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.
50th anniversary
- Thread starter polypal
- Start date
That's not the point Q.G., each shot is separate. The Multi-shot backs work in very small movements and record each shot cumulatively for increased color resolution, and Moire' free images. The tiff files are about 230 meg from a 4 shot 39 meg. The dual and quad adapters stitch together 2 or 4 individual view camera shots into a big one. Imacon has multi-shot backs and so does Sinar (others?).
Well, if the quality of lenses is concerned, lp/mm are indeed relevant.
Its rather the file size that would be irrelevant.
We need to know how the 'resolving power', in lp/mm, of such a multishot setup stacks up against that of 'average' lenses.
If a lens has a great resolving power (in lp/mm), its colour faults simply cannot be that bad.
The colour fringes that might cause problems on 'matrixed' (!) sensors are nothing else but a difference in magnification of the differently coloured images the lens produces.
The resulting spread of light not just appears as colour fringes, but will lower resolving power rather drastically.
Which is why "lens designers have been working to reduce [etc.]". And also why a lens with perfect colour correction, like the Superachromats, shows its best on B&W film (resolving more detail than colour film).
And if a lens 'outresolves' a digital back, or such a composite image, significantly, we can be very sure that colour correction is not where that lens needs to be improved to make it suitable for digital capture.
The closer the (spatial) resolving power of lens and digital capture, the more colour correction may (!) be a problem.
But what sensor (or multi-shot composite image) even appraoches the resolving power of lenses?
Such multi-shots images however will indeed reduce the sensor produced (!) fringing problems.
Its rather the file size that would be irrelevant.
We need to know how the 'resolving power', in lp/mm, of such a multishot setup stacks up against that of 'average' lenses.
If a lens has a great resolving power (in lp/mm), its colour faults simply cannot be that bad.
The colour fringes that might cause problems on 'matrixed' (!) sensors are nothing else but a difference in magnification of the differently coloured images the lens produces.
The resulting spread of light not just appears as colour fringes, but will lower resolving power rather drastically.
Which is why "lens designers have been working to reduce [etc.]". And also why a lens with perfect colour correction, like the Superachromats, shows its best on B&W film (resolving more detail than colour film).
And if a lens 'outresolves' a digital back, or such a composite image, significantly, we can be very sure that colour correction is not where that lens needs to be improved to make it suitable for digital capture.
The closer the (spatial) resolving power of lens and digital capture, the more colour correction may (!) be a problem.
But what sensor (or multi-shot composite image) even appraoches the resolving power of lenses?
Such multi-shots images however will indeed reduce the sensor produced (!) fringing problems.
Hi Paul,
> I had over 400 Meg tiff files when I got my first 25 Mp Sinar back. It > was used in 16 shot mode. > > Most older Carl Zeiss lenses clearly reached their limit when used > this way.
Why would multi-shot be any more taxing on a lense than a single shot? Unless something physically changes between shots... My understanding of multi-shot, which could very well be outdated, is that it simply adjusts the exposure to increase the dynamic range of the resultant merged image. This is of course a different multi-shot than a single filter camera does...which requires three shots, one R, one G, one B to get a full color spectrum.
Regards,
Austin
> I had over 400 Meg tiff files when I got my first 25 Mp Sinar back. It > was used in 16 shot mode. > > Most older Carl Zeiss lenses clearly reached their limit when used > this way.
Why would multi-shot be any more taxing on a lense than a single shot? Unless something physically changes between shots... My understanding of multi-shot, which could very well be outdated, is that it simply adjusts the exposure to increase the dynamic range of the resultant merged image. This is of course a different multi-shot than a single filter camera does...which requires three shots, one R, one G, one B to get a full color spectrum.
Regards,
Austin
>
Multi shot files are NOT more taxing on a lens than single shot files. They give improved quality by avoiding using the interpolation that any bayer sensor must use to de-mosaic the RAW file. In modern multi-shot systems the chip is moved by a peizo motor in one pixel steps (for a 4 shot) so that each pixel location is s&led four times. The resulting file is the exact same size as a single shot but with the benefits of increased detail, more accurate colour and no moire (with some very rare exceptions).
For a 16 shot file the chip is moved in 1/4 pixel steps giving a file size four times that of a single shot.
Nick-T
Multi shot files are NOT more taxing on a lens than single shot files. They give improved quality by avoiding using the interpolation that any bayer sensor must use to de-mosaic the RAW file. In modern multi-shot systems the chip is moved by a peizo motor in one pixel steps (for a 4 shot) so that each pixel location is s&led four times. The resulting file is the exact same size as a single shot but with the benefits of increased detail, more accurate colour and no moire (with some very rare exceptions).
For a 16 shot file the chip is moved in 1/4 pixel steps giving a file size four times that of a single shot.
Nick-T
Hi Nick,
> Multi shot files are NOT more taxing on a lens than single shot files. > They give improved quality by avoiding using the interpolation that > any bayer sensor must use to de-mosaic the RAW file.
If you are referring to the four shots being one R, one B and two G, of course in that scenario, I agree, this is not more taxing on the lense than a single shot.
> In modern > multi-shot systems the chip is moved by a peizo motor in one pixel > steps (for a 4 shot) so that each pixel location is s&led four > times.
What good does moving in one pixel steps do? I'm missing something about this scenario.
> For a 16 shot file the chip is moved in 1/4 pixel steps giving a file > size four times that of a single shot.
This makes sense, and I do believe this IS more taxing on the lense, since it requires the lense to resolve more than if you were to not move the sensor. Is this using four shots per location, one R, one B and two G, or four shots in each location with different exposures to increase dynamic range? Why do you believe this scenario is not more taxing on the lense? It would also appear it does four s&les at each location.
There are three reasons for multi-shot as I see it. One is using a different filter for each shot (except G which gets two). Second is to increase dynamic range by varying the exposure per shot, and three is to increase the resolution. Any combination of these three could be done as well. Also, multi-shots with Bayer for scenario 2 and 3 make sense as well.
Regards,
Austin
> Multi shot files are NOT more taxing on a lens than single shot files. > They give improved quality by avoiding using the interpolation that > any bayer sensor must use to de-mosaic the RAW file.
If you are referring to the four shots being one R, one B and two G, of course in that scenario, I agree, this is not more taxing on the lense than a single shot.
> In modern > multi-shot systems the chip is moved by a peizo motor in one pixel > steps (for a 4 shot) so that each pixel location is s&led four > times.
What good does moving in one pixel steps do? I'm missing something about this scenario.
> For a 16 shot file the chip is moved in 1/4 pixel steps giving a file > size four times that of a single shot.
This makes sense, and I do believe this IS more taxing on the lense, since it requires the lense to resolve more than if you were to not move the sensor. Is this using four shots per location, one R, one B and two G, or four shots in each location with different exposures to increase dynamic range? Why do you believe this scenario is not more taxing on the lense? It would also appear it does four s&les at each location.
There are three reasons for multi-shot as I see it. One is using a different filter for each shot (except G which gets two). Second is to increase dynamic range by varying the exposure per shot, and three is to increase the resolution. Any combination of these three could be done as well. Also, multi-shots with Bayer for scenario 2 and 3 make sense as well.
Regards,
Austin
>
> For a 16 shot file the chip is moved in 1/4 pixel steps giving a > > file > size four times that of a single shot.
Austin wrote: This makes sense, and I do believe this IS more taxing on the lense, since it requires the lense to resolve more than if you were to not move the sensor
You are quite right Austin sloppy post on my part. I should have said that a FOUR shot capture is not more taxing on the lense.
Austin wrote: What good does moving in one pixel steps do? I'm missing something about this scenario.
It means that actual colour data is s&led at each pixel as opposed to a single shot bayer capture where only 1/3 of actual colours are s&led and the rest being interpolated. Hope that makes sense and apologies for my earlier sloppy post.
Nick-T
> For a 16 shot file the chip is moved in 1/4 pixel steps giving a > > file > size four times that of a single shot.
Austin wrote: This makes sense, and I do believe this IS more taxing on the lense, since it requires the lense to resolve more than if you were to not move the sensor
You are quite right Austin sloppy post on my part. I should have said that a FOUR shot capture is not more taxing on the lense.
Austin wrote: What good does moving in one pixel steps do? I'm missing something about this scenario.
It means that actual colour data is s&led at each pixel as opposed to a single shot bayer capture where only 1/3 of actual colours are s&led and the rest being interpolated. Hope that makes sense and apologies for my earlier sloppy post.
Nick-T
Hi Nick,
Thanks for the 16 shot clarification...
> Austin wrote: What good does moving in one pixel steps do? I'm missing > something about this scenario. > > It means that actual colour data is s&led at each pixel as opposed > to a single shot bayer capture where only 1/3 of actual colours are > s&led and the rest being interpolated.
I still don't understand this. Is this four shot WITH a Bayer pattern sensor, or with one color per shot (RGGB)? One color per shot doesn't make sense to move the sensor. Bayer pattern, yes, I'd understand moving three times for four shots then.
Regards,
Austin
Thanks for the 16 shot clarification...
> Austin wrote: What good does moving in one pixel steps do? I'm missing > something about this scenario. > > It means that actual colour data is s&led at each pixel as opposed > to a single shot bayer capture where only 1/3 of actual colours are > s&led and the rest being interpolated.
I still don't understand this. Is this four shot WITH a Bayer pattern sensor, or with one color per shot (RGGB)? One color per shot doesn't make sense to move the sensor. Bayer pattern, yes, I'd understand moving three times for four shots then.
Regards,
Austin
Hi Nick.
Jeeze I didn't notice that the final 8 bit tiff file size for a 4 shot was the same as single shot.
What confused me is that with CF-39 single shot the 3FR compressed file is 50MB average, while the the multishot 3FR compressed is 200MB. ????
Where is all that data going when it's converted to a tiff?
The CF-22 16 shot RAW 3FR compressed is 480MB average but produces a 264 MB tiff ????
Jeeze I didn't notice that the final 8 bit tiff file size for a 4 shot was the same as single shot.
What confused me is that with CF-39 single shot the 3FR compressed file is 50MB average, while the the multishot 3FR compressed is 200MB. ????
Where is all that data going when it's converted to a tiff?
The CF-22 16 shot RAW 3FR compressed is 480MB average but produces a 264 MB tiff ????
I was thinking the same thing Q.G. ... conceptually sort of like layers in PhotoShop. It would still be interesting to know how this really works.
I think you are right Simon. In addition, no one is going to pay this kind of money for pure marketing hype improvements if its obvious that there are no visible results. "Don't tell me, Show Me" usually prevails as the final arbitrator. Honestly, these backs are so expensive it is a powerful deterent to impulsive "gotta have the latest greatest" so prevelant in consumer digital cameras.
For ex&le, Hasselblad marketing alerted me of the new 31 meg H3D, which improved maximum ISO sensitivity from 400 to 800 and added shooting speed. But before spending a King's ransom to secure this camera for the type of work I do, I needed to see it, hold it, shoot with it ... and compare the experience with my existing H2D/22 and H3D/39.
City Lights Digital jumped through hoops to provide me with ISO 800 files, and as soon as a demo camera was available, brought it to my studio to shoot with. The camera delivered on it's promises. No hype. Very good ISO 800 files, and shooting speed was indeed quicker... so I upgraded the H2D.
I think you are right Simon. In addition, no one is going to pay this kind of money for pure marketing hype improvements if its obvious that there are no visible results. "Don't tell me, Show Me" usually prevails as the final arbitrator. Honestly, these backs are so expensive it is a powerful deterent to impulsive "gotta have the latest greatest" so prevelant in consumer digital cameras.
For ex&le, Hasselblad marketing alerted me of the new 31 meg H3D, which improved maximum ISO sensitivity from 400 to 800 and added shooting speed. But before spending a King's ransom to secure this camera for the type of work I do, I needed to see it, hold it, shoot with it ... and compare the experience with my existing H2D/22 and H3D/39.
City Lights Digital jumped through hoops to provide me with ISO 800 files, and as soon as a demo camera was available, brought it to my studio to shoot with. The camera delivered on it's promises. No hype. Very good ISO 800 files, and shooting speed was indeed quicker... so I upgraded the H2D.