Ad
Ad
Ad
Pages: « 1 [2] 3 4 ... 10 »   Bottom of Page
Print
Author Topic: The future of medium format  (Read 19427 times)
Paul2660
Sr. Member
****
Online Online

Posts: 1910


WWW
« Reply #20 on: December 12, 2012, 03:20:34 PM »
ReplyReply

My rule of thumb on the D800 Live view is go to the max and back off three.  At a 100% view, it's way over kill and hard to determine anything.  Same as the image preview.  Nikon's what you see is what you get approach is great on exposure, where as Canon has a much much better noise buffering setup.  Just try to use a D800 in low light (even with the exposure correct), the noise on screen is very hard to work around and get a good focus.  In good light, this is not a problem.  Canon somehow buffers out the noise, enough so you can see the image and focus.  

I didn't really start using Live View until I moved to Nikon, as I found it critical to focus on the D800 and e.  I have started using on Canon (for night work) and love it.  As for implementation on MF, it will be interesting to see what the future brings.

I am still confused on the chip construction as many are using term stitched including myself.  I know from using a IQ180 briefly that this "stitching" can be sometimes seen on a actual image and is really impossible to remove in post.  It has been explained to me that the raw software is supposed to blend these areas and smooth them out. (Capture One, LR etc).  My 160 shows lines at times on LCC's when viewed on the camera LCD, but these same lines never show up in post.  On my loaner 180, they did show up in post and pretty much ruined the image.  Again I have been told this is a "tuning" issue and each back has to be tuned to get the blend correct.   I should also note, that on the 180 the images that caused the problem were taken with a Schneider 43mm lens and this lens is not considered a good solution on the 180.  (I own a 160, but it's in for work).  

Paul
Logged

Paul Caldwell
Little Rock, Arkansas U.S.
Photography > http://photosofarkansas.com
Blog> http://paulcaldwellphotography.com
torger
Sr. Member
****
Offline Offline

Posts: 1588


« Reply #21 on: December 12, 2012, 03:29:49 PM »
ReplyReply

The way landscape photographer likes to use live view is 1) framing (full view), 2) focus placement (100% zoom). If the live view is really good to simulate the final result one can get histograms and use it to set exposure too.

A good 100% focus check live view must demosaic the sensor data, preferably apply some sharpening etc, and if low light a good noise reduction, so it is not trivial. It must also update with a fast refresh rate and without delay, or else it becomes difficult to finetune manual focus. It should also preferably handle a wide range of light conditions.

As far as I know Canon is the leader on live view, much better than Nikon for example. Don't know about Sony. Maybe Canon's less-than-great low ISO performance is a price one have to pay for the excellent live view, at least so far. With the high ISO performance of the latest cameras can through live view in low light siutation actually make you see more than you do through the viewfinder.

Live view would be very useful for a tech camera, but not so meaningful for the typical 645 DSLR use case when you hand-hold the camera and look through the viewfinder.

Actually I don't think CMOS-quality live view would change that much in practical MFDB photography (although I would not cry if I did not need to use the slidning back), we do quite good without, but I think it would help sales and help market communication. Photographers of today do not start with film, they start with live-view capable DSLRs. It's a lot easier to sell something when you don't need to tell your customer that they need to drop the flexibility they are used to.
Logged
torger
Sr. Member
****
Offline Offline

Posts: 1588


« Reply #22 on: December 12, 2012, 03:43:14 PM »
ReplyReply

BJL, you seem to know a lot about sensors. As a tech cam user I worry about color cast, i e that the sensors cannot handle low angles of incoming light, and thus forces more or less strong retrofocus lens designs. I think it would be sad if the ability to handle non-retrofocus or weak-retrofocus lenses disappeared.

Do you know if there is something inherent in CMOS design that would make the color casts worse or not than we see in current CCDs?
Logged
BernardLanguillier
Sr. Member
****
Online Online

Posts: 8202



WWW
« Reply #23 on: December 12, 2012, 05:51:47 PM »
ReplyReply

Kind of...at the moment it is more like "what you see is what the camera interpolates out of the data read off the sensor"...hence the difference in LV image quality at 1:1 between cameras. Compare the 5DIII to the D800 to see the differences

Yair

The live view high magnification zoom mode of the D800 may not be sexy, but it does the job in terms of enabling perfect focus 100% of the time, even in pre-dawn and post-dusk dark conditions.

I'll stick to my view that usable live view is mandatory for MF backs to become credible contendors for some of my applications.

Cheers,
Bernard
« Last Edit: December 12, 2012, 05:54:54 PM by BernardLanguillier » Logged

A few images online here!
Doug Peterson
Sr. Member
****
Offline Offline

Posts: 2838


WWW
« Reply #24 on: December 12, 2012, 06:34:00 PM »
ReplyReply

3) The crop is fairly modest and gives "look around" capability on the viewfinder, which I've grown to like

Good point. Look around is a nice thing to have. Something to be kept in mind in future MF development.

D800 offers various crop abilities and gives you two options for indicating the crop in the finder. Either greyed out of a guide frame. Can be very handy.
The crop feature also increases frame rate.

Could be a good option for medium format too. 80MP at a slow frame rate and 15 or 20 much faster.


Sensor+ in the P40+, P65+, IQ140, IQ160, and IQ180 already offer faster frame rates (and two stops of free ISO) by reducing resolution but keeping the same frame size.

Aptus ii 10 and Aptus ii 12 already offer the option to shoot square raw files at faster frame rates.
Logged

DOUG PETERSON (dep@digitaltransitions.com), Digital Transitions
Dealer for Phase One, Mamiya Leaf, Arca-Swiss, Cambo, Profoto
Office: 877.367.8537
Cell: 740.707.2183
Phase One IQ250 FAQ
ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 7648


WWW
« Reply #25 on: December 12, 2012, 08:44:28 PM »
ReplyReply

Hi,

It normally means that a single chip is done using several exposures in the stepper. Normal steppers cannot expose 24x36 mm in a single exposure.

This paper from Canon discusses the issue http://www.robgalbraith.com/public_files/Canon_Full-Frame_CMOS_White_Paper.pdf

"Of course, there is more to this topic. For example, the circuit pattern of a fullframe
sensor is too large to be projected on the silicon wafer all at once; it requires
three separate exposures (See page 53). This means that the number of masks and
exposure processes is tripled. For now, appreciate that a full-frame sensor costs not
three or four times, but ten, twenty or more times as much as an APS-C sensor."

Best regards
Erik


I am still confused on the chip construction as many are using term stitched including myself.

Paul
Logged

FredBGG
Sr. Member
****
Offline Offline

Posts: 1651


« Reply #26 on: December 12, 2012, 08:47:09 PM »
ReplyReply

Sensor+ in the P40+, P65+, IQ140, IQ160, and IQ180 already offer faster frame rates (and two stops of free ISO) by reducing resolution but keeping the same frame size.

Aptus ii 10 and Aptus ii 12 already offer the option to shoot square raw files at faster frame rates.

The speeds in sensor plus are still quite slow. 1.2fps (1.8 fps in Sensor+ ) for the IQ140 and 0.7 fps (0.9 fps in Sensor+ ). However sensor plus is a
useful function producing a lower res image, but still using the full frame and still capturing the look of the whole projected image of the lens.
Usefull when 80MP files are overkill and 20MP is better but still with the advantage of oversampling.

I was suggesting significant crop to obtain even higher speed when needed. This may be a way for future MF cameras to reach 4 to 6 fps.

"(and two stops of free ISO)" That was tongue in-cheek right..... Wink
« Last Edit: December 12, 2012, 08:50:42 PM by FredBGG » Logged
FredBGG
Sr. Member
****
Offline Offline

Posts: 1651


« Reply #27 on: December 12, 2012, 09:07:12 PM »
ReplyReply

My rule of thumb on the D800 Live view is go to the max and back off three.  At a 100% view, it's way over kill and hard to determine anything.
Paul

Excuse the diversion...

Paul you can setup the center button of the multi function button to magnify to your preferred magnification. Very handy one click to go to
the magnification you want.
Logged
Doug Peterson
Sr. Member
****
Offline Offline

Posts: 2838


WWW
« Reply #28 on: December 12, 2012, 09:21:31 PM »
ReplyReply

The speeds in sensor plus are still quite slow. 1.2fps (1.8 fps in Sensor+ ) for the IQ140 and 0.7 fps (0.9 fps in Sensor+ ). However sensor plus is a
useful function producing a lower res image, but still using the full frame and still capturing the look of the whole projected image of the lens.
Usefull when 80MP files are overkill and 20MP is better but still with the advantage of oversampling.

I was suggesting significant crop to obtain even higher speed when needed. This may be a way for future MF cameras to reach 4 to 6 fps.

"(and two stops of free ISO)" That was tongue in-cheek right..... Wink

The noise at ISO 1600 in sensor plus is identical to the noise in an ISO400 full frame image. That is what I meant by "two free stops of iso". Going into sensor plus mode increases your native ISO by two stops, at the compromise of file size.

If I made a list of features/improvements I think that would help photographers (who use or would consider using medium format) produce better work, and produce it more easily and with more enjoyment of the process... 4-6fps wouldn't even be in the top 20.

Maybe that is just me. I don't claim to know the needs/wants of all medium format or potential medium format shooters.
Logged

DOUG PETERSON (dep@digitaltransitions.com), Digital Transitions
Dealer for Phase One, Mamiya Leaf, Arca-Swiss, Cambo, Profoto
Office: 877.367.8537
Cell: 740.707.2183
Phase One IQ250 FAQ
gerald.d
Sr. Member
****
Offline Offline

Posts: 388


« Reply #29 on: December 12, 2012, 10:15:14 PM »
ReplyReply

Live view would be very useful for a tech camera, but not so meaningful for the typical 645 DSLR use case when you hand-hold the camera and look through the viewfinder.

Imagine a combination of a really good EVF, along with further enhancement of the eye control focus-select system used in the Canon EOS-5.

Look in the viewfinder at the focus point you want to use whilst holding down one button, camera focuses using that focus point. Look at another point in the viewfinder whilst spinning a control wheel, zoom in and out to focus check any part of the image. Make it a clickable control wheel (like that on the Alpa FPS), and you can do it all with a single control.

Kind regards,

Gerald.
Logged
FredBGG
Sr. Member
****
Offline Offline

Posts: 1651


« Reply #30 on: December 12, 2012, 10:49:12 PM »
ReplyReply

The noise at ISO 1600 in sensor plus is identical to the noise in an ISO400 full frame image. That is what I meant by "two free stops of iso". Going into sensor plus mode increases your native ISO by two stops, at the compromise of file size.

What happens when the two files are printed with the subject of the photo printed to the same size?
Logged
ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 7648


WWW
« Reply #31 on: December 12, 2012, 10:57:42 PM »
ReplyReply

Hi,

You can check that on DxO-mark. Sensor+ makes a small improvement on DR.

Reason: Signal is summed over four pixels while readout noise is kept the same.

For tonal range (which by and large ignores read noise) the effect is very, very small.

Both samples I enclosed are in print mode, that is normalized to a smallish print.

Best regards
Erik




What happens when the two files are printed with the subject of the photo printed to the same size?
Logged

FredBGG
Sr. Member
****
Offline Offline

Posts: 1651


« Reply #32 on: December 12, 2012, 11:09:57 PM »
ReplyReply

Yes but look at the noise graphs for screen and print.




Logged
ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 7648


WWW
« Reply #33 on: December 12, 2012, 11:54:17 PM »
ReplyReply

Hi,

SNR 18% is dominated by shot noise and depends only on the number of photons captured. It scales perfectly with enlargement that is the reason it does little effect on noise. If you look at noise at 1% you would see the effect of Sensor+.

But I agree that most of the advantages by Sensor+ can be achieved by simply scaling down the picture to half linear dimension.

Best regards
Erik


Yes but look at the noise graphs for screen and print.





Logged

FredBGG
Sr. Member
****
Offline Offline

Posts: 1651


« Reply #34 on: December 13, 2012, 02:17:21 AM »
ReplyReply

What happens when the two files are printed with the subject of the photo printed to the same size?

Hi,

SNR 18% is dominated by shot noise and depends only on the number of photons captured. It scales perfectly with enlargement that is the reason it does little effect on noise. If you look at noise at 1% you would see the effect of Sensor+.

But I agree that most of the advantages by Sensor+ can be achieved by simply scaling down the picture to half linear dimension.

Best regards
Erik


I was not stating that, asking the question really.
« Last Edit: December 13, 2012, 02:24:47 AM by FredBGG » Logged
MrSmith
Sr. Member
****
Offline Offline

Posts: 908



WWW
« Reply #35 on: December 13, 2012, 06:44:52 AM »
ReplyReply

MFD has to change, being hamstrung by only having the hand-me-downs from chip manufacturers is not helping all the time 35mm makes advances in optics and sensors, the 'improvements' touted are abysmal when you look at the cost and these 'upgrades' are glacial in the speed they come to market, just look how long it took to get a passable rear screen and USB3 that doesn't work. 
Real live view that is usable like a canikon and 800-3200 iso that is perfectly usable like a current DSLR without silly crop formats.
If the next batch of 'new and improved' backs only offer even higher mpixels for a closer look at diffraction and the reps extolling the virtues of a dedicated button to add a star rating then the plot really has been lost.
I actually don't care if they cost 15-20k it just has to deliver the goods, MFD to me seems like a cartel that all the time they can get away with it they will keep pushing the same product.

I guess it's going to be a sensor manufacturer that is going to dictate and change for the better.
All IMHO.

While writing this I got a marketing email from hassleblad, they have an owners club now  Cheesy
Logged
Doug Peterson
Sr. Member
****
Offline Offline

Posts: 2838


WWW
« Reply #36 on: December 13, 2012, 08:58:20 AM »
ReplyReply

Down sampling in post in theory produces one stop gain in per pixel noise.
Pixel binning produces two stops.

In practice I find pixel binning to be a bit better than one stop difference; why? I could only speculate (maybe related to the modest ineffeciencies of demosaicing?)

But what do I know? I'm not looking at dXo charts, I'm only looking at raw files from said cameras in their native processor.
Logged

DOUG PETERSON (dep@digitaltransitions.com), Digital Transitions
Dealer for Phase One, Mamiya Leaf, Arca-Swiss, Cambo, Profoto
Office: 877.367.8537
Cell: 740.707.2183
Phase One IQ250 FAQ
theguywitha645d
Sr. Member
****
Offline Offline

Posts: 970


« Reply #37 on: December 13, 2012, 09:26:22 AM »
ReplyReply

Downsampling has very little impact on noise. It might be worth pointing out, we are comparing two pixels with a difference in linear resolution of 2x (or area of 4x). The visual system will really not perceive any difference between the conditions is it naturally does its own downsampling cause by its own limit in resolution. Comparing images at 100% monitor view is not going to show the reality of the situation, unless you put instructions under your pictures for viewers to fix two different viewing distances depending if you downsampled or not.

Binning on the other hand is making a difference as it impacts the actual signal.
Logged
ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 7648


WWW
« Reply #38 on: December 13, 2012, 11:18:38 AM »
ReplyReply

Hi,

I disagree in part.

To begin with, I presume that we really talk about prints. Whenever you print the image will be resampled to the native resolution of the printer driver. The resolution of the file may matter little, so far signal is dominated by shot noise only the total number of photons captured will matter and that number is not affected by binning.

Shadow noise can be dominated by readout noise. Lets do some very simple math:

Full well capacity is 65000
Nominal ISO is 25

So 1600 ISO = ln(1600 / 25) / ln 2 means six stops of underexposure. Let's also assume that we look at 12.5% grey tone, 12.5% is 1/8 or three stops.

So our grey tone will hold about  65000 * 25 / 1600 / 8 = 127 photons (well electrons to be correct). This will result in shot noise of 11 electron charges.

No readout noise on the DALSA chip used on the IQ180 is about 12 electrons (according to data sheet), noise is added in quadrature, so total noise would be:

sqrt(11*11 + 12 *12) = 16.3, leading to an Signal Noise Ratio of 127 / 16.3 = 7.8

Now, if we subsample four pixels the math will change a bit:

FWC = 4 * 65000
Electron count = 507

Giving shot noise of 22.5

Readout noise will become sqrt (4 * 12 *12) = 24

Total noise will be sqrt(22.5 * 22.5 + 24 *24) = 32.8 giving SNR of 15.4 (about twice the SNR)

If we use hardware binning everything will be the same, except that readout noise will not be added up but stay at 12.

So we get: total noise = sqr(22.5 * 22.5 + 12 * 12) = 25.5 leading to SNR of 507 / 25.5 = 19.8

So downsampling will improve SNR from 7.8 to 15.4. Binning in hardware will improve SNR to 19.8.

;-) This is the math ;-)

Keep in mind that downsampling will not improve image quality in print.

I'd also suggest that raw converters can take different parameters into account and do more ore less aggressive noise reduction. For instance, it would make a lot of sense to increase noise reduction in the darks, which often have impulse noise and probably little detail.

Best regards
Erik

Down sampling in post in theory produces one stop gain in per pixel noise.
Pixel binning produces two stops.

In practice I find pixel binning to be a bit better than one stop difference; why? I could only speculate (maybe related to the modest ineffeciencies of demosaicing?)

But what do I know? I'm not looking at dXo charts, I'm only looking at raw files from said cameras in their native processor.

.....

Downsampling has very little impact on noise. It might be worth pointing out, we are comparing two pixels with a difference in linear resolution of 2x (or area of 4x). The visual system will really not perceive any difference between the conditions is it naturally does its own downsampling cause by its own limit in resolution. Comparing images at 100% monitor view is not going to show the reality of the situation, unless you put instructions under your pictures for viewers to fix two different viewing distances depending if you downsampled or not.

Binning on the other hand is making a difference as it impacts the actual signal.
Logged

Wayne Fox
Sr. Member
****
Offline Offline

Posts: 2884



WWW
« Reply #39 on: December 13, 2012, 01:11:22 PM »
ReplyReply

Interesting thread, I quit following the other one some time ago.  That horse was beat to death when the d800 came out, and as one who owns both a d800e and a IQ180 system, for me MF is not dead.

As far as the future, the only feature I care about in MF is LiveView, as has been discussed.  I would be back on a technical camera in a flash, if MF had liveview that was as functional as my Canon or Nikons.  In fact, give me , my NEX 7's manual focus and I'd probably sell the DF and lens and move strictly to tech cameras.  I tried a nice tech camera system, and just couldn't stand the challenge of composing and focusing - just doesn't fit my "explorer" type shooting style.
« Last Edit: December 13, 2012, 01:18:50 PM by Wayne Fox » Logged

Pages: « 1 [2] 3 4 ... 10 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad