Ad
Ad
Ad
Pages: « 1 [2] 3 4 5 »   Bottom of Page
Print
Author Topic: "Diglloyd" is testing the Leica S2  (Read 17886 times)
ejmartin
Sr. Member
****
Offline Offline

Posts: 575


« Reply #20 on: July 15, 2010, 11:34:33 PM »
ReplyReply

Quote from: LKaven
There is a clearly some kind of unique technological achievement in this camera that brings the noise floor down to bedrock, and I'd like to know what it is.

Odds are it's the column parallel readout of the Sony Exmor sensor, which changes the readout rate from MHz to KHz.  The ~1000x slower readout allows lower read noise levels and thus cleaner shadows.  What puzzles me is how much more Nikon was able to achieve with this sensor than Sony themselves.
Logged

emil
ErikKaffehr
Sr. Member
****
Online Online

Posts: 7406


WWW
« Reply #21 on: July 15, 2010, 11:53:33 PM »
ReplyReply

Hi,

I got the impression that the Sony ADC:s are only 12 bit but the Nikon has obviously 14 bits. Nikon has either 12 bit readouts which is fast or 14 bit that is quite slow, as far as I understand. Has someone compared the methods?

It has been suggested that the Sony and Nikon sensors are similar but that the Nikon sensor may have design changes.

Best regards
Erik



Quote from: ejmartin
Odds are it's the column parallel readout of the Sony Exmor sensor, which changes the readout rate from MHz to KHz.  The ~1000x slower readout allows lower read noise levels and thus cleaner shadows.  What puzzles me is how much more Nikon was able to achieve with this sensor than Sony themselves.
Logged

ErikKaffehr
Sr. Member
****
Online Online

Posts: 7406


WWW
« Reply #22 on: July 16, 2010, 12:15:30 AM »
ReplyReply

Hi Guy,

I much appreciate the stuff you and Jack are doing at GetDPI.

Regarding Lloyds comparing the Leica S2 to Nikon it is quite valid, in my view. The interesting factor is what can be achieved not the price of the tools. Price tags don't make pictures, cameras do.

Lloyd's experience is with DSLRs, Canon and Nikon, mostly. Lloyd did test MF digital now and than, however.

Obviously, what you are shooting matters a lot. It seems that Lloyd has problems focusing the Leica S2 and also the Leica M9. He did communicate with Erwin Puts, a known Leica tester and it seems that focus bracketing is the only way of achieving optimal focus. Lloyd and many others see Live View as the ultimate solution. Live view also eliminates half the alignment problem as the actual sensor signal is used for focusing. The optical axis of the lens still needs be perpendicular to the sensor but longitudinal errors don't matter with Live View. Manual focus is also depending on alignment of sensor, mirror and viewing screen.

Your point on the market position of Leica S2 is dead on in my view.

Best regards
Erik


Quote from: fredjeang
On earth today, there are some gear that you can't go wrong with.
The D3x is one, the S2 is also on the list.
All will be a question of handling, style, economy and needs.

At this level, what matters is what do you want to acheive and how.

To me, the Leica is in a very hard position between a D3x and the best MFs,
without being as good as any of those in their strengh.

Can't really figure for who is this S2.  

Leica might knows.

Also, I find a little bit (diplomatically speaking) "ridiculous" comparing the S2 to the D3x, and sorry but this S2 should be compared to the same priced equipement from Phase and Hasselblad.
That is a good think indeed that Lu-La and the GetDPI did understand that from the very beginning.
See that Guy's pic I found on google? those are cameras in the same league-price-target. Not the D3x, despite being a top camera.
[attachment=23177:s2_getdp...merastud.jpg]

If I was a potential Leica buyer, I would not be interested at all to see how it compares to the Nikon or any 35mm. IMO, Lloyd missed the shot on that. (Lloyd is highly recommended anyway).

Take the Lu-La S2 testing, mixed with the GetDPI review, shake the all (in Bond's style) and you got a pretty extented and trustable information.

Conclusion (personal), why paying when you have good infos for free?
Logged

LKaven
Sr. Member
****
Offline Offline

Posts: 811


« Reply #23 on: July 16, 2010, 04:24:31 AM »
ReplyReply

Quote from: Nick Rains
BTW, the D3X arrived with the latest 24-70 zoom. I know it's a good lens but is it going to be good enough to show up the sensor vs the S2/35/70mm lenses. it's a new lens for me, not used it before but I hear good things about it.
The 24-70 is an extremely sharp lens, but it has the bokeh of a zoom lens.  
One intuition says that if the Leica gets a $5k lens, so does the Nikon.  I'd nominate the Coastal Optics 60mm Apochromat.  On the other hand, another intuition says that if the Nikon is half the price, the lens should be $2500.  Um, then again, yet another intuition says that you should use the best lenses Nikon has available -- at an eighth to one quarter of the price.  The Nikon 60mm AF-s Micro comes to mind for that, a favorite of some D3x fashion shooters.
Logged

LKaven
Sr. Member
****
Offline Offline

Posts: 811


« Reply #24 on: July 16, 2010, 04:39:54 AM »
ReplyReply

Quote from: ejmartin
What puzzles me is how much more Nikon was able to achieve with this sensor than Sony themselves.
That's just the question I was hoping you had the answer to.  For a minute, I thought someone had taken you up on your dual-processed active read-noise optimization strategy.  But you seemed to suggest somewhere earlier that the observed numbers from the camera did not show the telltale signs of this.
Logged

Nick Rains
Sr. Member
****
Offline Offline

Posts: 704



WWW
« Reply #25 on: July 16, 2010, 05:06:45 AM »
ReplyReply

Quote from: LKaven
The 24-70 is an extremely sharp lens, but it has the bokeh of a zoom lens.  
One intuition says that if the Leica gets a $5k lens, so does the Nikon.  I'd nominate the Coastal Optics 60mm Apochromat.  On the other hand, another intuition says that if the Nikon is half the price, the lens should be $2500.  Um, then again, yet another intuition says that you should use the best lenses Nikon has available -- at an eighth to one quarter of the price.  The Nikon 60mm AF-s Micro comes to mind for that, a favorite of some D3x fashion shooters.

Yes, agreed but I'll go with the best 'all Nikon' solution. The 24-70 at f5.6 is quite amazingly sharp, both to my eye and in the Photozone tests where it is as good at f5.6 as almost any other lens in the Nikon arsenal. I'd like to try the 60mm macro but it's not going to happen.

Today I shot this Nikon 24-70 lens @ 70mm against the Leica 75mm to get a sense of how the sensors resolve the same details. Details in both images appear much the same size at 100%, as you'd expect, but the Leica gives a better result - quite clearly. (This is from both ACR and C1 for both cameras.) This method eliminates the sensor size difference, so obviously matching the field of view of both cameras by using a 50mm lens on the Nikon would end up favouring the Leica even more as it's bringing 37.5MP vs 25MP to bear on the same field size.

It's impressive just now good the Nikon is though, brilliant metering accuracy and zero noise. The Leica is just better, but you have to pay a lot more (2-3 times)to get that step up in IQ. This is pretty much what I expected but it's certainly interesting to have arguably (or not!) the best 35mm DSLR currently available to compare to. I was hoping to have a Hassy or Phase to look at too but these have not been forthcoming from the distributors.
Logged

Nick Rains
Australian Photographer
Leica Akademie Instructor
www.nickrains.com
ErikKaffehr
Sr. Member
****
Online Online

Posts: 7406


WWW
« Reply #26 on: July 16, 2010, 05:26:28 AM »
ReplyReply

Hi,

Someone, perhaps me, suggested that two readings could be made with different pre-amp setings like 100 ISO and 400 ISO and the readings could be marged. CMOS sensors can be read nondestructively so this may be possible. Emil wrote that such mainipulation would have a very obvious signature in the file and that signature isn't there.

