total


PostPosted: Wed Apr 14, 2010 8:47 pm    Post subject: Reply with quote

Let we see I am curious very much! Congrats for your new gem!


PostPosted: Wed Apr 14, 2010 9:36 pm    Post subject: Reply with quote

6MP RD1
I can keep posting them all day.
The actual Pixel size and inherent quality of the CCD in the RD1 is not touched by the EP1 IMHO. 6MP v 12MP is not really the point.
Best of luck in your search for an RD1.

CZ Opton Contax RF Sonnar 2/85

ZM Biogon 2/35


Last edited by F16SUNSHINE on Thu Apr 15, 2010 12:27 am; edited 1 time in total


PostPosted: Wed Apr 14, 2010 9:49 pm    Post subject: Reply with quote

andy the pictures are great, and i have seen many great rd1 pix on this site. but i was hoping for more of an explanation: you say 'actual pixel size' of rd1-what does that mean? you say '6mp vs 10mp' not the point-ok, i believe you, but help me understand the point. Smile


PostPosted: Wed Apr 14, 2010 10:50 pm    Post subject: Reply with quote

Mpix does not make a good photographer (nor do lenses). Being a good photographer cannot be bought!

Andy, love the pictures and it makes me smile to see my "EX" in such good hands!


PostPosted: Wed Apr 14, 2010 11:04 pm    Post subject: Reply with quote

wow klaus, not sure where that came from, but i never claimed to be a good photographer; just an enthusiastic one.

and the questions i asked andy were searching for more definition in his response, which i would bet have a more technical explanation.

if you want to say my question doesnt matter because a good photographer can get good results with either, youre probably right. but i think i--and i think attila--would also be interested in this slightly rephrased question which might meet with your approval: rd1 and ep1 take the same photo with the same lens taken by the same photographer-which yields better results?


PostPosted: Wed Apr 14, 2010 11:24 pm    Post subject: Reply with quote

Define "better" then first! Better in terms of what?? That your print size is usually at least 6ft wide and this being
the reason for "mo pix needed"? Most of my exhibits so far were done using a 6Mpix Nikon, no complaints so far...
OK they were only 3ft, maybe that was it then...

Well, I guess it gets obvious that I'm not a pixel-peeper - I plead guilty.

To wrap it up and here I simply reiterate what Andy had stated already: It must be different reasons not
just the Mpix to make a decision Pro or Con a certain camera. Noise at high ISO could be one for instance
- depending on your needs.


PostPosted: Wed Apr 14, 2010 11:53 pm    Post subject: Reply with quote

klaus you seem a little too worked up about this; it was a pretty simple and rather benign question, and the term 'better' can be whatever the person who responds wants it to be. being a crappy photographer, im probably not the best person to ask. attila seemed to think this was an interesting topic; why dont you pm him and maybe he can explain it.

btw, to the moderator, since these pictures were posted the type spreads way out horizontally on my screen in an uncomfortable way. can that be corrected?
thanks


PostPosted: Thu Apr 15, 2010 12:25 am    Post subject: Reply with quote

Thank you Klaus you are way to kind.
My love affair with the RD1 may be my longest running love affair ever Laughing

Tony sorry for the oversize files. These are old images I have handy online. I did not realize that some were so big.
[EDIT: pic resized for Andy]

Here is my opinion and that of others.
Lots of pixels in a small area = poor DR and low light/high ISO production.
The smaller the pixels the noisier the resulting image.
The magic number with current technology is under 4mpx per square CM.
The epson has 1.6 MP/cm pixel density as does the Nikon D70 and some other earlier CCD cameras.
Canon 5Dii 2.4 MP/cm pixel density.
Olympus EP cameras 5.1 MP/cm pixel density.

I'm not an expert but it makes sense to me. Google for more info.
A friend sent me a Raw File from the EP1 for me to play with.
I can tell you that it was not possible to recover a poor exposure as well as I can do with the 2003 era file from the Epson sensor.

The bigger question for you is do you like the RF workflow.
It is certainly not for everyone and has many limits.
No Macro, no long lenses, lenses are expensive ($300 is considered cheap even for used lenses).
Sure you can buy FSU lenses but, it's a crap shoot and most often just plain crap.

I would suggest buying a decent FL RF and shooting 20-50 rolls.
Or just jump for an RD1. They are easy to sell as any used camera available.

I'll dump those big files now. Smaller versions are somewhere on the site or if you want I can send you a raw to mess around with.


PostPosted: Thu Apr 15, 2010 12:36 am    Post subject: Reply with quote

andy thanks for the detailed reply, its very appreciated. you know, when i first got the ep1 a couple months ago, i was very disappointed with the IQ and i posted some thoughts that the sensor was too small to support 10mp, and that i found better images if i dropped down to an 8mp quality level. wow, i got hammered! but just today i took the same scene at 10 and 8 mp and i liked the 8 much better. i dont shoot raw on it because of this.

again, i have seen great pix by you and others w epson, but you guys are all using very expensive lenses. if you put those lenses on the ep1, is it still your feeling the epson IQ would be demonstrably better--you know, to the naked eye better?

also, you originally said something about the 'actual pixel size' of the epson. ive heard this elsewhere that its really like 13mp or something. is that like the foveon implementation or something else?


PostPosted: Thu Apr 15, 2010 12:45 am    Post subject: Reply with quote

Yes it is my opinion and observation.
That said. I have seen some brilliant work with m4/3 cameras and M lenses.
As Klaus suggested. The camera and lens don't make the photographer.
I'll go one more and suggest that practice and protracted, thoughtful observation can.


PostPosted: Thu Apr 15, 2010 6:34 am    Post subject: Reply with quote

F16SUNSHINE wrote:

Here is my opinion and that of others.
Lots of pixels in a small area = poor DR and low light/high ISO production.
The smaller the pixels the noisier the resulting image.

This is not right.
Pixel size is almost irrelevant - the only thing that increases a little at image level with smaller pixels is noise in the deepest shadows. However, at the same time the signal to noise-ratio elsewhere increases.

In general pixel size is rather irrelevant noisewise - what counts is sensor size and the levell of technology used.

Quote:

The magic number with current technology is under 4mpx per square CM.
The epson has 1.6 MP/cm pixel density as does the Nikon D70 and some other earlier CCD cameras.
Canon 5Dii 2.4 MP/cm pixel density.
Olympus EP cameras 5.1 MP/cm pixel density.

This is absolutely not based on evidence and is simply false. The technology has improved and even changes from the CCD of D70 to modern CMOS-sensors.

There is no general optimum - what the optimum for given technology is also depends on what the sensor is optimized for - there are many parameters, and pixel size is just one of them.

DPReview used to have this campaign against increasing pixel density - too bad the folks at DPR are clueless when it comes to sensor technology. For example John Sheehy demonstrated well (on DPR forums), how the tiny sensors of pocket cameras are actually more efficient than the DSLR-sensors, even though they have tiny pixels!

Quote:

I'm not an expert but it makes sense to me. Google for more info.
A friend sent me a Raw File from the EP1 for me to play with.
I can tell you that it was not possible to recover a poor exposure as well as I can do with the 2003 era file from the Epson sensor.


That sensor has very bad high ISO by modern standards. I am also pretty certain it's low ISO is also not too good (relatively speaking). It's only advantage over m43-sensors is it's physical size - it is significantly larger, though almost certainly this will not be enough to give it better, or even equal signal/noise-ratio.

Two advantages the Epson has is in the amount of bokeh and the less cropped image compared to what the lenses used were designed for.

Also, if the camera feels good for you, and gives you pleasing images, who cares how antiquated it's sensor is Wink I myself would love to have the Epson - the crop-factor (and bokeh-factor Smile ) alone makes it very tempting. Too bad I'll never be able to afford to buy one (though it seems that for me the Epson will become an obsolete option for me soon as more APS-C (or larger) EVF-cameras start to appear on the market).


PostPosted: Thu Apr 15, 2010 6:48 am    Post subject: Reply with quote

rbelyell wrote:

also, you originally said something about the 'actual pixel size' of the epson. ive heard this elsewhere that its really like 13mp or something. is that like the foveon implementation or something else?


It is a regular sensor, and even when it was new, it wasn't that good. By today's standards it is very bad. I also imagine that close to the edges and corners of the sensor the IQ will deteriorate as more light hits the sensor at an angle that the sensor can not register or may even register into a wrong photosite.

Still, I am not saying that it can't be a wonderful tool. I for one would love to own one.


PostPosted: Thu Apr 15, 2010 1:13 pm    Post subject: Reply with quote

I had a feeling you would pop up Anu.


PostPosted: Thu Apr 15, 2010 4:24 pm    Post subject: Reply with quote

F16SUNSHINE wrote:
I had a feeling you would pop up Anu.


So sorry Wink


PostPosted: Thu Apr 15, 2010 4:45 pm    Post subject: Reply with quote

Anu wrote:
F16SUNSHINE wrote:
I had a feeling you would pop up Anu.


So sorry Wink


Don't be. Your absolute tones are always amusing for me Laughing

I still think your wrong. Theory against naked eye observation again.
My eye only cares about what it is looking at. So apparently do the eyes of others who feel the same way about this argument.
That appears to be Canikon who have gone to larger and fewer pixels in some of their latest offerings.
I surely don't imply that the D70/RD1 sensor is a marvel to hold all sensors against.
It does produce better "looking" files than many cameras that are more modern.
It is better than the EP1 (The OT in this case). Likely at least partially for the reasons I stated.


PostPosted: Thu Apr 15, 2010 5:15 pm    Post subject: Reply with quote

F16SUNSHINE wrote:

I still think your wrong. Theory against naked eye observation again.
My eye only cares about what it is looking at.


I think the problem is that you make statements about the theory and the technology and when those statements are argued against on scientific grounds, you revert to claiming that only the observed image matters. It's a perfectly valid viewpoint that what looks good to you is what makes a good camera for you, and I fully agree, and it looks to me that Anu does too.

However, once you claim that the camera's subjectively pleasing image is because of things like pixel density, you enter the realm of the theory and there it is possible to prove many such claims false (which is inherently absolute, resulting in very absolute statementsalthough admittedly statements such as very bad by modern standards are questionable due to no definition for very bad and lack of consensus on modern standards =).


I would suggest to just enjoy the image quality of the R-D1 (which looks quite pleasant) and let the images convince others of it, but not try to argue its inherent superiority to other cameras based on technical details unless you can support them with more than speculation. They are not matters of opinion.


PostPosted: Thu Apr 15, 2010 5:19 pm    Post subject: Re: r-d1 vs e-p1: opinions sought!!! Reply with quote

rbelyell wrote:
i
now i would really like to hear members thoughts on the following question: r-d1 and e-p1 (or panny gf1, same thing), with the same lens, shooting the same scene, at the same time: which will bring better results???


The larger sensor of the R-D1 is a huge bonus for using adapted lenses due to less crop factor. Of course there are modern APS-C-sized EVIL cameras coming, so if that is the deciding factor I would see about those.

However, for me, the really attractive thing about the R-D1 would be that it's a genuine digital rangefinder, not just EVF.


PostPosted: Thu Apr 15, 2010 5:36 pm    Post subject: Reply with quote

arkku, if theory is absolute, how do you reconcile your positive feelings about the rd1 sensor vs anu's negative feelings re same? not trying to be a smart ass, just trying to understand...


PostPosted: Thu Apr 15, 2010 8:22 pm    Post subject: Reply with quote

anybody else have an rd1/ep1/gf1????
would love to have some others weigh in!!!!


PostPosted: Thu Apr 15, 2010 9:01 pm    Post subject: Reply with quote

rbelyell wrote:
arkku, if theory is absolute, how do you reconcile your positive feelings about the rd1 sensor vs anu's negative feelings re same? not trying to be a smart ass, just trying to understand...


I don't have any feelings about the camera's sensor nor do I claim that there is some absolute theory by which to judge the subjective quality of images produced by a camera. I also didn't say that the sensor is good, nor did Anu say that the camera is bad (she even said she would love to own one).

I just look at the sample images posted and I think they are good enough, subjectively, without pixel-peeping, at least at this ISO, at this relatively small size (which is probably good enough for web use, amateur photography, etc). I also consider the large sensor (not large pixels) a major benefit, due to focal lengths of vintage lenses being made for full frame.

However, the claims made about why the camera supposedly has a better sensor than the proposed alternatives (i.e. larger pixels) are just speculation, and I this part in particular:

F16SUNSHINE wrote:
Lots of pixels in a small area = poor DR and low light/high ISO production.
The smaller the pixels the noisier the resulting image.
The magic number with current technology is under 4mpx per square CM


Is just false, and the magic number is downright ridiculous (as the term magic number should already suggest). To be fair, he said that it was an opinion, but that's a poor excuse considering that things like dynamic range and noise can be measured and therefore it is not a matter of opinion. I would speculate that these ideas are based on pixel peeping at 100% (also promoted by dpreview in their silly tests), which, when the pixel counts of the camera are not equal, is the same as making prints of two different sizes and then comparing their quality under a loupe to assess the filmanyone can see that it's not a fair comparison, but somehow some people have the idea that there's something magical about the 100% pixel size of an image as though it was a realistic size at which to view images. Then some, quite frequently the same people who refuse to make a fair comparison, claim that it's not the pixel count that matters when choosing a camera (and it isn't, which is why I don't see why they think pixel density is important).

(Note that I'm not referring to any particular person in the follow-up rant there; most of the claims I accuse some people of were not made in this thread nor do I mean to suggest that F16SUNSHINE is necessarily one such person.)


In any case, I think it's pointless to assess a camera, especially a special one like a digital rangefinder, based on pixel count or pixel density. If the sensor is of a good size for your lenses, the images are large enough, and look good enough for what you'd be using them for, why worry about some other camera being better by some measure that's irrelevant to you? Take the camera you have more fun using (of course after limiting your choices by whatever practical requirements you have).


Last edited by Arkku on Fri Apr 16, 2010 12:46 pm; edited 1 time in total


PostPosted: Thu Apr 15, 2010 9:10 pm    Post subject: Reply with quote

your reply is interesting, though intuitively i have a conceptual problem thinking that pixel density at some point does not adversely effect picture quality, but perhaps that is not what you are saying.

as for why compare cameras, because that's one of the main things we do here, compare stuff, whether its one lens vs another, one type of film vs another or two cameras. if i had a dime for every such comparison ive seen in the past 6 months on this forum i would be too wealthy to spend my time writing posts!

i think its fun and interesting to get people's points of view. hopefully you do as well, which is why you bothered to look at a thread that screams comparison in it's title. i also find it informative. if i'm thinking of spending money on an rd1, i would like a cross section of opinion on whether it's IQ is going to be less than a similar camera i already own.

there are independent reasons besides IQ to own an rd1 over an ep1, or gf1, but i really dont care so much about those, i care about IQ. so that's why i started the thread.


PostPosted: Fri Apr 16, 2010 12:59 am    Post subject: Reply with quote

rbelyell wrote:
your reply is interesting, though intuitively i have a conceptual problem thinking that pixel density at some point does not adversely effect picture quality, but perhaps that is not what you are saying.


There are many advantages in having more pixels, up to a limit.

The diffraction limit (or the limitations of a lens) cannot be overcome by adding pixels, i.e. at some point increasing the pixel density will be pointless. (But not necessarily harmful from a purely image quality point of view.)

Higher pixel density does not by itself make a camera perform worse. Yes, some cameras with higher pixel density perform worse in at least one aspect than a similar camera with lower pixel density. There are also counter-examplesthe quality of a sensor depends on many aspects, it cannot be judged by looking at a single figure, like the pixel density.

Comparing quality, e.g. high ISO performance, of images of different sizes is not fair or realistic. Many examples of high density leading to poor performance are based on such comparisons (e.g. all dpreview tests that I've seen). See my previous post for why it is wrong. Doing a fair comparison is a pre-requisite for drawing any meaningful conclusions.

A fair comparison does not automatically make higher pixel density winlike I said, it's more complex than that. However, e.g. with colour filter array sensors (pretty much all DSLRs/EVILs except Foveon) the image with more pixels already gains gets an advantage from having demosaicking artefacts smaller in proportion to the image size. And it's possible to simulate larger pixels by combining many smaller ones; e.g. skip demosaicking altogether and treat each GRGB group of four pixels as one full-colour pixel. It can also be done more intelligently than that, but even the trivial implementation has observable advantages at the expense of halving each pixel dimension of the imagegoes to show that (a smart version of) this becomes a viable alternative to demosaicking once the pixel count is sufficiently high to not lose significant amounts of resolution.

For some reason, many test sites and some users these days seem to equate high ISO noise with the image quality of a camera/sensor. It's more complex than that. =)

rbelyell wrote:

as for why compare cameras,


I didn't say anything against comparing cameras; I just said that it's pointless to assess a camera based on pixel count or pixel density. Look at real world images and compare those. If the pixel count is enough for you (e.g. you don't need more to print at a given size or to meet the requirements of a stock photo agency), why should you agonise over this aspect?

rbelyell wrote:

if i'm thinking of spending money on an rd1, i would like a cross section of opinion on whether it's IQ is going to be less than a similar camera i already own.


What is your definition of IQ? It's impossible for anyone else to know, that's why there are very lengthy tests of cameras, to try to show all aspects. For example, if you don't shoot high ISO, why would high ISO performance matter to you? If you don't print above a certain size, why should pixel count matter to you? etc.

