Home made camera designs?

Régine Weinberg · December 29th, 2004, 01:51 PM

Dear wayne
well only 2!!! days to New year
had a good red wine 2000
but not drunken at all
moving CCD|s and or prism|s Im not
on this path, well sometimes dreaming,.
but the shifting idea, and second Steve that are the ways that can be real....maybe
dreming is another idea, as waiting, waiting for next five years, allways waiting and time passes by

Wayne Morellini · December 31st, 2004, 10:05 AM

Yes. I thought I heard you mention some simular dream, but it's often hard to tell what your point is your piont is, if it's a dream. Most of these things canbe done, just really difficult or expensive.

Frank Monroe · February 8th, 2005, 07:02 PM

Hey Everyone,
I'm an indie film marker from Philadelphia. I've been following your posts now for some time, but didn't really have too much input add to the mix. I was wondering if anyone could tell me if an Idea I had is do-able...

I want to shoot my next project in 24p HD with a Home Grown camera on a budget of $2,500 to spend on parts. Is that possible, and what exactly would I need to make to workable (to a hard drive of course)?

I would prefer something that was capable of doing Higher speed shooting....maybe as high as 100fps at full res, and run off of battery power.
720p would be fine, but I'd love to make 1080p work.

Thanks,
Frank

*If not for 2,500...what would something like that cost?

Dan Diaconu · February 8th, 2005, 08:47 PM

Hi Frank,
I had in mind for a while something to deliver 90 to 180 fps full rez, but I did not want to start another.......
I do not think for the budget you have I could do it though. For something like 7-10K I see it possible (with a "film look")
I think I have a good theoretical/practical solution to accomplish it! .....I might get to it ...some day......

Frank Monroe · February 8th, 2005, 09:35 PM

Hey Dan,
Thanks for the info. I'm only hoping there is a solution for 100fps, but as long as I can get 1080p at 24fps (max) that'd be fine to start. I mean...Kreiner has seemed to find a way to deliver this res on his Kinetta camera, but who knows what it will cost, or if it will ever be out.

This is the basic idea I'd like:
-1080p 24fps (hopefully capable of being faster)
-Large Internal Hard Disk Capable of dumbing captured image sequences off to a firewire drive after the fact.
-Simple Eye Piece to see what is being captured (review/playback not needed)
-Easy capture starting
-Maybe a secondard capture res eventually so you can start an offline edit without having to deal with the bigger pictures.
-Priced to build at around $2,500.

Anyone with ideas on how to make this would be awesome; however, I have no budget for experimentation.

Thanks Dan, and everyone else on the blog.

Dan Diaconu · February 8th, 2005, 11:46 PM

You are welcome and
If this is of any more help....

http://www.compumodules.com/image-processing/hitachi-hv-d30.shtml

http://www.chori-america.com/Chori_10/body_color.HTM#3CCDcolor
.......... go for it.
Those cameras could be looking at this GG:
http://dandiaconu.com/gallery/FIRST-PICTURES/IMGA0144
to give you this:

http://dandiaconu.com/gallery/ALL-CLIP-TESTS/IMGA0287

good luck achieving what you want.

Frank Monroe · February 9th, 2005, 12:41 AM

There has to be a way to get a good CMOS chip to liked up with a hard drive. I mean Kinetta managed to do it. Are there any good camera heads like those that shoot HD images, and have a USB or firewire interface? I mean in theory could I just link that up to a MiniMac or something and house it inside of a camera body?

Dan Diaconu · February 9th, 2005, 12:55 AM

Sorry Frank, not my field.

Rob Lohman · February 9th, 2005, 07:47 AM

Frank: this is exactly what this thread and the other thread
(4:4:4 uncompressed camera) are for. Please read through those
first.

To give you a short version: yes this is possible (at least part of
your list). It will cost more probably. It is not something you can
easily build yourself unless you have massive knowledge of
computers and programming (which it looks like you don't). Two
(and half) teams are currently working on this very thing and at
least one has plans to sell units (which I think will cost much more
than $2500).

In the end it isn't a simple thing. If it was simple everybody would
be building and selling this things, no wouldn't they?

If you can build such things yourself read through the threads and
familiarize yourself with what has been done and is being done.

If not you'll have to wait till someone starts selling (which I doubt
will be before the summer, or perhaps even 2006) and hope you
can afford one of these systems.

Do keep in mind that this is cutting edge technology and requires
time, patience and money.