Best regards
Erik

Quote from: LKaven
That's just the question I was hoping you had the answer to.  For a minute, I thought someone had taken you up on your dual-processed active read-noise optimization strategy.  But you seemed to suggest somewhere earlier that the observed numbers from the camera did not show the telltale signs of this.
Logged

Guy Mancuso
Sr. Member
****
Offline Offline

Posts: 1110


WWW
« Reply #27 on: July 16, 2010, 05:42:30 AM »
ReplyReply

Quote from: ErikKaffehr
Hi Guy,

We had a lengthy discussion on the forum on both DR and it's interpretation with quite a few members chiming in who had impressive knowledge in image processing and all essentially say that the DxO figures are correct. Lloyd Chambers also found that DxO numbers are correct. There were some theories trying to explain why experienced photographers see a DR advantage.

Please remember that DxO-mark is about raw data that has not been processed. It's data prior to conversion.


Best regards
Erik


Not sure you read me right but yes that is what I am saying not that the numbers are wrong but it is before raw conversion and as a shooter they really mean nothing at all because everything we do is after the raw conversion. On a scientific level sure DXO numbers and testing are okay but we deal with life after the pure raw data so in effect it is kind of meaningless. To me it is what C1 does to my Phase files as they are brought in what counts. And those files do change a lot when brought into there dedicated software

 Not that we need to get into a whole discussion here but Phase , Hassy, Leaf and Sinar files are all tuned to there converters and when brought in those big changes happen. Nikon, Canon and all of those that produce in camera jpegs I believe have some type of raw conversion going (or algorithms built into the file) in camera on the raws before they are considered raw data. I believe the reason why is the jpegs in camera are obviously being processed as well and the raws follow along to a certain level. Something maybe a new thread to explore for sure. The MF only put out raw files so they have no need for any algorithms built into the file in the back so that data is very linear and the software adds those algorithms . My guess and it is a guess is these backs do absolutely nothing at the shooting stage with regards to the raw but the Nikons and in camera jpegs actually do something to the raws even before they are considered raw data. Maybe something worth discussing in a thread. Obviously this is the science part behind what is going on that we really never hear about. Hope that made sense very early in the morning and on first espresso. LOL
Logged

Guy Mancuso
Sr. Member
****
Offline Offline

Posts: 1110


WWW
« Reply #28 on: July 16, 2010, 05:50:16 AM »
ReplyReply

Quote from: ErikKaffehr
Hi Guy,

I much appreciate the stuff you and Jack are doing at GetDPI.

Regarding Lloyds comparing the Leica S2 to Nikon it is quite valid, in my view. The interesting factor is what can be achieved not the price of the tools. Price tags don't make pictures, cameras do.

Lloyd's experience is with DSLRs, Canon and Nikon, mostly. Lloyd did test MF digital now and than, however.

Obviously, what you are shooting matters a lot. It seems that Lloyd has problems focusing the Leica S2 and also the Leica M9. He did communicate with Erwin Puts, a known Leica tester and it seems that focus bracketing is the only way of achieving optimal focus. Lloyd and many others see Live View as the ultimate solution. Live view also eliminates half the alignment problem as the actual sensor signal is used for focusing. The optical axis of the lens still needs be perpendicular to the sensor but longitudinal errors don't matter with Live View. Manual focus is also depending on alignment of sensor, mirror and viewing screen.

Your point on the market position of Leica S2 is dead on in my view.