If someone else says I tried cameras A and B, B was unusable., it does not mean that camera A might not still have the best image quality for you. For example, that person might have only tried a high ISO setting that they regularly need, but the situation might be reversed at low ISO

This is why I suggest comparison by looking at images similar to ones you would be using the camera for and judging the quality from those.

rbelyell wrote:
there are independent reasons besides IQ to own an rd1 over an ep1, or gf1, but i really dont care so much about those, i care about IQ


Personally I think the larger sensor is a deciding factor in image quality between these cameras; it allows you more freedom with shallow depth of field, and lets you get away with lower-quality lenses (as any problems will be smaller in proportion to the image sizeagain a reason why meaningful comparisons cannot be made at 100% of different images, and why old photos taken on 45 and 810 inch sheet film look incredibly detailed despite using ancient, simple, uncoated lenses).


Last edited by Arkku on Fri Apr 16, 2010 12:44 pm; edited 1 time in total


PostPosted: Fri Apr 16, 2010 2:29 am    Post subject: Reply with quote

This debate could go on for some time.
If you search online you will see there are 2 camps. Both feel they are correct.

I understand why Anus and Arkkus opinions are so strong that I am absolutely wrong. I've read both sides of the argument.
The fact is I have yet to see a high ISO image from a "over-dense" sensor that looks clean or at the very least grain like in it's noise.
The subject came up for me at a party in Seattle last fall. My RD1 got some attention from some Adobe nerds in attendance (guys that work for Adobe).
Believe me guys there was little consensus among these 3 fellows other than the feeling that the technology has a way to go.
The best anology I heard was from my friend Mika an Adobe tester.

"You can look at a window pane and believe it is flat. If you actually measure it you will find it is not flat nor is it even in thickness."
A photo from a low res sensor can in fact look cleaner. The laboratory can prove this wrong but our eye tells us different.
I have yet to see a pic from a sensor higher than 4mp per cm2 that looks clean at high ISO.
I have seen plenty of frames from the 5D and 5Dii Much lower than 4mp per cm2) that are high ISO and look quite clean.
I trust my eyes when it comes to aesthetics not lab measurements.

Of course the next post will be in conflict with this and chop up what has been said.
That's fine have fun with it. For those who want to capture images.
Trust your eyes not the Pundits.

OK, chop away Very Happy


PostPosted: Fri Apr 16, 2010 8:15 am    Post subject: Reply with quote

F16SUNSHINE wrote:
This debate could go on for some time.
If you search online you will see there are 2 camps. Both feel they are correct.


Well, one camp feels and bases it's evidence on 100% crops of sensors with different pixel counts. The other camp analyzes the normalized images.

F16SUNSHINE wrote:

I understand why Anus and Arkkus opinions are so strong that I am absolutely wrong. I've read both sides of the argument.


Reading is not enough. One has to follow the trail of logic too.

If you measure where the water drops, and you divide your large water collecting boxes into smaller water collecting boxes, your signal to noise ratio goes up, not down. This is of course very simplified example as there are other issues than just photon collecting involved.

I recommend you go to dpreview forums, search for people like ejmartin, John Sheehy and perhaps even Eric Fossum (I'm not sure if he's touched this issue). All those three know a lot more on this than you or I will ever know. I especially enjoy ejmartin's texts - John may be a bit agressive in his style and Eric tends to remind people a bit too often of him inventing the CMOS sensor Smile

F16SUNSHINE wrote:

The fact is I have yet to see a high ISO image from a "over-dense" sensor that looks clean or at the very least grain like in it's noise.


As it's been demonstrated by mr. Sheehy, having and DSLR-sized sensor fiilled with compact camera pixels, one would easily blow away the current DSLR-sensors at high-ISO. Having such a sensor would mean mayne 100 million pixels on APS-C and awesome pixel density Smile

F16SUNSHINE wrote:

I have yet to see a pic from a sensor higher than 4mp per cm2 that looks clean at high ISO.


Which DSLRs are those? Have you compared them to images from DSLRs with same sensor size? How did you compare?

F16SUNSHINE wrote:

I have seen plenty of frames from the 5D and 5Dii Much lower than 4mp per cm2) that are high ISO and look quite clean.
I trust my eyes when it comes to aesthetics not lab measurements.


I guess the fact that 5D and 5Dii (which btw is far superior at all ISOs to 5D even though it has twice the pixel count) are full frame sensors does not matter? Wink

It is not the pixel size, but the sensor size that is relavent.