4:4:4 10bit single CMOS HD project

Steve Nordhauser · September 1st, 2004, 09:19 AM

Rob L - bin/subsample mode on Altasens:
You are correct - the Altasens is the first sensor that I have seen that can do that combo mode - binning and subsampling together to downsize 1920x1080 (native sensor size) to 1280x720 using the full sensor. The 3300 only supports subsample and ROI resizing.

Altasens:
A *preliminary* spec sheet is here:
http://www.siliconimaging.com/SI1920/SI1920%20Spec.PDF
Still due around the end of October.

Obin:
Sorry we don't have a hefty price tag on the SI-1920HD, just a low price. If you want, I'll create a hefty price tag for you.
Now if you want to know the low price, it is $3995, single piece.

Jason Rodriguez · September 1st, 2004, 10:12 AM

Steve,

are you incorporating all those features in the brochure? If so, how will we access them? Will XCAP be able to access them, or will we have to send out serial commands and edit registers?

Steve Nordhauser · September 1st, 2004, 10:29 AM

Jason:
Life is never easy. We will give you access to all those features. You will be able to get most of them from the XCAP GUI (not what you want) but if you use the SDK, you need to send strings, just like the other cameras. Depending how many people want to use the SDK for independent projects, I might put a camera discount bounty on the developer who releases an open source wrapper that lets people use higher level calls.

Gary McClurg · September 1st, 2004, 10:34 AM

Steve do you have a web site so we can check out your camera or when is it do to hit the market?

Wayne Morellini · September 1st, 2004, 10:48 AM

<<<-- Originally posted by Steve Nordhauser : independent projects, I might put a camera discount bounty on the developer who releases an open source wrapper that lets people use higher level calls. -->>>

I thought that was the original objective for Rob's software? To just use camera specification files to let the software know what features are available and where they go for a new camera.

Steve Nordhauser · September 1st, 2004, 10:55 AM

Wayne:
It is not clear what on Rob's software will be open source and what is not, but I think he said the camera specific stuff may not be. It is up to him how that works.

Gary:
My sig file has my company web address with all the released information. Every once in awhile I toss out test files or preliminary information on DVInfo. I just posted the SI-1920HD prelim spec a few posts ago. Delivery is expected around the end of October. The other cameras - SI-1300, 3300 are on the web site and available now.

Jason Rodriguez · September 1st, 2004, 10:56 AM

Well I'm not sure if Rob's software is going to give us access to all those goodies listed in the brochure. Of course I know that Rob is working very hard, and there are only so many features that are really necessary for this project to work-if he tried to incorporate every feature this chip has, he might be programming for th next year before we get a good beta release, and I wouldn't want him slaving away for more obscure features. What I think would be interesting is some access to these calls so that if you did want to really tweak your camera you could.

BTW, why is XCAP so bad?? Of course it's not great, and there are better things out there, but it doesn't seem as though it's horrible, like totally unusable. Oh well, I haven't used it, so I can't really comment, now can I :)

Rob Scott · September 1st, 2004, 10:57 AM

Quote:

Wayne Morellini wrote:
I thought that was the original objective for Rob's software? To just use camera specification files to let the software know what features are available and where they go for a new camera.

That was one of the original ideas, yes. However, I'm not sure how practical it is -- XCAP (for example) tries to do this, and ends up being so flexible that it's hard to use. To be readily used for filmmaking, I think the ObscuraCam software needs to know a lot about the individual cameras that it supports, which would be very difficult to reduce to capability profiles. That's not to say it would be out of the question, but I don't see it happening any time soon.

-- EDIT --

Quote:

It is not clear what on Rob's software will be open source and what is not, but I think he said the camera specific stuff may not be. It is up to him how that works.

That's correct, Steve. My intention right now is to keep the "firmware" application closed-source.

Quote:

Well I'm not sure if Rob's software is going to give us access to all those goodies listed in the brochure.