Best regards
Erik


Agree the issue is the S2 is a tweener neither 35 nor MF. Kind of a two part cam as far as comparing. The ergos and features against the 35mm  and image quailty against the MF . It does not win in either court but combined it is very good but as a tweener. Obviously it is a strange duck because it does not fit either but compares against both. On one hand certainly not two times better over the D3x but the price is and no better than the Hasy , Phase 40mpx offerings but more expensive. When buying it throws a big curve ball into the decision process because of it's mix.
Logged

wildlightphoto
Sr. Member
****
Offline Offline

Posts: 628


« Reply #29 on: July 16, 2010, 06:59:56 AM »
ReplyReply

Quote from: Guy Mancuso
Agree the issue is the S2 is a tweener neither 35 nor MF. Kind of a two part cam as far as comparing. The ergos and features against the 35mm  and image quailty against the MF . It does not win in either court but combined it is very good but as a tweener. Obviously it is a strange duck because it does not fit either but compares against both. On one hand certainly not two times better over the D3x but the price is and no better than the Hasy , Phase 40mpx offerings but more expensive. When buying it throws a big curve ball into the decision process because of it's mix.

For me the S2 fixes problems I have with both small-format cameras and medium-format cameras: it has a bigger sensor than the 35mm-based DSLRs and better handling than the typical medium-format camera.  Except for the price and the lack of long lenses it hits a bullseye for me.
Logged
ErikKaffehr
Sr. Member
****
Online Online

Posts: 7406


WWW
« Reply #30 on: July 16, 2010, 07:09:22 AM »
ReplyReply

Guy,

My post was also early morning after first espresso...

Best regards
Erik

Quote from: Guy Mancuso
Not sure you read me right but yes that is what I am saying not that the numbers are wrong but it is before raw conversion and as a shooter they really mean nothing at all because everything we do is after the raw conversion. On a scientific level sure DXO numbers and testing are okay but we deal with life after the pure raw data so in effect it is kind of meaningless. To me it is what C1 does to my Phase files as they are brought in what counts. And those files do change a lot when brought into there dedicated software

 Not that we need to get into a whole discussion here but Phase , Hassy, Leaf and Sinar files are all tuned to there converters and when brought in those big changes happen. Nikon, Canon and all of those that produce in camera jpegs I believe have some type of raw conversion going (or algorithms built into the file) in camera on the raws before they are considered raw data. I believe the reason why is the jpegs in camera are obviously being processed as well and the raws follow along to a certain level. Something maybe a new thread to explore for sure. The MF only put out raw files so they have no need for any algorithms built into the file in the back so that data is very linear and the software adds those algorithms . My guess and it is a guess is these backs do absolutely nothing at the shooting stage with regards to the raw but the Nikons and in camera jpegs actually do something to the raws even before they are considered raw data. Maybe something worth discussing in a thread. Obviously this is the science part behind what is going on that we really never hear about. Hope that made sense very early in the morning and on first espresso. LOL
Logged

eronald
Sr. Member
****
Offline Offline

Posts: 4026



« Reply #31 on: July 16, 2010, 07:38:47 AM »
ReplyReply

Quote from: ErikKaffehr
Guy,

My post was also early morning after first espresso...

Best regards
Erik

Any company which expects its clients to use LR should do the processing onboard. Adobe can not be expected to spend a lot of time fine tuning their algorithms or profiles for a particular camera. Assuming they knew how to, which given the quality of their results ... I find the quality of the D3x Jpegs good enough for almost every use now -

Edmund
« Last Edit: July 16, 2010, 07:40:38 AM by eronald » Logged
rainer_v
Sr. Member
****
Offline Offline

Posts: 1130


WWW
« Reply #32 on: July 16, 2010, 08:15:23 AM »
ReplyReply


if there comes our just one file for each image there is done processing on  board of the camera.


Logged

rainer viertlböck
architecture photographer
munich / germany

www.tangential.de
Guy Mancuso
Sr. Member
****
Offline Offline

Posts: 1110


WWW
« Reply #33 on: July 16, 2010, 08:18:21 AM »
ReplyReply

Quote from: ErikKaffehr
Guy,

My post was also early morning after first espresso...

Best regards
Erik


Let's move on to number two. LOL

Have a great morning
Logged

Guy Mancuso
Sr. Member
****
Offline Offline

Posts: 1110


WWW
« Reply #34 on: July 16, 2010, 08:31:10 AM »
ReplyReply

Quote from: eronald
Any company which expects its clients to use LR should do the processing onboard. Adobe can not be expected to spend a lot of time fine tuning their algorithms or profiles for a particular camera. Assuming they knew how to, which given the quality of their results ... I find the quality of the D3x Jpegs good enough for almost every use now -

Edmund


 I think this is worth exploring at some point in a thread. OEM's like Nikon, Canon, Leica and some others do not have sophisticated raw converters( yes some do). That ONLY shoot raw like the MF back and these companies put everything in the software(algorithms). But companies that use jpeg in the camera the question is are they doing some onboard raw processing(or algorithms) because they did not design there software(Nikon and Canon) like that or rely on third party so they have to put some in to get it correctly in different programs . We all know Hassy, Phase , Sinar and Leaf design there software specifically for there backs and my bet put those algorithms in the software not the files themselves and why we see there raws as being pretty poorly looking or linear without there own software package. I know Phase files do not look good in ACR or LR for instance and need work. This alone makes me believe this theory. Sorry I know we are OT but it is a interesting thought because it never comes up in conversations at the root level. Sorry for rambling here.
Logged

ejmartin
Sr. Member
****
Offline Offline

Posts: 575


« Reply #35 on: July 16, 2010, 08:31:10 AM »
ReplyReply

Quote from: Nick Rains
Yes, agreed but I'll go with the best 'all Nikon' solution. The 24-70 at f5.6 is quite amazingly sharp, both to my eye and in the Photozone tests where it is as good at f5.6 as almost any other lens in the Nikon arsenal. I'd like to try the 60mm macro but it's not going to happen.

Today I shot this Nikon 24-70 lens @ 70mm against the Leica 75mm to get a sense of how the sensors resolve the same details. Details in both images appear much the same size at 100%, as you'd expect, but the Leica gives a better result - quite clearly. (This is from both ACR and C1 for both cameras.) This method eliminates the sensor size difference, so obviously matching the field of view of both cameras by using a 50mm lens on the Nikon would end up favouring the Leica even more as it's bringing 37.5MP vs 25MP to bear on the same field size.

It's impressive just now good the Nikon is though, brilliant metering accuracy and zero noise. The Leica is just better, but you have to pay a lot more (2-3 times)to get that step up in IQ. This is pretty much what I expected but it's certainly interesting to have arguably (or not!) the best 35mm DSLR currently available to compare to. I was hoping to have a Hassy or Phase to look at too but these have not been forthcoming from the distributors.


In what aspects is the Leica result better?  Not being argumentative, just curious.  Can one put the difference down to the AA filter?  That will certainly rob the Nikon of a bit of MTF near Nyquist, while cutting down on aliasing artifacts.
Logged

emil
ErikKaffehr
Sr. Member
****
Online Online

Posts: 7406


WWW
« Reply #36 on: July 16, 2010, 09:13:53 AM »
ReplyReply

Hi,

Lloyd made some raws available for download. It's his mosaic target which contains lot of fine detail close to Nyquist (I think). I made a quick comparison with strong capture sharpening on Nikon and little on the Leica. No doubt Leica resolves better, but the Leica had a lot of artifacts. I could not see a significant difference in sharpness, Nikon was cleaner and Leica did resolve detail the Nikon did not resolve.

Best regards
Erik



Quote from: ejmartin
In what aspects is the Leica result better?  Not being argumentative, just curious.  Can one put the difference down to the AA filter?  That will certainly rob the Nikon of a bit of MTF near Nyquist, while cutting down on aliasing artifacts.
Logged