Wayne Morellini · February 9th, 2005, 09:53 AM

<<<-- Originally posted by Frank Monroe : There has to be a way to get a good CMOS chip to liked up with a hard drive. I mean Kinetta managed to do it. Are there any good camera heads like those that shoot HD images, and have a USB or firewire interface? I mean in theory could I just link that up to a MiniMac or something and house it inside of a camera body? -->>>

Mmmm.. I had the same idea about the Mini Mac. Now the problem, it has very slow network (need Gigbe) and the firewire is only 400. So it is speed restricted and may present problems on any camera that is not frame buffered on the camera (and pixel packed). If it is not frame buffered many cameras will try to feed data out faster than the shutter speed, which easily maxes out FW400, and even 800, and Gige. So if you want to do any higher I suggest you contact Sumix about there buffered, lossless comrpessed (and I think pixel packed) Altasens camera (and whatever else they have). that shoudl get the max speed at a cheap price.

There was somebody in the "Drake" camera thread saying they had a hi-speed cheap.

Most of he stuff is in those threads, but the Digital Cinema Technical thread is the only general thread for different designs here. Some of the more knowledgeable members roam there.

I forgot to mention, the xbox next is rumoured to be comming out with a multi cored Power PC (either 21Ghz combined power or 10.5Ghz, I don't know which). So you can probably expect a simular chip to turn up in the Mac line. I would not be surprised if at least a single cored 3.5Mhz turned up in a improved Mini-mac type computer this year, but I also wouldn't be too surpised if Applle didn't do it.

I have just go more details, and it is all over the place, but I think the 3 core chip is probably likely:

http://www.theregister.co.uk/2004/02..._powerpc_chip/
http://www.theregister.co.uk/2004/02...o_sport_three/
http://forums.xbox-scene.com/index.php?showtopic=231928
http://news.spong.com/detail/news.asp?prid=6994

I have yet to read these all (too late).

There might be a low powered G5 noebook next quater, I beleive that this sort of tech will form the basis of the xbox2 processor. So we might see improvements after that.

Thanks

Wayne.

Frank Monroe · February 9th, 2005, 06:52 PM

Does anyone have a price for a SI-3170-CL from Silicon Images? Any ideas for that it might cost?

Thanks

Rob Lohman · February 10th, 2005, 05:18 AM

Frank: just contact SI, they will be happy to send you a quote on
the camera and other stuff you may need (cameralink board etc.)

Frank Monroe · February 13th, 2005, 01:44 PM

Ok say I am shooting with a board camera that captures 1280 x 1024, and I want my final image to be widescreen....which is the better solution. Getting an anamorphic lens (which I'm sure it'd have to be pretty compressed to turn that res into a widescreen shot) or simply cropping the images to 1280 x 720?

Thanks.

Oh...and what exactly is the problem with the bayer color pattern... there is a lot of talk on here about how "crappy" it is, but nobody ever seems to say why.

One other thing...is it better to use a global shutter at 24p, or use a rolling shutter at 48fps then drop off every other frame to avoid the artifact? What is the artifact? I only ask cause to seems to cost so much more to work in 48fps with the drop off then it is to work in 24p global.

Thanks again.

Rob Lohman · February 14th, 2005, 07:29 AM

The jury is still out on the rolling/global shutter issues and higher
capture rates etc.

Bayer isn't crappy at all, it is just lower resolution since you are
using one chip, and not three. A single chip (unless you have a
foveon chip) is monochrome and has a color filter placed on the
chip to let just one filter through for a pixel. This produces a
bayer pattern where:

- the first line contains green and red pixels
- the second line contains blue and green pixels

This gives the famous bayer pattern:

GR
BG

So of the full resolution 50% of the pixels contain the green
channel, 25% for red and 25% for blue. In a 3 chip camera (of
the same resolution chips) there are twice as much green pixels
and four times as much red and blue pixels.

However, with good algorithms the resolution loss (compared to
three chips) isn't that much. So you can argue whether the
resulting signal is truly HD or not.

Another benefit is 10 or 12 bit processing, which is in my opinion
a far nicer thinger to have, but that may be just me....

In regards to an anamorphic attachtment, that might be possible,
but it also gives you 304 lines of more data to work and process
with. This results in 389.120 extra bytes per frame or 9.338.880
(roughly 9 MB) bytes at 24 frames per second. I'd rather don't
have it to have a better chance of getting it to the harddisk in the
first place.

Can always add an anamorphic attachment option later....

BTW, Obin is working on a 1920 x 1080 system at this moment

Wayne Morellini · February 14th, 2005, 08:12 AM

Every bodies going to have a different opinion on this one, but the reason you do anamorphic is to increase the resolution (maybe collect more light, but I don't know if there is an adaptor that can actually achieve that) because the resolution is inadequate (such as on normal DV going to cinema).

Anamorphic is going to be a hassle, the extra bandwidth of using the whole frame will take down your frame rate used to eliminate rolling shutter, and shutter.

The bayer has the problem that it samples red and blue once, and green twice, every block of four pixels, and it guesses these colours in the other pixels it doesn't sample. To smooth this out (estimate the missing colour) they use processing routines, the better ones use more processing power. This stuffs things around a bit, requiring a better processor to do this, and maybe using the 3D GPU to compensate (in particular for preview). At the same time, it reduces the amount of data compared to three chip by three (less disks) and the higher resolution will hide some of the bayerness of the image. You can also use a wider aperture with a single chip (though a microlense can also restricts this). I think this is why Cinema cameras from the major companies (not video ones like Sony) are using single chip bayers. In the end a chip like the Foveon will eventually (hopefully) solve this. But at the same time the image of a camera is defined by sensitivity and range (how little it can accurately detect, and how many photons a pixel can take before it saturates) and I don't know if there is limits on these things in the Foveon design compared to normal designs, but I am yet to see Foveon produce a high speed design, and I have noticed less performance in comparison in a review. I'm just saying don't expect it to be better than a 3CCD, but better than bayer.

Another advantage of three chip is manufacturers can easily use pixel shift. With this we can get 4 to 9 times max resolution increase. In single chip you have to use a very complicated full spectrum filter pattern (then I don't know how good it is going to be). It would just be hard enough to do a complimentary filter pattern (I haven't seen anybody even offer one on a box camera yet). I have been working on ideas to do single chip pixel shift here, and might consider it if I get around to the commercial lens adaptor. One of the problems in this insidious industry, is that 150MB's links off a chip pushes the price way up compared to two mainboard GIGE ports (200MB's) and other higher speed standard off chip busses :( It is complicated and you can only get what you can get, sort of thing.

The rolling shutter artifact is where the image moves while the image is being scanned out, what happens is that the image now slants away from the direction of movement. Going 48th of a second alleviates the problems, but it is not the best, faster will be better. Dropping every second frame is a no brainier in software, and emulates a 48th a second shutter at 24fps (used in the movie industry). Global shutter at 24fps, 24th's shutter. The Global shutter of the IBIS sacrifices some of the light gathering power (presumably reducing the shutter time, by how much, I don't know). But unless that time is reduced to something like 48th of a second, you are going to get a lot of blur.

Now what people don't consider is what happens when you want to use high shutter speed, most of he chips do not have buffering to read out at high speeds and then send accross the slow interface at 24fps. There are, and will be cameras with memory buffering on camera to do this. Otherwise you need a mechanical shutter (with it's own problems, maybe even rolling shutter again) or high speed interface and chip (more money). There are cameras with on

These cameras will not be a perfect solutions, so it will be interesting to see how they compare to commercial cameras. In cmos the only near ideal, cheap enough, solution is the Altasens in three chip using a memory buffer and cinema quality compression, I don't think any has been planned for us yet.

There is a Altasens single chip (I think buffered and comrpessed) and a pixel shifted (4x I think) three chip IBIS5a (much better than the bayer one) from Sumix, so we will see (wait for specs).

Wayne.

February 8th, 2005, 07:02 PM	#303
Frank Monroe New Boot Join Date: Feb 2005 Location: Baltimore, Maryland Posts: 9	How to go about this, and cost? Hey Everyone, I'm an indie film marker from Philadelphia. I've been following your posts now for some time, but didn't really have too much input add to the mix. I was wondering if anyone could tell me if an Idea I had is do-able... I want to shoot my next project in 24p HD with a Home Grown camera on a budget of $2,500 to spend on parts. Is that possible, and what exactly would I need to make to workable (to a hard drive of course)? I would prefer something that was capable of doing Higher speed shooting....maybe as high as 100fps at full res, and run off of battery power. 720p would be fine, but I'd love to make 1080p work. Thanks, Frank *If not for 2,500...what would something like that cost? __________________ ~Frank

February 8th, 2005, 09:35 PM	#305
Frank Monroe New Boot Join Date: Feb 2005 Location: Baltimore, Maryland Posts: 9	Thanks Dan Hey Dan, Thanks for the info. I'm only hoping there is a solution for 100fps, but as long as I can get 1080p at 24fps (max) that'd be fine to start. I mean...Kreiner has seemed to find a way to deliver this res on his Kinetta camera, but who knows what it will cost, or if it will ever be out. This is the basic idea I'd like: -1080p 24fps (hopefully capable of being faster) -Large Internal Hard Disk Capable of dumbing captured image sequences off to a firewire drive after the fact. -Simple Eye Piece to see what is being captured (review/playback not needed) -Easy capture starting -Maybe a secondard capture res eventually so you can start an offline edit without having to deal with the bigger pictures. -Priced to build at around $2,500. Anyone with ideas on how to make this would be awesome; however, I have no budget for experimentation. Thanks Dan, and everyone else on the blog. __________________ ~Frank

February 9th, 2005, 12:41 AM	#307
Frank Monroe New Boot Join Date: Feb 2005 Location: Baltimore, Maryland Posts: 9	That's not exaclty what i mean There has to be a way to get a good CMOS chip to liked up with a hard drive. I mean Kinetta managed to do it. Are there any good camera heads like those that shoot HD images, and have a USB or firewire interface? I mean in theory could I just link that up to a MiniMac or something and house it inside of a camera body? __________________ ~Frank

February 9th, 2005, 07:47 AM	#309
Rob Lohman RED Code Chef Join Date: Oct 2001 Location: Holland Posts: 12,514	Frank: this is exactly what this thread and the other thread (4:4:4 uncompressed camera) are for. Please read through those first. To give you a short version: yes this is possible (at least part of your list). It will cost more probably. It is not something you can easily build yourself unless you have massive knowledge of computers and programming (which it looks like you don't). Two (and half) teams are currently working on this very thing and at least one has plans to sell units (which I think will cost much more than $2500). In the end it isn't a simple thing. If it was simple everybody would be building and selling this things, no wouldn't they? If you can build such things yourself read through the threads and familiarize yourself with what has been done and is being done. If not you'll have to wait till someone starts selling (which I doubt will be before the summer, or perhaps even 2006) and hope you can afford one of these systems. Do keep in mind that this is cutting edge technology and requires time, patience and money. __________________ Rob Lohman, visuar@iname.com DV Info Wrangler & RED Code Chef Join the DV Challenge \| Lady X Search DVinfo.net for quick answers \| Buy from the best: DVinfo.net sponsors

February 9th, 2005, 09:53 AM	#310
Wayne Morellini Inner Circle Join Date: May 2003 Location: Australia Posts: 2,762	Re: That's not exaclty what i mean <<<-- Originally posted by Frank Monroe : There has to be a way to get a good CMOS chip to liked up with a hard drive. I mean Kinetta managed to do it. Are there any good camera heads like those that shoot HD images, and have a USB or firewire interface? I mean in theory could I just link that up to a MiniMac or something and house it inside of a camera body? -->>> Mmmm.. I had the same idea about the Mini Mac. Now the problem, it has very slow network (need Gigbe) and the firewire is only 400. So it is speed restricted and may present problems on any camera that is not frame buffered on the camera (and pixel packed). If it is not frame buffered many cameras will try to feed data out faster than the shutter speed, which easily maxes out FW400, and even 800, and Gige. So if you want to do any higher I suggest you contact Sumix about there buffered, lossless comrpessed (and I think pixel packed) Altasens camera (and whatever else they have). that shoudl get the max speed at a cheap price. There was somebody in the "Drake" camera thread saying they had a hi-speed cheap. Most of he stuff is in those threads, but the Digital Cinema Technical thread is the only general thread for different designs here. Some of the more knowledgeable members roam there. I forgot to mention, the xbox next is rumoured to be comming out with a multi cored Power PC (either 21Ghz combined power or 10.5Ghz, I don't know which). So you can probably expect a simular chip to turn up in the Mac line. I would not be surprised if at least a single cored 3.5Mhz turned up in a improved Mini-mac type computer this year, but I also wouldn't be too surpised if Applle didn't do it. I have just go more details, and it is all over the place, but I think the 3 core chip is probably likely: http://www.theregister.co.uk/2004/02..._powerpc_chip/ http://www.theregister.co.uk/2004/02...o_sport_three/ http://forums.xbox-scene.com/index.php?showtopic=231928 http://news.spong.com/detail/news.asp?prid=6994 I have yet to read these all (too late). There might be a low powered G5 noebook next quater, I beleive that this sort of tech will form the basis of the xbox2 processor. So we might see improvements after that. Thanks Wayne.

December 29th, 2004, 01:51 PM	#301
Régine Weinberg Major Player Join Date: Jan 2004 Location: Bordeaux, going to Bangkok, 2011 Posts: 232	Dear wayne well only 2!!! days to New year had a good red wine 2000 but not drunken at all moving CCD\|s and or prism\|s Im not on this path, well sometimes dreaming,. but the shifting idea, and second Steve that are the ways that can be real....maybe dreming is another idea, as waiting, waiting for next five years, allways waiting and time passes by

December 31st, 2004, 10:05 AM	#302
Wayne Morellini Inner Circle Join Date: May 2003 Location: Australia Posts: 2,762	Yes. I thought I heard you mention some simular dream, but it's often hard to tell what your point is your piont is, if it's a dream. Most of these things canbe done, just really difficult or expensive.

February 8th, 2005, 08:47 PM	#304
Dan Diaconu Major Player Join Date: Dec 2004 Location: Vancouver, Canada Posts: 804	Hi Frank, I had in mind for a while something to deliver 90 to 180 fps full rez, but I did not want to start another....... I do not think for the budget you have I could do it though. For something like 7-10K I see it possible (with a "film look") I think I have a good theoretical/practical solution to accomplish it! .....I might get to it ...some day......

February 8th, 2005, 11:46 PM	#306
Dan Diaconu Major Player Join Date: Dec 2004 Location: Vancouver, Canada Posts: 804	You are welcome and If this is of any more help.... http://www.compumodules.com/image-processing/hitachi-hv-d30.shtml http://www.chori-america.com/Chori_10/body_color.HTM#3CCDcolor .......... go for it. Those cameras could be looking at this GG: http://dandiaconu.com/gallery/FIRST-PICTURES/IMGA0144 to give you this: http://dandiaconu.com/gallery/ALL-CLIP-TESTS/IMGA0287 good luck achieving what you want.

February 9th, 2005, 12:55 AM	#308
Dan Diaconu Major Player Join Date: Dec 2004 Location: Vancouver, Canada Posts: 804	Sorry Frank, not my field.

February 9th, 2005, 06:52 PM	#311
Frank Monroe New Boot Join Date: Feb 2005 Location: Baltimore, Maryland Posts: 9	SI-3170-CL Does anyone have a price for a SI-3170-CL from Silicon Images? Any ideas for that it might cost? Thanks __________________ ~Frank

February 10th, 2005, 05:18 AM	#312
Rob Lohman RED Code Chef Join Date: Oct 2001 Location: Holland Posts: 12,514	Frank: just contact SI, they will be happy to send you a quote on the camera and other stuff you may need (cameralink board etc.) __________________ Rob Lohman, visuar@iname.com DV Info Wrangler & RED Code Chef Join the DV Challenge \| Lady X Search DVinfo.net for quick answers \| Buy from the best: DVinfo.net sponsors

February 13th, 2005, 01:44 PM	#313
Frank Monroe New Boot Join Date: Feb 2005 Location: Baltimore, Maryland Posts: 9	Anamorphic Versus Cropping Ok say I am shooting with a board camera that captures 1280 x 1024, and I want my final image to be widescreen....which is the better solution. Getting an anamorphic lens (which I'm sure it'd have to be pretty compressed to turn that res into a widescreen shot) or simply cropping the images to 1280 x 720? Thanks. Oh...and what exactly is the problem with the bayer color pattern... there is a lot of talk on here about how "crappy" it is, but nobody ever seems to say why. One other thing...is it better to use a global shutter at 24p, or use a rolling shutter at 48fps then drop off every other frame to avoid the artifact? What is the artifact? I only ask cause to seems to cost so much more to work in 48fps with the drop off then it is to work in 24p global. Thanks again. __________________ ~Frank

February 14th, 2005, 07:29 AM	#314
Rob Lohman RED Code Chef Join Date: Oct 2001 Location: Holland Posts: 12,514	The jury is still out on the rolling/global shutter issues and higher capture rates etc. Bayer isn't crappy at all, it is just lower resolution since you are using one chip, and not three. A single chip (unless you have a foveon chip) is monochrome and has a color filter placed on the chip to let just one filter through for a pixel. This produces a bayer pattern where: - the first line contains green and red pixels - the second line contains blue and green pixels This gives the famous bayer pattern: GR BG So of the full resolution 50% of the pixels contain the green channel, 25% for red and 25% for blue. In a 3 chip camera (of the same resolution chips) there are twice as much green pixels and four times as much red and blue pixels. However, with good algorithms the resolution loss (compared to three chips) isn't that much. So you can argue whether the resulting signal is truly HD or not. Another benefit is 10 or 12 bit processing, which is in my opinion a far nicer thinger to have, but that may be just me.... In regards to an anamorphic attachtment, that might be possible, but it also gives you 304 lines of more data to work and process with. This results in 389.120 extra bytes per frame or 9.338.880 (roughly 9 MB) bytes at 24 frames per second. I'd rather don't have it to have a better chance of getting it to the harddisk in the first place. Can always add an anamorphic attachment option later.... BTW, Obin is working on a 1920 x 1080 system at this moment __________________ Rob Lohman, visuar@iname.com DV Info Wrangler & RED Code Chef Join the DV Challenge \| Lady X Search DVinfo.net for quick answers \| Buy from the best: DVinfo.net sponsors

February 14th, 2005, 08:12 AM	#315
Wayne Morellini Inner Circle Join Date: May 2003 Location: Australia Posts: 2,762	Every bodies going to have a different opinion on this one, but the reason you do anamorphic is to increase the resolution (maybe collect more light, but I don't know if there is an adaptor that can actually achieve that) because the resolution is inadequate (such as on normal DV going to cinema). Anamorphic is going to be a hassle, the extra bandwidth of using the whole frame will take down your frame rate used to eliminate rolling shutter, and shutter. The bayer has the problem that it samples red and blue once, and green twice, every block of four pixels, and it guesses these colours in the other pixels it doesn't sample. To smooth this out (estimate the missing colour) they use processing routines, the better ones use more processing power. This stuffs things around a bit, requiring a better processor to do this, and maybe using the 3D GPU to compensate (in particular for preview). At the same time, it reduces the amount of data compared to three chip by three (less disks) and the higher resolution will hide some of the bayerness of the image. You can also use a wider aperture with a single chip (though a microlense can also restricts this). I think this is why Cinema cameras from the major companies (not video ones like Sony) are using single chip bayers. In the end a chip like the Foveon will eventually (hopefully) solve this. But at the same time the image of a camera is defined by sensitivity and range (how little it can accurately detect, and how many photons a pixel can take before it saturates) and I don't know if there is limits on these things in the Foveon design compared to normal designs, but I am yet to see Foveon produce a high speed design, and I have noticed less performance in comparison in a review. I'm just saying don't expect it to be better than a 3CCD, but better than bayer. Another advantage of three chip is manufacturers can easily use pixel shift. With this we can get 4 to 9 times max resolution increase. In single chip you have to use a very complicated full spectrum filter pattern (then I don't know how good it is going to be). It would just be hard enough to do a complimentary filter pattern (I haven't seen anybody even offer one on a box camera yet). I have been working on ideas to do single chip pixel shift here, and might consider it if I get around to the commercial lens adaptor. One of the problems in this insidious industry, is that 150MB's links off a chip pushes the price way up compared to two mainboard GIGE ports (200MB's) and other higher speed standard off chip busses :( It is complicated and you can only get what you can get, sort of thing. The rolling shutter artifact is where the image moves while the image is being scanned out, what happens is that the image now slants away from the direction of movement. Going 48th of a second alleviates the problems, but it is not the best, faster will be better. Dropping every second frame is a no brainier in software, and emulates a 48th a second shutter at 24fps (used in the movie industry). Global shutter at 24fps, 24th's shutter. The Global shutter of the IBIS sacrifices some of the light gathering power (presumably reducing the shutter time, by how much, I don't know). But unless that time is reduced to something like 48th of a second, you are going to get a lot of blur. Now what people don't consider is what happens when you want to use high shutter speed, most of he chips do not have buffering to read out at high speeds and then send accross the slow interface at 24fps. There are, and will be cameras with memory buffering on camera to do this. Otherwise you need a mechanical shutter (with it's own problems, maybe even rolling shutter again) or high speed interface and chip (more money). There are cameras with on These cameras will not be a perfect solutions, so it will be interesting to see how they compare to commercial cameras. In cmos the only near ideal, cheap enough, solution is the Altasens in three chip using a memory buffer and cinema quality compression, I don't think any has been planned for us yet. There is a Altasens single chip (I think buffered and comrpessed) and a pixel shifted (4x I think) three chip IBIS5a (much better than the bayer one) from Sumix, so we will see (wait for specs). Wayne.