stngoldberg
Well-known member
One can eliminate the noise almost completely in Phocus.
Read the Ohocus Manuel to learn how to reduce/eliminate he noise
Stanley
The GetDPI Photography Forum
Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!
Hasselblad 50 mp backs superior on tech cams! Why didn't anyone tell us?
- Thread starter torger
- Start date
Read the Ohocus Manuel to learn how to reduce/eliminate he noise
Stanley
stngoldberg
Well-known member
The Breakers, the Cornelius Vanderbilt house in Newport, taken in great light this afternoon with the H5D50 and the Arca Swiss Rm3di back and the Rodenstock 32mm lens with one degree tilt.
Stanley
Stanley
DDudenbostel
Active member
I used my CFV39 on a Technikardan 23 with wides to long glass with extreme movements and had no problems. I don't believe the sensor has micro lenses either. In any case it performed very well.
I'm actually selling it through Popflash if anyone is interested. It's mint with all the goodies. Low frame count too.
I'm actually selling it through Popflash if anyone is interested. It's mint with all the goodies. Low frame count too.
Yes, it will require center filter. The sensor does not crosstalk much, but instead vignettes so even with center filter you lose a couple of stops to the edge.
Here's an image that illustrates crosstalk:
this is when the angle is large enough so you get a leak over to the next pixel. However if the angle is smaller it instead hits the light shield and it does not get registered, this is "pixel vignetting". The Dalsa doesn't seem to have much of a light shield at all, thus you have less pixel vignetting than the Kodak, but instead a massive amount of crosstalk on the symmetric wides.
If the Kodak has had the same low noise levels on the shadows as the new Sony CMOS MF sensor the pixel vignetting on the SK28 would be a non-issue. Unfortunately Kodak is among the noisier sensors with today's standards which means that in some occassions you might want to bracket and merge to get the image quality you want.
Another more elegant solution would be to use a stronger center filter which also compensates for the pixel vignetting. The current center filter is 2 stops (which actually is a 0.5 stops weaker than the Rodenstock 32 filter, so the 28 is most likely a bit undercorrected), and this sensor would need a 4-4.5 stop center filter to function optimally. There is no such filter on the market though, and you would get an awfully dark image if you focus on ground glass.
Another option to shoot this lens is to skip the center filter and always bracket for your large-shifted images. As it's a 4 stop difference or so with this sensor, you'd probably shoot three images, 0 +2 and +4 stops. There can be a good reason to do so as a center filter on a lens this extremely wide is prone for reflections and contrast issues.
Here's an image that illustrates crosstalk:
this is when the angle is large enough so you get a leak over to the next pixel. However if the angle is smaller it instead hits the light shield and it does not get registered, this is "pixel vignetting". The Dalsa doesn't seem to have much of a light shield at all, thus you have less pixel vignetting than the Kodak, but instead a massive amount of crosstalk on the symmetric wides.
If the Kodak has had the same low noise levels on the shadows as the new Sony CMOS MF sensor the pixel vignetting on the SK28 would be a non-issue. Unfortunately Kodak is among the noisier sensors with today's standards which means that in some occassions you might want to bracket and merge to get the image quality you want.
Another more elegant solution would be to use a stronger center filter which also compensates for the pixel vignetting. The current center filter is 2 stops (which actually is a 0.5 stops weaker than the Rodenstock 32 filter, so the 28 is most likely a bit undercorrected), and this sensor would need a 4-4.5 stop center filter to function optimally. There is no such filter on the market though, and you would get an awfully dark image if you focus on ground glass.
Another option to shoot this lens is to skip the center filter and always bracket for your large-shifted images. As it's a 4 stop difference or so with this sensor, you'd probably shoot three images, 0 +2 and +4 stops. There can be a good reason to do so as a center filter on a lens this extremely wide is prone for reflections and contrast issues.
Last edited:
Enda Cavanagh
Member
+1
+1 absolutely.
Phase One / Leaf and Hasselblad should all hang their heads in shame. Last time I checked there were about a 'staggering' 12 or so lenses in the (latest) range of 23 - 60 from S/K and R/S designed for digital use. When you're asking $30k - $40k for your product, would it be too much to ask for just a couple of days testing with these so that potential buyers are fully informed as to exactly how that new DB being promoted is going to perform? Yes, users of these lenses are a small % overall, but they're the ones left in the dark. For $30k - $40k a pop, surely they can afford to find someone to switch the lights on.
Can't speak to Hasselblad, but Phase One for US customers, has done an OK job with their dealer channel. Phase is setup that way and I don't feel you are ever going to see much more reach out.
Digital Transitions and Capture Integration both offer a lot of information on the various tech cameras and lenses. Both offer software tools and demos. I realize that if you don't live close to one of their dealer locations, getting a demo can be a bit difficult.
I have to also say, that I finally was able to pull down the article that Quentin pointed to on Hasselbald and their tech offerings and overall besides being a bit dated on available cameras, the info was very well presented and informative. A lot of good solutions were listed and mapped out very well.
The "crosstalk" phenomenon has never been indorsed by any dealer or manufacturer of MFD that I have seen. Not to say that means that is doesn't exist, it's just that very little has been published by the camera companies that I can find. No doubt the effect is real and Torger's work does go a long way to explaining both the cause and ways to get around it. In fact "crosstalk" didn't seem to exist on the web forums until the IQ250 shots started to show up, even though it's been a problem with the 60MP and 80MP backs from day one. I however always assumed this loss of color/saturation was due instead to the lens/image circle and not the chip design.
It's always a trade off, and for me, the massive increase in DR with the Dalsa chips is worth it. Having shot the 51000's smaller cousin in the P45+ for almost 4 years, I for one prefer the ability to have less noise, much less noise and useable details at base iso. Neither Dalsa or Kodak offer much push past base iso, from my experience.
Paul
Digital Transitions and Capture Integration both offer a lot of information on the various tech cameras and lenses. Both offer software tools and demos. I realize that if you don't live close to one of their dealer locations, getting a demo can be a bit difficult.
I have to also say, that I finally was able to pull down the article that Quentin pointed to on Hasselbald and their tech offerings and overall besides being a bit dated on available cameras, the info was very well presented and informative. A lot of good solutions were listed and mapped out very well.
The "crosstalk" phenomenon has never been indorsed by any dealer or manufacturer of MFD that I have seen. Not to say that means that is doesn't exist, it's just that very little has been published by the camera companies that I can find. No doubt the effect is real and Torger's work does go a long way to explaining both the cause and ways to get around it. In fact "crosstalk" didn't seem to exist on the web forums until the IQ250 shots started to show up, even though it's been a problem with the 60MP and 80MP backs from day one. I however always assumed this loss of color/saturation was due instead to the lens/image circle and not the chip design.
It's always a trade off, and for me, the massive increase in DR with the Dalsa chips is worth it. Having shot the 51000's smaller cousin in the P45+ for almost 4 years, I for one prefer the ability to have less noise, much less noise and useable details at base iso. Neither Dalsa or Kodak offer much push past base iso, from my experience.
Paul
No doubt about it - if you're in the US you're lucky to have dealers like CI and DT. Especially if something goes wrong - they have clout. If you live on the remaining 93.4% of the earth's land mass, P1 dealers are a lottery.
IMHO P1 shouldn't be asking their dealer network to do work they should be doing themselves. Outwith the US, that dealer network just doesn't exist at anything like the same level, so it essentially leaves global testing to CI/DT (plus those with the skills - and access to equipment - to do it properly, like a number of folks here on GetDPI).
As much as I welcome this testing, more often than not the results are on blog posts with 'email for RAWs'. Remember CIs testing of the IQ180 and the S/K 28? A prize to anyone who finds the link to download the RAWs from that test (here's a link to the page, but unfortunately the form which you have to fill in is missing).
If anyone from CI or DT (or any P1 dealer) is reading this - please just post the resultant RAWs from testing to dropbox (or similar) and leave them up there permanently with a readme.txt file listing what's what. Dropbox has 250GB per day of bandwidth available to business users - enough for over 3000 IQ180 RAW downloads per day - that should cover it (and you can still post the links to the files on your own site in order to drive traffic towards you initially and put the testing in context).
Just my 2c.
IMHO P1 shouldn't be asking their dealer network to do work they should be doing themselves. Outwith the US, that dealer network just doesn't exist at anything like the same level, so it essentially leaves global testing to CI/DT (plus those with the skills - and access to equipment - to do it properly, like a number of folks here on GetDPI).
As much as I welcome this testing, more often than not the results are on blog posts with 'email for RAWs'. Remember CIs testing of the IQ180 and the S/K 28? A prize to anyone who finds the link to download the RAWs from that test (here's a link to the page, but unfortunately the form which you have to fill in is missing).
If anyone from CI or DT (or any P1 dealer) is reading this - please just post the resultant RAWs from testing to dropbox (or similar) and leave them up there permanently with a readme.txt file listing what's what. Dropbox has 250GB per day of bandwidth available to business users - enough for over 3000 IQ180 RAW downloads per day - that should cover it (and you can still post the links to the files on your own site in order to drive traffic towards you initially and put the testing in context).
Just my 2c.
I live in the remaining 93.4% of the world, as mentioned above. The dealer network in my neck of the woods is truly atrocious. There was a brief moment when we had the option of going through a really good, enthusiastic guy who knew his stuff and, shock horror, actually USED the equipment for his own business. Phase decided he wasn't selling enough and gave sole distributorship to a company that most would rather walk over broken glass and negotiate alligator pits to avoid dealing with. Truly horrible service, not to mention inflated prices. In terms of Hasselblad, things are a lot better and that alone sways me to use their products.
With regards to Hasselblad CFV backs, can anyone here state categorically that the sensor in the H5D-50c and CFV-50c is spec'd exactly the same as the one in the Phase IQ250? Do they use the same micro lenses and everything? I guess I'm asking on the off chance it might perform better on technical cameras compared to the competition.
With regards to Hasselblad CFV backs, can anyone here state categorically that the sensor in the H5D-50c and CFV-50c is spec'd exactly the same as the one in the Phase IQ250? Do they use the same micro lenses and everything? I guess I'm asking on the off chance it might perform better on technical cameras compared to the competition.
photomgraphy
Member
thank you torger for confirming my decision to invest in a CFV-50 last spring...
it has been an unbeatable priced option and i like it very much.
i primarily use it on my arca rm3di, and just for fun bought a hasselblad 500cm in addition to it. i did always love the series 500 and now i'm so happy to be back to this lovely system again ;-)
the only nuisance is the already mentioned lack of power indication.
it's unbelievable that such a high engineered product doesn't offer this simple feature. :loco:
but just in case i carry a spare battery and also my silvestry external battery which i used to use on my h3d-39 too. so no worries at all
it has been an unbeatable priced option and i like it very much.
i primarily use it on my arca rm3di, and just for fun bought a hasselblad 500cm in addition to it. i did always love the series 500 and now i'm so happy to be back to this lovely system again ;-)
the only nuisance is the already mentioned lack of power indication.
it's unbelievable that such a high engineered product doesn't offer this simple feature. :loco:
but just in case i carry a spare battery and also my silvestry external battery which i used to use on my h3d-39 too. so no worries at all
Shashin
Well-known member
The CFA array is probably the same. The profile and processing is most likely different.
Haven't seen any specific tests of the 50c, but it would be extremely surprising if it's not exactly the same sensor. I don't think CFAs differ either, they could be different from Sonys 35mm sensors though. At some point I hope for a side-by-side raw test shot from Pentax 645z / IQ250 and H5D-50c, by processing them in RawTherapee you can apply the exact same processing and then see if sensors differ or it's just the raw converter's pipelines. I don't think it's likely that any dealer would provide such shots to play with though .
I'm no sensor design specialist, but it does seem to me that having an option to buy the Sony sensor without microlenses would not make much sense if you could not also get it with proper light shields. As being without microlenses is supposed to increase angular response (at least without ripple artifacts) you need light shields too to make it worthwhile otherwise you just get lots of crosstalk. While not having microlenses probably just removes one layer from the sensor and doesn't change much at all, having light shields would be a substantial design change. All indications from the IQ250 raw files I've looked at is that the sensor has very weak shielding between pixels, thus it would still perform very bad with tech wides even without microlenses. The microlenses that is there don't ripple either as on the Dalsa, so probably there's not much gain at all from removing them.
Additionally, the CFV-50c would perform worse with the V system without microlenses, so no I have no hopes there that it would perform better than an IQ250 with a tech cam :-\.
.I'm no sensor design specialist, but it does seem to me that having an option to buy the Sony sensor without microlenses would not make much sense if you could not also get it with proper light shields. As being without microlenses is supposed to increase angular response (at least without ripple artifacts) you need light shields too to make it worthwhile otherwise you just get lots of crosstalk. While not having microlenses probably just removes one layer from the sensor and doesn't change much at all, having light shields would be a substantial design change. All indications from the IQ250 raw files I've looked at is that the sensor has very weak shielding between pixels, thus it would still perform very bad with tech wides even without microlenses. The microlenses that is there don't ripple either as on the Dalsa, so probably there's not much gain at all from removing them.
Additionally, the CFV-50c would perform worse with the V system without microlenses, so no I have no hopes there that it would perform better than an IQ250 with a tech cam :-\.