Yeah, I'm going to have to pick and choose based on what I think is best for filmmaking. 1000 fps with a tiny window? Probably not. 120fps @ 720p? Almost certainly. Interlaced? Not sure.

Wayne Morellini · September 1st, 2004, 11:27 AM

<<<-- Originally posted by Steve Nordhauser : Wayne:
It is not clear what on Rob's software will be open source and what is not, but I think he said the camera specific stuff may not be. It is up to him how that works.
-->>>

Yes, your right, sorry my fault.

Rob, I agree with Jason, common (+good) features version first. But a profile file system in future, would make it easier for camera makers (or advanced users) to "register" their cameras for use with your sofware. I scribbled up an idea for an extensible interface and menue system the other week, that sorted out all these problems. If you like I can share it with you for the next version (though somebody is also talking to me about doing something, so maybe I'll have to check with them first).

Obin Olson · September 2nd, 2004, 10:04 PM

jason or anyone who would know..why did a variCam 35mm transfer look really really dark today when I ran it at the local cinema? would it be the projector ? or the print transfer CC is wrong? all the darks are really really DARK

Jason Rodriguez · September 2nd, 2004, 10:12 PM

There's a lot of things that could have gone wrong. Typically most transfer facilities should have some calibrated monitor systems to give you an idea of what you'll get on film.

When you say the darks are really dark, are you talking about the blacks being crushed? Or just the gamma of the picture in general is dark? Crushed blacks could have come from you trying to make the picture look contrasty on the monitor system, since monitors don't respond the same way as print film, especially Vision Premiere, which tends to have very rich blacks (to offset the very flat look of color negative). So something that looks nice with the blacks on your monitor will look crushed on the screen.

BTW, other than crushed blacks, how did everything else look? Highlight handling, colors, compression, etc.?

Obin Olson · September 2nd, 2004, 10:29 PM

compression is good..i don't see it colors are ok...one shot that was lit with nothing but clouds in the sky looked really great!

the darks are BLACK not dark like on 2 TV sets one HD one non and on the monitor in the field..or could it be the projector bulb?

I gota say that it felt like a cheap contrast was on the footage...if I think in terms of what I know it felt like a bad and really heavy contrast...the lab did a "CC" for all the footage and a "CC LOOK" for 3 shots

I will try atleast one more projector..they have a good one at ScreenGems Studio here and also JDC has one

720P is enough resolution but 1080p will be better ;)

Wayne Morellini · September 2nd, 2004, 11:18 PM

For everybody that wants to know more about sensor specs I have discovered this page:

http://www.ccd.com/ccdu.html

Silicon Imaging also has some basic tutorial pages on their site.

Juan M. M. Fiebelkorn · September 3rd, 2004, 12:21 AM

I work with film transfers every day of my life ( I do them :) or to be clearer I'm the guy who could make your footage look good or dark...) and If I'm allowed to tell anything, I would say it is very normal to see your video footage darker on 35 mm...
Around the internet is a common saying of giving cameras (video) darker blacks that what it is supposed for TV standards , and that's not right..
The standard black for NTSC is 7.5 IRE and for PAL it is 16 IRE, ( I suggest always use 16) and they work very well..
May be some of that things got messed up somewhere in the chain, don't know really...
Anyway, film material always looks darker (or with more contrast) than the video master...
Remember that your video master goes thru a standard gamma correction on the film recorder to be transfered to film, and if the video original isn't captured within TV Standards limits (From 7.5 or 16 to 235 IRE) or corrected within that limits anything can go wrong...
Also remember that anything blue will look darker on film against video most of the times...
Hope this helps..

BTW, did you use a MAC somewhere in your workflow before sending the images to the facility?

Obin Olson · September 3rd, 2004, 07:54 AM

no I shot and sent it off..NO editing at all jsut a tape

September 1st, 2004, 09:19 AM	#1561
Steve Nordhauser Silicon Imaging, Inc. Join Date: May 2004 Location: Troy, NY USA Posts: 325	Rob L - bin/subsample mode on Altasens: You are correct - the Altasens is the first sensor that I have seen that can do that combo mode - binning and subsampling together to downsize 1920x1080 (native sensor size) to 1280x720 using the full sensor. The 3300 only supports subsample and ROI resizing. Altasens: A preliminary spec sheet is here: http://www.siliconimaging.com/SI1920/SI1920%20Spec.PDF Still due around the end of October. Obin: Sorry we don't have a hefty price tag on the SI-1920HD, just a low price. If you want, I'll create a hefty price tag for you. Now if you want to know the low price, it is $3995, single piece. __________________ Silicon Imaging, Inc. We see the Light! http://www.siliconimaging.com

September 1st, 2004, 10:29 AM	#1563
Steve Nordhauser Silicon Imaging, Inc. Join Date: May 2004 Location: Troy, NY USA Posts: 325	Jason: Life is never easy. We will give you access to all those features. You will be able to get most of them from the XCAP GUI (not what you want) but if you use the SDK, you need to send strings, just like the other cameras. Depending how many people want to use the SDK for independent projects, I might put a camera discount bounty on the developer who releases an open source wrapper that lets people use higher level calls. __________________ Silicon Imaging, Inc. We see the Light! http://www.siliconimaging.com

September 1st, 2004, 10:55 AM	#1566
Steve Nordhauser Silicon Imaging, Inc. Join Date: May 2004 Location: Troy, NY USA Posts: 325	Wayne: It is not clear what on Rob's software will be open source and what is not, but I think he said the camera specific stuff may not be. It is up to him how that works. Gary: My sig file has my company web address with all the released information. Every once in awhile I toss out test files or preliminary information on DVInfo. I just posted the SI-1920HD prelim spec a few posts ago. Delivery is expected around the end of October. The other cameras - SI-1300, 3300 are on the web site and available now. __________________ Silicon Imaging, Inc. We see the Light! http://www.siliconimaging.com

September 1st, 2004, 10:12 AM	#1562
Jason Rodriguez Trustee Join Date: Mar 2003 Location: Virginia Beach, VA Posts: 1,095	Steve, are you incorporating all those features in the brochure? If so, how will we access them? Will XCAP be able to access them, or will we have to send out serial commands and edit registers?

September 1st, 2004, 10:34 AM	#1564
Gary McClurg Major Player Join Date: Jul 2004 Location: Springfield, MO, USA Posts: 389	Steve do you have a web site so we can check out your camera or when is it do to hit the market?

September 1st, 2004, 10:48 AM	#1565
Wayne Morellini Inner Circle Join Date: May 2003 Location: Australia Posts: 2,762	<<<-- Originally posted by Steve Nordhauser : independent projects, I might put a camera discount bounty on the developer who releases an open source wrapper that lets people use higher level calls. -->>> I thought that was the original objective for Rob's software? To just use camera specification files to let the software know what features are available and where they go for a new camera.

September 1st, 2004, 10:56 AM	#1567
Jason Rodriguez Trustee Join Date: Mar 2003 Location: Virginia Beach, VA Posts: 1,095	Well I'm not sure if Rob's software is going to give us access to all those goodies listed in the brochure. Of course I know that Rob is working very hard, and there are only so many features that are really necessary for this project to work-if he tried to incorporate every feature this chip has, he might be programming for th next year before we get a good beta release, and I wouldn't want him slaving away for more obscure features. What I think would be interesting is some access to these calls so that if you did want to really tweak your camera you could. BTW, why is XCAP so bad?? Of course it's not great, and there are better things out there, but it doesn't seem as though it's horrible, like totally unusable. Oh well, I haven't used it, so I can't really comment, now can I :)

September 1st, 2004, 11:27 AM	#1569
Wayne Morellini Inner Circle Join Date: May 2003 Location: Australia Posts: 2,762	<<<-- Originally posted by Steve Nordhauser : Wayne: It is not clear what on Rob's software will be open source and what is not, but I think he said the camera specific stuff may not be. It is up to him how that works. -->>> Yes, your right, sorry my fault. Rob, I agree with Jason, common (+good) features version first. But a profile file system in future, would make it easier for camera makers (or advanced users) to "register" their cameras for use with your sofware. I scribbled up an idea for an extensible interface and menue system the other week, that sorted out all these problems. If you like I can share it with you for the next version (though somebody is also talking to me about doing something, so maybe I'll have to check with them first).

September 2nd, 2004, 10:04 PM	#1570
Obin Olson Trustee Join Date: Jan 2003 Location: Wilmington NC Posts: 1,414	jason or anyone who would know..why did a variCam 35mm transfer look really really dark today when I ran it at the local cinema? would it be the projector ? or the print transfer CC is wrong? all the darks are really really DARK

September 2nd, 2004, 10:12 PM	#1571
Jason Rodriguez Trustee Join Date: Mar 2003 Location: Virginia Beach, VA Posts: 1,095	There's a lot of things that could have gone wrong. Typically most transfer facilities should have some calibrated monitor systems to give you an idea of what you'll get on film. When you say the darks are really dark, are you talking about the blacks being crushed? Or just the gamma of the picture in general is dark? Crushed blacks could have come from you trying to make the picture look contrasty on the monitor system, since monitors don't respond the same way as print film, especially Vision Premiere, which tends to have very rich blacks (to offset the very flat look of color negative). So something that looks nice with the blacks on your monitor will look crushed on the screen. BTW, other than crushed blacks, how did everything else look? Highlight handling, colors, compression, etc.?

September 2nd, 2004, 10:29 PM	#1572
Obin Olson Trustee Join Date: Jan 2003 Location: Wilmington NC Posts: 1,414	compression is good..i don't see it colors are ok...one shot that was lit with nothing but clouds in the sky looked really great! the darks are BLACK not dark like on 2 TV sets one HD one non and on the monitor in the field..or could it be the projector bulb? I gota say that it felt like a cheap contrast was on the footage...if I think in terms of what I know it felt like a bad and really heavy contrast...the lab did a "CC" for all the footage and a "CC LOOK" for 3 shots I will try atleast one more projector..they have a good one at ScreenGems Studio here and also JDC has one 720P is enough resolution but 1080p will be better ;)

September 2nd, 2004, 11:18 PM	#1573
Wayne Morellini Inner Circle Join Date: May 2003 Location: Australia Posts: 2,762	For everybody that wants to know more about sensor specs I have discovered this page: http://www.ccd.com/ccdu.html Silicon Imaging also has some basic tutorial pages on their site.

September 3rd, 2004, 12:21 AM	#1574
Juan M. M. Fiebelkorn Major Player Join Date: Jun 2004 Location: Buenos Aires , Argentina Posts: 444	I work with film transfers every day of my life ( I do them :) or to be clearer I'm the guy who could make your footage look good or dark...) and If I'm allowed to tell anything, I would say it is very normal to see your video footage darker on 35 mm... Around the internet is a common saying of giving cameras (video) darker blacks that what it is supposed for TV standards , and that's not right.. The standard black for NTSC is 7.5 IRE and for PAL it is 16 IRE, ( I suggest always use 16) and they work very well.. May be some of that things got messed up somewhere in the chain, don't know really... Anyway, film material always looks darker (or with more contrast) than the video master... Remember that your video master goes thru a standard gamma correction on the film recorder to be transfered to film, and if the video original isn't captured within TV Standards limits (From 7.5 or 16 to 235 IRE) or corrected within that limits anything can go wrong... Also remember that anything blue will look darker on film against video most of the times... Hope this helps.. BTW, did you use a MAC somewhere in your workflow before sending the images to the facility?

September 3rd, 2004, 07:54 AM	#1575
Obin Olson Trustee Join Date: Jan 2003 Location: Wilmington NC Posts: 1,414	no I shot and sent it off..NO editing at all jsut a tape