Back when I was printing 9 foot prints and big lighted chromes like Duratrans and using large and format like the 617.....I could clearly see the advantage of larger film (and cameras). Today my largest exhibition of images are on slide/sound shows with LED screens about 55 inches and lesser desirable canon projected images which can be say 9 ft x 12 ft. The difference between FX and DX digital Nikons is just not there. In the field our D700 was not getting used much. D90, D300, D7000 and even our D200 are all producing images impossible to tell the FX images from the DX images. Recently I received a reply from Nikon about DX versus FX. I do not think Nikon is going to abandon DX. I do feel my least useable of all Nikon DSLRs is the D40X and yet some photos taken with that were important and the bigger cameras got left that day due to field hazards and the chance of drowning a camera. I think the next D7100 or D400 will outsell even the possible D600 although that may be a especially good camera. I do believe not every one is going to buy or use a D800. So I am sure FAR from assuming FX is better.
Dx v. Fx Why is full frame better?
(96 posts) (35 voices)-
Posted 10 months ago #
-
I note that DxOMark ranks the new DX D3200 sensor equal to the FX D3 sensor. Imagine that, the "bottom of the line" DX sensor is equal to last generations base FX sensor! Check it out yourself. Hard to believe, but true. Surely the sensor Nikon is putting in the D7200 and the D400 will be better than the "bottom of the line" one in the D3200. Maybe a D400 sensor equal to the D3x or the D3s? Hope so.
Posted 10 months ago # -
FX sensor has more than double the area of a DX sensor.
Posted 10 months ago # -
donaldejose said:
I note that DxOMark ranks the new DX D3200 sensor equal to the FX D3 sensor.One must be very careful in equating a "Total" score with quality. The argument was made with the 5d that the high resolution was giving it an inflated score, I would say that is happening here and with even the D800. To me, I think the common press puts way too much focus on resolution and that is reflected in DXO's and others tests which inflates "scores" but looses focus on quality. The tests also pick and choose "focus" areas. Many have argued the focus areas are too few or the relevancy placed on them is limiting. Not to get off topic or into a debate about that, but it must be stated. The D3200 is not going to be at the AF, WB, Metering, Durability, etc. of a D3. I'm sure it is a good camera, but the numbers only reflect a very, very small portion of what stacks together to achieve a great image.
The other thing that has to be pointed out is that most all (but the Dx00 line) have parsed down AF, Metering, processing, Auto WB, etc. which adds to the IQ as well. FX cameras are PRO-level systems. Start downgrading the back-end stuff, and I believe the difference will be barely visible. Maybe not this time around, but in 3-4 years, I doubt there will much of a difference in 90% of shooting situations. Because of the period of technology we are in (first couple iterations of FX), there is a difference - that gap will close.
This is my main "wait to be seen" curiosity of the D600. Processing, metering, WB, AF of a D7000 with an FX chip, in my mind is just a D7000 with more resolution. It's the entire guts that make the camera, not just one part (sensor) that helps us capture what we envision. It will work for many, but will not quite get there for others.
Posted 10 months ago # -
TaoTeJared has I believe pretty much stated this issue as well as can be done. I would take some exception to DaveO's comment on sensor size of the FX compared to APS-C sensor size. In terms of the relative size, yes. But recalling results of shooting 8x10 for years and then looking at 35mm shots taken of the same subjects with a Nikon F5 I do believe that simple comparison of surface size of the sensor or film size leaves a little to be desired.
8x10 film size compared to a 35mm frame (24mmx36mm) would seem to dictate that the 8x10 would rule in every comparative shot. This was actually not the case at all. Sometimes the 35mm image would be far superior to the 8x10. Same for 617, 6x9, etc. The cameras and lens in the medium and large formats was about as good as it gets. There is a host of things going on when we photograph from catching the right moment, getting the exposure right and all the other requirements.
I have seen some stunningly good FX digital images. I have also seen some DX images taken under conditions that larger formats would be hard pressed to be able to pull off. One of the worst issues for FX Nikon filed use is Nikons dearth of really practical FX field glass.
In large format lens 5.6 and even F8 might be the starting point on fine lens. I fail to see why 400mm and 500 mm lens with Nikkor labels are not more available in 5.6 or even narrower apertures.
Posted 10 months ago # -
"I fail to see why 400mm and 500 mm lens with Nikkor labels are not more available in 5.6 or even narrower apertures."
Perhaps it is an historical artifact. In the past days of slow film, you needed a large f-stop when you wanted to use a fast shutter speed to freeze motion because you were forced to work with a low ISO. In the early days of autofocus you needed a fast f-stop to put enough light on the focus sensor. Only today are we now getting cameras which give us good image quality at high ISO and which can focus well at F5.6 or F8. Today it would be possible for Nikon to produce 400mm and 500 mm lenses in 5.6 and we could use them. Perhaps they will. The need for big fast glass is decreasing with advancements in sensor technology.
Posted 10 months ago # -
-
I like what DaveyJ & DaveO are touching on. Sensor size, better put, pixel micron size and the density of pixels.
A bit more on my rational of DX and FX becoming indistinguishable: Film is just dependent on resolution of the negative (i.e. at the same image print size, 8x10 has more available resolution than a 35mm negative).
Unlike film, sensors ability to capture an image is dependent on:
1) Number of pixels (resolution),
2) Their ability to gather light, and
3) The camera's processing engine.Take a well know example of the D300 vs. D3. Same number of pixels, same processing back bone, but the Sensor pixels ability to gather light (efficiency) is different. Pixel "well" size with, micro lenses, and a whole plethora of technologies. The size of the pixel pitch is the difference - larger is better.
Now take the D7000 vs. D800. Essentially the same pixel pitch but now has an upgraded sensor design with new processing engine. The D800 although is scoring better at high ISOs and many other scores. So what is different? There are certainly new sensor design techniques, but not some major ground breaking tech event. So now it is image processing that is updated.
We can't forget image sensors capture RGB per pixel and the image data is then combined into the final image. That is all processing by the camera. I believe IQ is and will become more dependent on the processing algorithmic functions and less on the sensor as it does now and in the past. At this point though as the D4 shows, Sensor design is still a big part of the equation.
Posted 10 months ago # -
-
DaveyJ said:
I fail to see why 400mm and 500 mm lens with Nikkor labels are not more available in 5.6 or even narrower apertures.I'll kick this can down the street as well. Added to what sevencrossing said, you also have to consider the nyquest frequency, where the shutter speed is faster than almost any shaking. Slower F-stops means more opportunity for outside factors that cause shaking to enter into the equation.
I think sevencrossing probably hit the main reason though. Good thing on the horizon is that the D4 can focus in more dim light than before. If that filters down the line, 5.6+ glass may become more feasible.
Posted 10 months ago # -
TaoTeJared said:
We can't forget image sensors capture RGB per pixel and the image data is then combined into the final image. That is all processing by the camera. I believe IQ is and will become more dependent on the processing algorithmic functions and less on the sensor as it does now and in the past. At this point though as the D4 shows, Sensor design is still a big part of the equation.Dxo is measuring the raw sensor data and we raw shooters out there we too only work with that. The d800 scores significantly higher than the d7000, because the downsizing equalizes noise.
I too think DX could become equal with FX, even depth of field wise, but it never will, nikon decided not to go that direction. Think about the lenses for a moment. Nikon's best primes on fx are all f1.4s, theoretically it's possible to make dx lenses that are of similar size, but because of the smaller image circle are faster by a stop, so if Nikon would have started making F1.0 primes and F2 zooms instead of releasing the d3 we would have had a similar experience in almost all aspects, but the size of the viewfinder, which was strangely fine for many ppl in the proDX era. (just take a look at the chinese SLRmagic noktor lenses for other Aps-c platforms) One drawback I could see, that it's very hard to design lenses that have good bokeh on an Aps-c sensor, I'm speaking of quality not quantity here. Nikon really makes sharp instruments, but even the best modern 1.4 lenses, like the 24, or 85 are suffering from bokeh nervousness, green outlining and just generally they are not as creamy as zeiss optics, on an FX camera they are great however and I imagine similar quality for dx would be economically less optimal for Nikon and Canon.
I'm using a d7000 and I very often have to try to fix bokeh nervousness in post. (like selectively reducing clarity and sharpness in LR4 with localA tools)Posted 10 months ago # -
Because it costs more? 3x cost = better performance :P
Sorry I have no constructive comments to add so I thought I would add the obvious.
Posted 10 months ago # -
Gabbb said:
Dxo is measuring the raw sensor data and we raw shooters out there we too only work with that. The d800 scores significantly higher than the d7000, because the downsizing equalizes noise.If that is implying they are reading RGB, that isn't true. NEF (RAW) files are just an processed image file that stores more information. Some have implied lately that it is true sensor data, which is not the same thing, nor does anyone but Nikon can read the 1s & 0s of that. The files are processed by the camera, and all of the tweaks are stored in the file. It's no different than having a PS file stored with layers.
I agree completely with the downsizing. DXO Mark: "Sensor Overall Score is normalized for a defined printing scenario—8Mpix printed on 8"x12" (20cmx30cm) at 300dpi resolution. Any other normalization, even at a higher resolution, would lead to the same ranking, given that any camera that could not deliver the chosen resolution would be eliminated from the comparison."
Gabbb said:
I too think DX could become equal with FX, even depth of field wise...I have been watching this too as I chase bokeh quite a bit and it is one of my sticking points with lenses. One observation is that going much below 1.4 on DX, the focal plane becomes so tight (narrow) that it almost becomes unusable other than shooting low light distance shooting. At 6' with a (75mm equiv)50mm f1.0 you have a focus area of only 2". With a 1.4 you gain an inch to just under 3". Doesn't sound like much but the average distance between two pupils is 3". Depth of the tip of a nose to the eyes is about 2". Take a photo of someone at any angle other than straight on, you can miss the zone easily. If the focus lands on the nose, the eyes will start to fall out. There is a practical use of f1.0s is limited. The other practical side I believe in is the 2-stops down for optimal sharpness. Most Canon shooters with 1.2s I meet shoot those at f/2. They say that is when it is as sharp as it can really get and they get the bokeh.
I like Zeiss bokeh as well from what I have seen. My eyes are not up to the task of MF at those narrow apertures. I did just picked up the 50mm F/2 for my Bessa R2. I did want the 1.5 but I got a sneaky low bid on one - $400 off of new on a few month old lens. We will see what comes of it.
Bokeh colors: Nikon has sways from a blueish with slight purple to green, Zeiss seems to have magenta and greens, Sigma is really purple but some newer designs are green. All deal with the types of coatings, glass used and are are the remnants of CAs and are always there. Shaky Bokeh annoys me as well. In my experience, the more rounded the aperture opening is, the better bokeh - much more so than the f-stop. Rounded blades have helped this but from my experience, the more blades there are, the better bokeh result. If you look at a Leica 50mm f/1.4 lens - they have 12 blades.
If you follow Rangefinders, Leica & Voiglander, with their f/0.95-1.2 lenses and see what they can do, envy does enter in. I followed a guy who bought a Noct f/1.0 on flickr with is M8 and the story changed from "how cool is this" to "how many times can I miss a shot." It was quite funny. I think he got only about 1 out of 100 that was close to focus. When he got his M9 it improved some.
Posted 10 months ago # -
TaoTeJared said:
nor does anyone but Nikon can read the 1s & 0s of that. The files are processed by the camera, and all of the tweaks are stored in the file. It's no different than having a PS file stored with layers.That's true, but to my knowledge the reason behind this is not actual processing, but to fix predetected dead pixels and such, it's not doing noise reduction and image processing like that, that would make no sense to me, a PC could always do a better job at that..
TaoTeJared said:
I have been watching this too as I chase bokeh quite a bit and it is one of my sticking points with lenses. One observation is that going much below 1.4 on DX, the focal plane becomes so tight (narrow) that it almost becomes unusable other than shooting low light distance shooting. At 6' with a (75mm equiv)50mm f1.0 you have a focus area of only 2". With a 1.4 you gain an inch to just under 3". Doesn't sound like much but the average distance between two pupils is 3". Depth of the tip of a nose to the eyes is about 2". Take a photo of someone at any angle other than straight on, you can miss the zone easily. If the focus lands on the nose, the eyes will start to fall out. There is a practical use of f1.0s is limited.Well, use contrast detection and focus will be nailed every time with ANY lens, I do that with my 1.8s when I don't want travel with the hit and miss train (or LV manual).
The usability is the same as using an 1.4 on FX, for example if you compare a 35mm 1.0 on dx with an 50mm 1.4 on fx you will get exactly the same deph of field whenever you fill the frame with your subject equally, obtaining perfect focus requires the same care, obviously FX cameras got better focusing systems, but that's not relevant in this comparison.TaoTeJared said:
In my experience, the more rounded the aperture opening is, the better bokeh - much more so than the f-stop. Rounded blades have helped this but from my experience, the more blades there are, the better bokeh result. If you look at a Leica 50mm f/1.4 lens - they have 12 blades.I'm not exactly certain in this area, but I've seen fantastic bokeh done by cinema lenses having as little as 5 blades. And my personal experience with the 50mm 1.8g (7blades) and 50mm 1.4g(9 blades) suggests that it may only affect the shape of the circles, in this case the 1.8g has a much better bokeh.
TaoTeJared said:
If you follow Rangefinders, Leica & Voiglander, with their f/0.95-1.2 lenses and see what they can do, envy does enter in. I followed a guy who bought a Noct f/1.0 on flickr with is M8 and the story changed from "how cool is this" to "how many times can I miss a shot." It was quite funny. I think he got only about 1 out of 100 that was close to focus. When he got his M9 it improved some.
Well it's a typical problem really, they don't have live view for focusing, but only the split screen rangefinder thingy, whatever it's called. I heard that needs to be calibrated regularly and I imagine using that it's a pain to nail focus generally with a full frame 0,95 lens.
Posted 10 months ago # -
Gabbb said:
That's true, but to my knowledge the reason behind this is not actual processing, but to fix predetected dead pixels and such, it's not doing noise reduction and image processing like that, that would make no sense to me, a PC could always do a better job at that..It's all processing. It take the reading of the r/g/b pixel and the one's next to it and combines them all into a pixel value that is 14bit color. Over simplified, 1 image pixel we see = a calculation of all the read r/g/b values in the area and is an educated guess what that color should be, like brown or the shade of yellow. It is actually kicking out an image. Only Foveon sensors are 1:1 data. Computers would be faster but not any better as the algorithm is essentially the same. That is oversimplified but essentially what the camera is doing.
The raw file just includes more information (larger color space, body set contrast, hue, sharpness settings etc., and doesn't clip off data like what happens with jpeg. It does actually apply many user settings in the raw file.
Gabbb said:
Well, use contrast detection and focus will be nailed every time with ANY lens, I do that with my 1.8s when I don't want travel with the hit and miss train (or LV manual).Kind of missed the point. To get the sharpest results you should be in the middle of the focus area. At a 45deg angle to a subject at F/1 with a 50mm lens, you can not get both eyes in focus on an adult. The distance between the eyes is larger than the DOF. At f/1.8 = 3.5", f/1.4 =3", F/1 = 2", distance of the edges of the eyes is 3 inches and at a 45deg, you are looking somewhere near 2.5" DOF needed at 6ft. To give some leeway (natural body sway, etc.), you should shoot at F/2 which gives you 4" of DOF.
Of course there are artistic merits that break from that as well as shooting studio work with a tripod and control of environmental nuances.
Bokeh is always subjective - each of us have our tolerances on various items on it and rarely have I ever seen two people completely agree - even describing the same lens - which always cracks me up. Personally, I actually like the 50mm 1.4g bokeh better but the new 1.8g is very good as well. If that was out a few years ago, that would be in my bag rather than the 1.4 - mainly due to price and not being that much different.
Rangefinder focus did need calibrated in the olden days, as well as parallax, but that has been eliminated 20 years ago and continues to get better. It is a different way of shooting for sure but really not much different than shooting with any split screen focus MF camera.
Posted 10 months ago # -
TaoTeJared said:
It's all processing. It take the reading of the r/g/b pixel and the one's next to it and combines them all into a pixel value that is 14bit color. Over simplified, 1 image pixel we see = a calculation of all the read r/g/b values in the area and is an educated guess what that color should be, like brown or the shade of yellow. It is actually kicking out an image. Only Foveon sensors are 1:1 data. Computers would be faster but not any better as the algorithm is essentially the same. That is oversimplified but essentially what the camera is doing.I believe demosaicing is done by your raw converter. It's not just the speed, there are smart ways do the demosaicing, computing in the actual image data, and a computer has processing power for that and also expensive adobe and non adobe algorithms, Nikon is not a software company, just compare in camera jpegs vs. basic raw converted ones.
TaoTeJared said:
Kind of missed the point. To get the sharpest results you should be in the middle of the focus area. At a 45deg angle to a subject at F/1 with a 50mm lens, you can not get both eyes in focus on an adult. The distance between the eyes is larger than the DOF. At f/1.8 = 3.5", f/1.4 =3", F/1 = 2", distance of the edges of the eyes is 3 inches and at a 45deg, you are looking somewhere near 2.5" DOF needed at 6ft. To give some leeway (natural body sway, etc.), you should shoot at F/2 which gives you 4" of DOF.
Of course there are artistic merits that break from that as well as shooting studio work with a tripod and control of environmental nuances.
Bokeh is always subjective - each of us have our tolerances on various items on it and rarely have I ever seen two people completely agree - even describing the same lens - which always cracks me up. Personally, I actually like the 50mm 1.4g bokeh better but the new 1.8g is very good as well. If that was out a few years ago, that would be in my bag rather than the 1.4 - mainly due to price and not being that much different.
Rangefinder focus did need calibrated in the olden days, as well as parallax, but that has been eliminated 20 years ago and continues to get better. It is a different way of shooting for sure but really not much different than shooting with any split screen focus MF camera.
I get it, but the same is true for f1.4 on fx. With a Dx f1 you would need to move back a few steps to get similar field of view so the dof would be the same.
About the nikon 50 1.4-1.8g here is comparison shot i've made when I had them:
Left one is the 1.8g..
(100% crops)
Posted 10 months ago # -
@Gabbb I may not be with the younger folks here, but I find comparisons easier if a full frame is posted full size to Flickr in order that we can cruise all over the image area and compare. One can post the links to the full size and then we can open a couple windows and get a real nice comparison.
For example, my photo... http://www.flickr.com/photos/fantinesfotos/7014984231/sizes/o/in/photostream/ allows one to examine the 24-120mm f/4.0 all the way to the corners.
Just my suggestion and of course the image is full frame FX so my opinion is, well, the resolution is "good enough" for me and I think the D800 would be for those going to absolutely huge images.... Maybe the comparison between D800/E and D4 would be a nice thread.
Posted 10 months ago # -
msmoto said:
@Gabbb I may not be with the younger folks here, but I find comparisons easier if a full frame is posted full size to Flickr in order that we can cruise all over the image area and compare. One can post the links to the full size and then we can open a couple windows and get a real nice comparison.For example, my photo... http://www.flickr.com/photos/fantinesfotos/7014984231/sizes/o/in/photostream/ allows one to examine the 24-120mm f/4.0 all the way to the corners.
Just my suggestion and of course the image is full frame FX so my opinion is, well, the resolution is "good enough" for me and I think the D800 would be for those going to absolutely huge images.... Maybe the comparison between D800/E and D4 would be a nice thread.
I think this image is showing perfectly the differences in green outlining. On the full image it wouldn't be that obvious and at the time of flickr-ing this I thought it would be better this way and now I won't look up the raw files unless someone would really want to see them, assuming I haven't deleted them.
Also I'm not a flickr pro member, so I cannot post full-size images.
Posted 10 months ago # -
TaoTeJared said:
I'll kick this can down the street as well. Added to what sevencrossing said, you also have to consider the nyquest frequency, where the shutter speed is faster than almost any shaking.Might be a little bit of a pedantic point, but having a longer lens does not increase the frequency of camera shake -- it just increases its amplitude. The 1/focal length rule is a guideline intended to keep camera motion blur amplitude within a certain range, not deal with the frequency of camera shake. (The Nyquist frequency comes into play when choosing a sampling rate based on the frequencies found in the function being sampled, which in this instance don't change when you change lenses.)
Also note that a higher resolution sensor will show more camera shake if one follows that rule, because it's more sensitive to small amplitudes of camera shake. This is why a lot of people find it frustrating that their handheld images seem somewhat soft at 1:1 on the D800.
Posted 10 months ago # -
DaveyJ said:
In large format lens 5.6 and even F8 might be the starting point on fine lens. I fail to see why 400mm and 500 mm lens with Nikkor labels are not more available in 5.6 or even narrower apertures.Nikon sees your 500mm f/5.6 and raises you an additional 300mm and VR :)
Latest rumor: Nikon to announce a new AF-S 800mm f/5.6 VR lens
Posted 10 months ago # -
DaveyJ said:
I fail to see why 400mm and 500 mm lens with Nikkor labels are not more available in 5.6 or even narrower apertures.well it seems Nikon does read this forum. Will an 800mm f 5.6 be OK
http://imaging.nikon.com/lineup/lens/mf/singlefocal/telephoto/800mmf_56/index.htm
Posted 10 months ago #
Reply
You must log in to post.