BernardLanguillier
Sr. Member
****
Offline Offline

Posts: 7975



WWW
« Reply #37 on: July 16, 2010, 09:54:31 AM »
ReplyReply

Quote from: Guy Mancuso
Bernard we have been going round and round on DXO but the bottom line on it it simply does not account for raw processing software dedicated to the backs in question. No question C1 does a lot of stuff to the Phase files as soon as you bring them in. So how can we effectively understand the DXO marks when this is done in the raw converters and bottom line we use the raw converters to process so you can't ignore it as the working tool and end of day that is really what counts is what is going on with the output from the raw converter, I could careless what happens before it is brought in and that is there measurement. No matter how we slice the cheese that is the file going to print, to client or to press. I'm not here to argue DXO frankly I could care less about it and need to get back to Hospice with my mother in law but I think DXO is good and a nice measurement but I also believe whole heartedly people are forgetting the power of dedicated back to it's dedicated software which no question about it Phase and Hassy, Sinar and Leaf that is the life bread of there backs is the software. There is a lot of baseline tweaking going on at the sensor level with these backs. Anyway gotta run and obviously that was OT concerning the S2. I just don't put as much faith in those numbers, I look at it as a guide some think it is the bible. Whatever makes one happy

Guy,

I am a long term C1 user and use it excusively for my D3x files... because I get even cleaner shadows than with other RAW converters.

The magic appliced by the software is real, but does not only apply to Phase backs files.

Cheers,
Bernard
Logged

A few images online here!
ejmartin
Sr. Member
****
Offline Offline

Posts: 575


« Reply #38 on: July 16, 2010, 10:11:01 AM »
ReplyReply

Quote from: BernardLanguillier
Guy,

I am a long term C1 user and use it excusively for my D3x files... because I get even cleaner shadows than with other RAW converters.

The magic appliced by the software is real, but does not only apply to Phase backs files.

Cheers,
Bernard

When it comes to converters, it's a rather subjective thing.  I used to like the detail and default colors from C1, as well as the good shadows, but after using it for a while I became disenchanted with some of the demosaic artifacts, and the fact that much of the shadow noise performance comes from strongly desaturating them.  I never had the pro version, just the low end software up through 4.8, so I don't know if there are extras in the high end version that would change that conclusion.

Each converter has its plusses and minuses, and I agree that the software from the camera manufacturer may be tuned a bit better to the camera, especially when it comes to color profiling, they know very well the spectral properties of their color filters and could and should spend more effort adjusting the profile.  Other than that, noise characteristics differ from camera to camera, so default NR from the converter can be tuned to suppress any idiosyncrasies from the camera, and again one would expect the hardware manufacturer to spend a bit more effort with that adjustment.  Perhaps the MFDB manufacturers treat moire with more attention (I know C1 has a moire suppression tool; I don't know if it is enabled by default).  OTOH, demosaic algorithms play a big part in output quality, and that is independent of what camera is being used so long as its a standard Bayer array.  
Logged

emil
JeffKohn
Sr. Member
****
Offline Offline

Posts: 1671



WWW
« Reply #39 on: July 16, 2010, 10:19:25 AM »
ReplyReply

Quote from: Guy Mancuso
I think this is worth exploring at some point in a thread. OEM's like Nikon, Canon, Leica and some others do not have sophisticated raw converters( yes some do). That ONLY shoot raw like the MF back and these companies put everything in the software(algorithms). But companies that use jpeg in the camera the question is are they doing some onboard raw processing(or algorithms) because they did not design there software(Nikon and Canon) like that or rely on third party so they have to put some in to get it correctly in different programs . We all know Hassy, Phase , Sinar and Leaf design there software specifically for there backs and my bet put those algorithms in the software not the files themselves and why we see there raws as being pretty poorly looking or linear without there own software package. I know Phase files do not look good in ACR or LR for instance and need work. This alone makes me believe this theory. Sorry I know we are OT but it is a interesting thought because it never comes up in conversations at the root level. Sorry for rambling here.
Just because DSLR's have internal JPEG engine doesn't mean that the raw files are any less raw than from MFDB. I know of no evidence at all to suggest there is any processing happening on the RAW files for Nikon DSLR's. In fact there's plenty of evidence to the contrary. Nikon's CaptureNX RAW converter has things like dynamic lighting and automatic lens correction (for CA, vignetting, etc), and you don't get that functionality unless you use their converter. It's not baked into the RAW.
Logged

Pages: « 1 [2] 3 4 5 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad