Ahhh....now I don't feel so bad about the res of my CFV-II.....whew, just saved me the heartburn of trying to come up with $14k that I don't have and shouldn't spend if I did! Probably saved my marriage too!
Gary
Medium Format Forum
Register a free account now!
If you are registered, you get access to the members only section, can participate in the buy & sell second hand forum and last but not least you can reserve your preferred username before someone else takes it.
First results from the new CFV 39 part 2
- Thread starter polypal
- Start date
- Status
Ahhh....now I don't feel so bad about the res of my CFV-II.....whew, just saved me the heartburn of trying to come up with $14k that I don't have and shouldn't spend if I did! Probably saved my marriage too!
Gary
CFVI-16 vs CFV-39, does size really matter that much?
Hi everyone, I hope you are prepared for a semi detailed observation of what we loose or not loose with the smaller pixel of the CFV39 as apposed to the fat pixel of the CFVI16.
All the comparison photos were taken Hand held. I used my 205TCC on this endeavor with my 40IF and CF180. Apertures were always at f5.6, I focused, put the mirror up, and gently squeezed the trigger. Shutter speed was between 1/750 and 1/1000 and in manual mode. Almost all the pictures were taken at iso 100.
The pictures were taken in Jeddah, Saudi Arabia. Outside temperature was 38C, and it was mid-afternoon, not such good lighting. Really hot, but the process was so exciting that I didn’t care at all, but I did go back in the car every 10 minutes to cool off.
My expectation was that the pictures from the CFV39 would be clearly and undeniably superior to the CFV16. After downloading the pictures my first impression was that the CFV16 was more appealing in terms of saturation, contrast and dynamic range. This was depressing. In normal viewing the CFV39 was sharper and much more appealing in that respect. In actual pixel the CFV16 was sharper with the 40IF (but strangely enough it was the same with the 180CF, needs more testing), to get the CFV39 to the same level of sharpness I had to add between 70-110 points in Phocus. That even depressed me more, so I comforted myself by saying ok, for Euro 10,000 I get a rectangular image, which I like, and a 1.1 crop factor. But I was really pissed.
It took several hours of examination to come to the following conclusions:
1. Noise. As Paul the Great Moderator said, the noise is probably higher on the CFV39, but, and this is a big but, the noise is higher but only noticeable if you blowout the shadows by a lot, and view in actual pixels, and in some cases it is very faintly visible on certain well lit surfaces.
In macro photography I have confirmed that at iso 50 the noise is higher on the CV39 than that on the CFV16, and that is one sure benefit of the bigger pixel. However, given that the CFV39 has smaller and more pixels, you tend to look at things on the CFV39, more enlarged, and this by itself will make the appearance of noise more visible.
2. Color. While the immediate visual impact of the CFV16 was more pleasing, I found that the color accuracy of the CFV39 is much more. The colors of the CFV16 were more saturated, and a bit more contrast and of course more yellow. Was this because of the new filter in the CFV39? Was it that Phocus added these automatically? Or was it the internal processing of each of the digital backs? I don’t know the answer to any of them.
3. Sharpness. After some examination I found that at actual pixel the CFV16’s sharpness was a little on the harsh side, and the CFV39’s was slightly softer. By adding the maximum level of sharpening to both, in Phocus, I found that the CFV16 accepted that intensity less, perhaps, indication that sharpening was added to the pictures from the CFV16 either in-camera or in-Phocus. Unfortunately this cannot be detected in small jpegs.
4. Chromatic Aberration. This is something that amazed me the most. I found that Hasselblad managed to reduce or control CA in the CFV39 by about 50% when compared to the CFV16. The technical reasons for this is unknown to me, but I love it. Lens used was the 40IF. Unfortunately this cannot be detected in the small jpegs.
5. Pixel Quality. This I have established for sure. For me pixel quality is measured in terms of clarity and cleanliness when viewed in the actual pixel level. The CFV39 pixel quality is at least equal to the CFV16. This has been examined in macro photography. So pixel size has no effect on pixel quality from my perspective.
6. Dynamic & Tonal Range. The dynamic range of the two CFVs appeared to be the same given that the CFV16 is more saturated and appears to have more contrast. As far as the tonal range is concerned, well, there is nothing in any of my images that clearly indicated any differences, one could assume that the much higher mpx will provide better tonal gradation.
There is one additional difference in the images other than saturation, color and contrast which I have found. The CFV39 and at the same shutter speed was 70% or 80% of the time, about 0.30 – 0.50 brighter or more exposed, which is great because it brings it closer to the light metering of my 205FCC. Why is this, I have no idea, further testing will be required, but it may be because of the new sensor filter is better at transmitting light. Blowupster may have been right when he said in his post “Does it mean that the real iso setting of the 50 asa mode is 80 asa with the CFV-39”
Conclusion
For those with the CFV16, I would like to say that you have a fantastic photographic equipment, the differences between the two CFVs is visually not that big, but technically very big. For me this difference is substantial, but then again, I am that kind of a person, totally irrational, and I love it.
On the computer screen, the image of the CFV16 viewed at normal size, looks a little bit out of focus, when compared to the CFV39, and that is the bottom line with that additional 23mpx, this is only from my point of view. For printing it may apply differently. The iso 800 is important, but most important is the 1.1 crop factor. That 40IF is now able to stand up and stretch it’s legs. I got out of that feeling of claustrophobia. Boy that feels good. Waiting for a Zeiss 28 f2.8.
I hope nobody fell asleep. I could post some pictures but they will not show much unless they are raw. If no one will blast me out of the water for the quality of the jpeg’s I will post the images in my evaluation.
I hope others will confirm, comment on, or disagree with my observations so far. The punch line is that the science of the bigger pixel has less of an effect on the final image than was the case before current advances in software, firmware and hardware designs.
Best regards
Saad
Hi everyone, I hope you are prepared for a semi detailed observation of what we loose or not loose with the smaller pixel of the CFV39 as apposed to the fat pixel of the CFVI16.
All the comparison photos were taken Hand held. I used my 205TCC on this endeavor with my 40IF and CF180. Apertures were always at f5.6, I focused, put the mirror up, and gently squeezed the trigger. Shutter speed was between 1/750 and 1/1000 and in manual mode. Almost all the pictures were taken at iso 100.
The pictures were taken in Jeddah, Saudi Arabia. Outside temperature was 38C, and it was mid-afternoon, not such good lighting. Really hot, but the process was so exciting that I didn’t care at all, but I did go back in the car every 10 minutes to cool off.
My expectation was that the pictures from the CFV39 would be clearly and undeniably superior to the CFV16. After downloading the pictures my first impression was that the CFV16 was more appealing in terms of saturation, contrast and dynamic range. This was depressing. In normal viewing the CFV39 was sharper and much more appealing in that respect. In actual pixel the CFV16 was sharper with the 40IF (but strangely enough it was the same with the 180CF, needs more testing), to get the CFV39 to the same level of sharpness I had to add between 70-110 points in Phocus. That even depressed me more, so I comforted myself by saying ok, for Euro 10,000 I get a rectangular image, which I like, and a 1.1 crop factor. But I was really pissed.
It took several hours of examination to come to the following conclusions:
1. Noise. As Paul the Great Moderator said, the noise is probably higher on the CFV39, but, and this is a big but, the noise is higher but only noticeable if you blowout the shadows by a lot, and view in actual pixels, and in some cases it is very faintly visible on certain well lit surfaces.
In macro photography I have confirmed that at iso 50 the noise is higher on the CV39 than that on the CFV16, and that is one sure benefit of the bigger pixel. However, given that the CFV39 has smaller and more pixels, you tend to look at things on the CFV39, more enlarged, and this by itself will make the appearance of noise more visible.
2. Color. While the immediate visual impact of the CFV16 was more pleasing, I found that the color accuracy of the CFV39 is much more. The colors of the CFV16 were more saturated, and a bit more contrast and of course more yellow. Was this because of the new filter in the CFV39? Was it that Phocus added these automatically? Or was it the internal processing of each of the digital backs? I don’t know the answer to any of them.
3. Sharpness. After some examination I found that at actual pixel the CFV16’s sharpness was a little on the harsh side, and the CFV39’s was slightly softer. By adding the maximum level of sharpening to both, in Phocus, I found that the CFV16 accepted that intensity less, perhaps, indication that sharpening was added to the pictures from the CFV16 either in-camera or in-Phocus. Unfortunately this cannot be detected in small jpegs.
4. Chromatic Aberration. This is something that amazed me the most. I found that Hasselblad managed to reduce or control CA in the CFV39 by about 50% when compared to the CFV16. The technical reasons for this is unknown to me, but I love it. Lens used was the 40IF. Unfortunately this cannot be detected in the small jpegs.
5. Pixel Quality. This I have established for sure. For me pixel quality is measured in terms of clarity and cleanliness when viewed in the actual pixel level. The CFV39 pixel quality is at least equal to the CFV16. This has been examined in macro photography. So pixel size has no effect on pixel quality from my perspective.
6. Dynamic & Tonal Range. The dynamic range of the two CFVs appeared to be the same given that the CFV16 is more saturated and appears to have more contrast. As far as the tonal range is concerned, well, there is nothing in any of my images that clearly indicated any differences, one could assume that the much higher mpx will provide better tonal gradation.
There is one additional difference in the images other than saturation, color and contrast which I have found. The CFV39 and at the same shutter speed was 70% or 80% of the time, about 0.30 – 0.50 brighter or more exposed, which is great because it brings it closer to the light metering of my 205FCC. Why is this, I have no idea, further testing will be required, but it may be because of the new sensor filter is better at transmitting light. Blowupster may have been right when he said in his post “Does it mean that the real iso setting of the 50 asa mode is 80 asa with the CFV-39”
Conclusion
For those with the CFV16, I would like to say that you have a fantastic photographic equipment, the differences between the two CFVs is visually not that big, but technically very big. For me this difference is substantial, but then again, I am that kind of a person, totally irrational, and I love it.
On the computer screen, the image of the CFV16 viewed at normal size, looks a little bit out of focus, when compared to the CFV39, and that is the bottom line with that additional 23mpx, this is only from my point of view. For printing it may apply differently. The iso 800 is important, but most important is the 1.1 crop factor. That 40IF is now able to stand up and stretch it’s legs. I got out of that feeling of claustrophobia. Boy that feels good. Waiting for a Zeiss 28 f2.8.
I hope nobody fell asleep. I could post some pictures but they will not show much unless they are raw. If no one will blast me out of the water for the quality of the jpeg’s I will post the images in my evaluation.
I hope others will confirm, comment on, or disagree with my observations so far. The punch line is that the science of the bigger pixel has less of an effect on the final image than was the case before current advances in software, firmware and hardware designs.
Best regards
Saad
The filter on the CFV-II and CFV-39 are the same type and should not contribute to any color differences. The original CFV back had a different filter.
I'd be interested in seeing the difference in CA between the two backs since that is a function of the lens not the back. If the same lens was used it shouldn't be a factor.
Sensor bloom or purple fringing may be different since that's a function of the sensor performance.
Marc
I'd be interested in seeing the difference in CA between the two backs since that is a function of the lens not the back. If the same lens was used it shouldn't be a factor.
Sensor bloom or purple fringing may be different since that's a function of the sensor performance.
Marc
bradleygibson
Member
Thank you, Saad for the helpful observations! Thank you for taking the time to do the write-up.
The color differences could be coming from anywhere--the sensors may have different sensel filters, native color space, analog characteristics and/or ADC's. Plus, Phocus may have a different software path depending on the file being developed, so yes it is very difficult to say. It's even likely that ALL of these things are different, thus making it a minor miracle that the colors are even close!
Still, I have found that the 'default' color provided by manufacturer's software applications is usually excellent. That being said, those colors are indeed arbitrary--they are someone's (hopefully expert) opinion of what 'good' color is. In many applications, there is a way to tweak the 'default' color setting to save time and reduce repetition (but I haven't yet played with Phocus enough).
One thing to consider is that when viewing a CFV-39 at 1:1, on, say, a 2MP monitor, you are looking at ~5% of the image (2MP out of 39). When viewing a CFV-16 at 1:1 on the same 2MP monitor, you are looking at ~ 12% of the image (2MP out of 16). Or put another way, you are effectively using a higher power loupe on a higher resolution sensor. You will see the flaws much more readily under the increased magnification. It is interesting to note that despite this, your overall conclusion is that the CFV's individual pixel quality holds its own against the fat pixels of the CFV-16. Perhaps newer techniques are being used in the CFV-39 to partially offset the smaller pixels (and lower SNR).
In actuality the sensors are different sizes, so this isn't the whole story--taking into account the CFV-16's smaller physical size, the CFV-39 at 1:1 is showing about 7% of the image vs. the CFV-16's 12%.
Bottom line is that 1:1 on a CFV-39 is at least 70% more effective magnification than showing 1:1 on a CFV-16, which is quite a bit! Don't be too disappointed if you see more flaws, more lens abberations, etc., because you are looking 70% harder! (The fact that you saw only slightly more noise and less CA is quite remarkable, actually.)
If you have raws you would like to share, please PM me--I'll be happy to have you e-mail them to me, and I will make them available to folks.
I just bought a 50 FE and a 150 FE, and am definitely chomping at the bit to receive my back...
Hopefully more of us can join your ranks soon, Saad!
Best regards,
-Brad
The color differences could be coming from anywhere--the sensors may have different sensel filters, native color space, analog characteristics and/or ADC's. Plus, Phocus may have a different software path depending on the file being developed, so yes it is very difficult to say. It's even likely that ALL of these things are different, thus making it a minor miracle that the colors are even close!
Still, I have found that the 'default' color provided by manufacturer's software applications is usually excellent. That being said, those colors are indeed arbitrary--they are someone's (hopefully expert) opinion of what 'good' color is. In many applications, there is a way to tweak the 'default' color setting to save time and reduce repetition (but I haven't yet played with Phocus enough).
One thing to consider is that when viewing a CFV-39 at 1:1, on, say, a 2MP monitor, you are looking at ~5% of the image (2MP out of 39). When viewing a CFV-16 at 1:1 on the same 2MP monitor, you are looking at ~ 12% of the image (2MP out of 16). Or put another way, you are effectively using a higher power loupe on a higher resolution sensor. You will see the flaws much more readily under the increased magnification. It is interesting to note that despite this, your overall conclusion is that the CFV's individual pixel quality holds its own against the fat pixels of the CFV-16. Perhaps newer techniques are being used in the CFV-39 to partially offset the smaller pixels (and lower SNR).
In actuality the sensors are different sizes, so this isn't the whole story--taking into account the CFV-16's smaller physical size, the CFV-39 at 1:1 is showing about 7% of the image vs. the CFV-16's 12%.
Bottom line is that 1:1 on a CFV-39 is at least 70% more effective magnification than showing 1:1 on a CFV-16, which is quite a bit! Don't be too disappointed if you see more flaws, more lens abberations, etc., because you are looking 70% harder!
(The fact that you saw only slightly more noise and less CA is quite remarkable, actually.)If you have raws you would like to share, please PM me--I'll be happy to have you e-mail them to me, and I will make them available to folks.
I just bought a 50 FE and a 150 FE, and am definitely chomping at the bit to receive my back...
Hopefully more of us can join your ranks soon, Saad!Best regards,
-Brad
Thank you very much Brad, I agree with everything you say, and will be delighted to send some raw files if you promise not to judge me on my non-artistic photographic capabilities. Please tell me how and where to send those files.
Without a doubt the colors of the CFV39 is much more true than those of my CFV16 (first version).
On the CA, I was looking at actual pixel size and was pleasantly surprised. This is the file I would really like you see.
And thank you for your response.
After a few days with the CFV39, I am beginning to really like it.
Regards
Saad
Without a doubt the colors of the CFV39 is much more true than those of my CFV16 (first version).
On the CA, I was looking at actual pixel size and was pleasantly surprised. This is the file I would really like you see.
And thank you for your response.
After a few days with the CFV39, I am beginning to really like it.
Regards
Saad
Hi Brad & Marc,
I have attached two actual pixel crops to demonstrate the CA or color fringing, or maybe its moire. Please look at the stainless steel bars, letters and the writing SAM. The big picture is of course the CFV39.
I have attached two actual pixel crops to demonstrate the CA or color fringing, or maybe its moire. Please look at the stainless steel bars, letters and the writing SAM. The big picture is of course the CFV39.
- Hasselblad - Hasselblad 503CWD
- ƒ/2
- 30.0073260073 sec
- ISO 100
Attachments
These are too small ... I can barely see what you are talking about.
Also, try processing the files as DNGs in Lightroom or Adobe Camera RAW and see if you get the same artifacts.
That's what I'm curious about Saad. Strangely, when I processed one of your other shots I got artifacts in the side area of the speed boat using Phocus, but it wasn't there using DNGs in Adobe Camera RAW ... which didn't make any sense to me.
bradleygibson
Member
The only things I'm seeing in these shots that might be mistaken for CA are demosaicing artifacts--a slight reddish or magenta/green mottling of color in some areas of the image. Is that what you're referring to?
If so, it's the result of only sampling one color per sensel in the sensor combined with the algorithm used to reconstruct the (full-color) pixel. Hasselblad and Adobe are almost certainly using different techniques to reconstruct the information.
With Bayer demosaicing algorithms there is often a tradeoff between image detail and color quality (not to mention speed). Why is this so? Because at any given sensel site, only one color is recorded. The software algorithm must look at a number of neigboring sensel sites and interpolate the other two colors for each pixel it wants to create. Unfortunately, neighboring sites saw the scene ever so slightly differently, so the process cannot be perfect.
Raw demosaicers will tend to improve with age (ie. new software releases), so be sure to keep those raw files around, even after you've made TIFFs.
-Brad
If so, it's the result of only sampling one color per sensel in the sensor combined with the algorithm used to reconstruct the (full-color) pixel. Hasselblad and Adobe are almost certainly using different techniques to reconstruct the information.
With Bayer demosaicing algorithms there is often a tradeoff between image detail and color quality (not to mention speed). Why is this so? Because at any given sensel site, only one color is recorded. The software algorithm must look at a number of neigboring sensel sites and interpolate the other two colors for each pixel it wants to create. Unfortunately, neighboring sites saw the scene ever so slightly differently, so the process cannot be perfect.
Raw demosaicers will tend to improve with age (ie. new software releases), so be sure to keep those raw files around, even after you've made TIFFs.
-Brad
bradleygibson
Member
Raw files!!
At last, a CFV-39 raw file!
Saad will have to chime in with the particulars of camera, lens and exposure settings, but he has kindly agreed to share this with everyone who has been looking for a raw file from this back. The CFV-16 file is provided as comparison of the same scene.
CFV-39 raw file (.fff)
CFV-16 raw file (.fff)
If you don't have Phocus or don't want to install it, I have converted the above two files to .dng:
CFV-39 raw file (as .dng)
CFV-16 raw file (as .dng)
Thank you in advance for being respectful of my webspace bandwidth (please do not share these links to the world), and thank you to Saad for making these files available.
Best regards,
-Brad
P.S. At least for me, Internet Explorer is renaming the above files to .tiff, even though they are stored as .fff and .dng. I assume you can just download and rename the file(s) of interest, but be aware this appears to be an issue with IE. Safari (and presumably Firefox) works as expected, and does not rename these files to .tiff.
At last, a CFV-39 raw file!
Saad will have to chime in with the particulars of camera, lens and exposure settings, but he has kindly agreed to share this with everyone who has been looking for a raw file from this back. The CFV-16 file is provided as comparison of the same scene.
CFV-39 raw file (.fff)
CFV-16 raw file (.fff)
If you don't have Phocus or don't want to install it, I have converted the above two files to .dng:
CFV-39 raw file (as .dng)
CFV-16 raw file (as .dng)
Thank you in advance for being respectful of my webspace bandwidth (please do not share these links to the world), and thank you to Saad for making these files available.
Best regards,
-Brad
P.S. At least for me, Internet Explorer is renaming the above files to .tiff, even though they are stored as .fff and .dng. I assume you can just download and rename the file(s) of interest, but be aware this appears to be an issue with IE. Safari (and presumably Firefox) works as expected, and does not rename these files to .tiff.
bradleygibson
Member
On a Mac, you CTRL clicked a link and selected Save As or Save Link As? (My Mac is down right now so I can't recall the precise wording...) Doing so will force the browser to save the file and you should be good to go.
Hi Dear Brad, you are so good.
The camera used was 205TCC.
Lens 40IF
shutter speed was between 750/1 and 1000/1.
Settings for iso was 100.
Aperture f5.6.
Hand held
Please everyone these pictures were taken in harsh sunlight ONLY for comparison purposes. These particular ones were sent to Brad for his evaluation of whether what we see is CA related or moire or something else.
The camera used was 205TCC.
Lens 40IF
shutter speed was between 750/1 and 1000/1.
Settings for iso was 100.
Aperture f5.6.
Hand held
Please everyone these pictures were taken in harsh sunlight ONLY for comparison purposes. These particular ones were sent to Brad for his evaluation of whether what we see is CA related or moire or something else.
- Status