Ad
Ad
Ad
Pages: « 1 2 [3] 4 »   Bottom of Page
Print
Author Topic: MF Digital newbie needs a little help please  (Read 10025 times)
ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 7888


WWW
« Reply #40 on: April 07, 2013, 02:10:45 AM »
ReplyReply

Hi,

I would suggest that you try to arrange so you can test the camera you intend to buy, or try to get more representative samples.

As I said before DR and histogram says little about the quality of noise. There is in all probability also a sample variation between backs. Different vendors may go to different depths to calibrate their sensors.

Marc McCalmont, a very helpful person on these forums, has found that his Pentax K5 had better DN than his Phase One P45+. With the IQ180 he had similar DR to his Pentax K5. Marc has posted raw images that illustrated this.

If you need best DR in DSLRs, Nikon and Sony seems to be best in that are, and Canon has done little progress on DR. In APS-C, Pentax seems to making the best of the Sony sensors they use.

Best regards
Erik


Thanks for taking the time for your detailed responses Erik; Though... i am well and truely confused now. A Dm33  recording only 12bits and the Aptus II 8 and 12 recording 12-14bits. The canon i belive records the same... So the MF cameras are recording less or equal to a 5D2...

Not sure if i am mis interpreting this and if going to MFD something like the DM33 will really be of any benefit or would i be going backwards.

Again, thanks for the information and replies, im just not sure what to make of this info and what direction to go in.

Regards

Andrew

« Last Edit: April 07, 2013, 02:38:59 AM by ErikKaffehr » Logged

ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 7888


WWW
« Reply #41 on: April 07, 2013, 02:14:33 AM »
ReplyReply

Here is Marc's posting:

http://www.luminous-landscape.com/forum/index.php?topic=50895.0

Best regards
Erik
Logged

ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 7888


WWW
« Reply #42 on: April 07, 2013, 02:21:36 AM »
ReplyReply

Hi,

Now that I have some more insight in your needs I would point a bit at some tests by Miles Hecker.

http://wyofoto.com/Pentax_645D/Pentax_645D_review_pt1.html

http://wyofoto.com/Pentax_645D/Pentax_645D_review_pt2.html

And also this: http://www.luminous-landscape.com/forum/index.php?topic=50977.0

Just to confuse a bit more, I have downloaded a pair of comparison images between Canon 1DII and Phase One P65+, and the P65+ had much better noise and shadow detail. That comparison was shot by Peter Eastway and I don't think it was biased. I don't think the copyright allows me to share the images, unfortunately.

http://echophoto.dnsalias.net/ekr/index.php/photoarticles/41-phase-one-images-for-download

I also added some plots from DxO mark.

Best regards
Erik
« Last Edit: April 07, 2013, 02:38:13 AM by ErikKaffehr » Logged

jerome_m
Sr. Member
****
Offline Offline

Posts: 656


« Reply #43 on: April 07, 2013, 06:23:26 AM »
ReplyReply

Rawdigger is an interesting piece of software, but I tried it on largely identical images taken with the H3D-31 and the D800 (the church pictures I posted in another thread) and I find the results puzzling. I don't think that the results are really relevant for real photographic uses.
Logged
jerome_m
Sr. Member
****
Offline Offline

Posts: 656


« Reply #44 on: April 07, 2013, 06:46:29 AM »
ReplyReply

Erik, sorry, i should be a little clearer on what i mean on DR (most probably not the best term to use). Really if i could get the sameor slightly greater DR that i get from my 5D2, that would be great, though i am more concerned in how it renders tones. With the 5D2 when we are getting close to pure white its like it hits a wall and looks very abrupt. I am chasing subtle tonal changes from light to dark, not a hard banging, clipping looking tonal transition. I hope this makes sense, i do apologise for the my lack of technical description.

You will get better DR than the one from the 5D2 with about any other camera. The 5D2 is notoriously famous for having more noise than the competition in its shadow. Due to the structure of this noise ("banding"...), it seems to be read noise. You won't get that particular noise from a CCD (all MFs use CCDs as opposed to CMOS) and you won't get it from Sony full frame sensors, which duplicated the read-out lines to measure and remove that particular noise from the raw data. The first sensor to use that system was the one in the A900, BTW.

What this means in photographic practice is that it is possible to "raise the shadows" of MF cameras and 24x36 cameras with a Sony sensor without having structured noise appearing. In turn, that allows one to underexpose a little bit so as to have smoother highlights. In my experience with the A900 at base iso some years ago, the dynamic range of that camera was already sufficient for most photographic uses. In other words, if I tried to use the complete dynamic range of that camera (by raising shadows and recovering highlights), it would already be too much and look like an HDR picture with typical HDR artefacts like halos around high contrast transitions.

What does this mean for us as photographers? It means that the real difference is not in the DR of the sensor, but in the way the data is processed. The real secret lies in processing the DR of the camera to fit in the limited DR of our screens or prints while preserving smooth transitions and good details. This is, traditionally, a strong point of MF manufacturers, not that Nikon/Canon/Sony can't do it, but because these brands compete on other metrics (e.g. high iso).
Logged
alosurdo
Newbie
*
Offline Offline

Posts: 40


« Reply #45 on: April 07, 2013, 06:47:40 AM »
ReplyReply

Thanks guys for replying.

Jerome, im also not sure about the Rawdigger software either, though the 2 sample images that leaf have of the DM33 on the website (see previous post) are pretty crummy. ive tried raw processing these in both Capture one and ACR and expected Capture one to do a much better job, but was suprised when ACR won hands down.

RE: the Pentax 645. I am not 100% sold on this system, yesit it does have better resolution then the sytems i am looking at, though as i understand, it is not really upgradeable when the time comes.

Regards

Andrew
Logged
FredBGG
Sr. Member
****
Offline Offline

Posts: 1651


« Reply #46 on: April 07, 2013, 04:00:20 PM »
ReplyReply

What does this mean for us as photographers? It means that the real difference is not in the DR of the sensor, but in the way the data is processed. The real secret lies in processing the DR of the camera to fit in the limited DR of our screens or prints while preserving smooth transitions and good details. This is, traditionally, a strong point of MF manufacturers, not that Nikon/Canon/Sony can't do it, but because these brands compete on other metrics (e.g. high iso).

Shadow recovery with the Sony 35mm sensor.

Here is the original file jpeg right out of camera


click on the link below to see the recovery that demonstrates the dynamic range advantages in the shadows.

http://farm9.staticflickr.com/8106/8628371435_23ef91c167_o.gif
pixel peeping at 1:2 Underexposed and same file with recovered

Edit: I made the above animated gif into a link rather than displaying the image directly as the animation
is a bit distracting while reading.


And this is processing the jpeg in photoshop and not even using a raw converter... so this is about as bad as it gets.
« Last Edit: April 08, 2013, 11:38:29 AM by FredBGG » Logged
ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 7888


WWW
« Reply #47 on: April 07, 2013, 10:30:20 PM »
ReplyReply

Hi,

The value of Rawdigger is that it shows the actual data in the raw file without processing. So it gives good indication about clipping (as it doesn't attempt highlight recovery) and you can also estimate the dynamic range of the scene. It does say very little about the looks of noise.

Best regards
Erik


Rawdigger is an interesting piece of software, but I tried it on largely identical images taken with the H3D-31 and the D800 (the church pictures I posted in another thread) and I find the results puzzling. I don't think that the results are really relevant for real photographic uses.
Logged

ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 7888


WWW
« Reply #48 on: April 07, 2013, 10:39:37 PM »
ReplyReply

Hi,

What was essentially new with the Exmoor sensor is that it has massively parallell Analog Digital Converters, there is one for each column, so there are over 6000 of them, working in parallell. So the signal coming out from the chip is actually digital. Canon and some Nikons still have off chip converters. Nikon D4 has off chip converters and D800/D800E have the Exmoor chip.

Correlated double sampling, measuring voltage after chip reset and subtracting from voltage after exposure is used by all CMOS sensors, but it is not possible on CCD as readout on a CCD is destructive.

I think the same design was used on the Alpha 700.

Best regards
Erik


You will get better DR than the one from the 5D2 with about any other camera. The 5D2 is notoriously famous for having more noise than the competition in its shadow. Due to the structure of this noise ("banding"...), it seems to be read noise. You won't get that particular noise from a CCD (all MFs use CCDs as opposed to CMOS) and you won't get it from Sony full frame sensors, which duplicated the read-out lines to measure and remove that particular noise from the raw data. The first sensor to use that system was the one in the A900, BTW.


Logged

jerome_m
Sr. Member
****
Offline Offline

Posts: 656


« Reply #49 on: April 08, 2013, 01:00:35 AM »
ReplyReply

Shadow recovery with the Sony 35mm sensor.


pixel peeping at 1:2 Underexposed and same file with recovered

And this is processing the jpeg in photoshop and not even using a raw converter... so this is about as bad as it gets.


When I see your continuous stream of answers trying to prove the same point again and again and the time you must take to produce these flashing comparisons I wonder what is wrong with you.

Yes, one can recover underexposed shadows with a Sony sensor. I know, I have been doing so for the past 4 years.
Logged
jerome_m
Sr. Member
****
Offline Offline

Posts: 656


« Reply #50 on: April 08, 2013, 01:07:28 AM »
ReplyReply

The value of Rawdigger is that it shows the actual data in the raw file without processing. So it gives good indication about clipping (as it doesn't attempt highlight recovery) and you can also estimate the dynamic range of the scene. It does say very little about the looks of noise.

I saw that the clipping routines of rawdigger give different results on the same scene taken by two different cameras, which I find puzzling. The appearance of the histograms in some display modes is also difficult to explain from sensor data alone (but could be an effect of how raw are compressed).
Logged
ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 7888


WWW
« Reply #51 on: April 08, 2013, 01:37:20 AM »
ReplyReply

Hi,

Can you give some examples?

I know there are some bugs, but I have not seen major issues, like false results.

Best regards
Erik

I saw that the clipping routines of rawdigger give different results on the same scene taken by two different cameras, which I find puzzling. The appearance of the histograms in some display modes is also difficult to explain from sensor data alone (but could be an effect of how raw are compressed).
Logged

jerome_m
Sr. Member
****
Offline Offline

Posts: 656


« Reply #52 on: April 08, 2013, 02:31:13 AM »
ReplyReply

The correct place to discuss rawdigger issues is the associated forum from its developer, not lula. I will post my observations there.
Logged
FredBGG
Sr. Member
****
Offline Offline

Posts: 1651


« Reply #53 on: April 08, 2013, 11:32:08 AM »
ReplyReply

When I see your continuous stream of answers trying to prove the same point again and again and the time you must take to produce these flashing comparisons I wonder what is wrong with you.

Yes, one can recover underexposed shadows with a Sony sensor. I know, I have been doing so for the past 4 years.

Now you have me really puzzled.
You correctly bring up the point that how you use large dynamic range to move and fit it into the display medium, be it a monitor or a print.
You go on to say that this was traditionally a strong point of MF cameras, and I agree on that because there was a time when they had more dynamic range than
other formats. You then go on to say that that does not mean that Nikon/Canon/Sony can't do that. On this I agree to a certain point.
Canon unfortunately currently is not quite up there with the Sony/Nikon sensors... as a result I chose to move (in part) from Canon to Nikon.

What really puzzles me is why does a visual example of what you are saying upset you and make you think something is wrong with me?
You question the time I spend making the example...... It probably took me less time to take that example, load it into photoshop and export an animated gif.
After all don't they say an image is worth a thousand words?

alosurdo stated he is on the fence regarding the d800e, but is not sure. He also said that one of the main things he is looking for is dynamic range in the shadows.
I posted the example for him. If I recall correctly I have not posted an example I have taken of shadow recovery, just highlight recovery.
« Last Edit: April 08, 2013, 11:41:53 AM by FredBGG » Logged
ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 7888


WWW
« Reply #54 on: April 08, 2013, 11:46:30 AM »
ReplyReply

OK,

But that leaves us without a clue what you meant.

Best regards
Erik


The correct place to discuss rawdigger issues is the associated forum from its developer, not lula. I will post my observations there.
Logged

jerome_m
Sr. Member
****
Offline Offline

Posts: 656


« Reply #55 on: April 08, 2013, 11:54:40 AM »
ReplyReply

What I was trying to say but was not understood is that, in photographic practice (as opposed to internet tests), it is much easier for me to get smoother highlights on my H3D than my D800. The reason is not that the DR of the sensor itself is higher (it is sufficient for both cameras), but that the camera meters are calibrated differently (I already adjusted the D800 meter down) and that Phocus is optimized for that task. Nikon apparently chose to optimize their cameras for better low light performance, it is an area they compete for.

This will not appear on internet "tests", since they make sure that exposure and processing is the same.
Logged
ErikKaffehr
Sr. Member
****
Offline Offline

Posts: 7888


WWW
« Reply #56 on: April 08, 2013, 12:56:17 PM »
ReplyReply

Hi,

OK, my take is that to make the best use of any camera we need to expose correctly to the right. That is the only way to fully utilize the sensor.

Using Rawanalyzer, possible bugs/issues notwithstanding, is helpful in learning how close we get to saturation in the highlights. The way I work is that I make an exposure, check histogram and try to expose just short of blinking highlights.

Best regards
Erik

What I was trying to say but was not understood is that, in photographic practice (as opposed to internet tests), it is much easier for me to get smoother highlights on my H3D than my D800. The reason is not that the DR of the sensor itself is higher (it is sufficient for both cameras), but that the camera meters are calibrated differently (I already adjusted the D800 meter down) and that Phocus is optimized for that task. Nikon apparently chose to optimize their cameras for better low light performance, it is an area they compete for.

This will not appear on internet "tests", since they make sure that exposure and processing is the same.
Logged

jerome_m
Sr. Member
****
Offline Offline

Posts: 656


« Reply #57 on: April 08, 2013, 05:04:10 PM »
ReplyReply

OK, my take is that to make the best use of any camera we need to expose correctly to the right. That is the only way to fully utilize the sensor.

In theory, we should expose to the right. In practice, as opposed to theory, we tend to oversaturate a small range of pixels by doing so and don't notice them as they are too few to show on the histogram. These are the values needed for the smooth high tones that the O.P. is seeking.
Logged
alosurdo
Newbie
*
Offline Offline

Posts: 40


« Reply #58 on: April 08, 2013, 05:44:35 PM »
ReplyReply

Hi, sorry to get off track here. I am looking at a few raw files from a p30+ back. I have loaded them in raw digger and this time get the horizontal showing approx 60000. However when i go to the exif under bits ber sampel it says 8 8 8 ... wondering what this means and why not 16.

TO some of the most recent comments, i do apologise as i am not as technically learned in sensor technology and breaking it down. FOr my images, DR is important, thought i would say this means in the application of post processing. i.e. what can i get back from the raw files. With my 5d2, i expose way to the right, sometimes a little to far just because if it is a large DR scene i know the shadows will be gone if not. One of the qualities i am looking for is tonal smoothness in the highlights, as well as recoverable shadows, though highlight tonal graduation or smoothness (what ever you want to call it) is something of importance to me.
Logged
FredBGG
Sr. Member
****
Offline Offline

Posts: 1651


« Reply #59 on: April 08, 2013, 09:59:19 PM »
ReplyReply

Here is a follow up to my previous post showing the dynamic range and shadow recovery of the D800 using an in camera jpeg.
This time I used Adobe Camera Raw. Also this is full frame showing highlight all the way to shadows in the recovered image.



Despite being a very dark capture it can be recovered with very natural colors and gradation form the detailed off light beige wall to the shadows behind the plant.
« Last Edit: April 08, 2013, 10:02:58 PM by FredBGG » Logged
Pages: « 1 2 [3] 4 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad