3 channel 36 bit 1280 X 720 low $ camera - Viper?

[Rob Scott]

<<<--- I'd really like to get a mac with FCP HD. -->>>
<<<--- About the Codec, I don't think anybody has decided yet. I suggested that somebody could make a plugin to support -->>>

Wayne/Laurence,

If we do any compression in the camera itself (maybe in version 2.x?) it will need to be an open-source codec ... perhaps a 10- or 12-bit version of HuffYUV. (HuffYUV is 4:2:2, but otherwise lossless and provides about 2:1 compression.)

As far as the codec support goes on the "host" PC or Mac goes, I'd like to see if we can support any codec available on the machine.

I have a cousin in the industry who uses FCP, so I'm planning to work with him to make sure the software is compatible. We'll see ...

A summary of my ideas so far is available at http://www.obscuracam.com/wiki/wiki/...RobScottDesign

[David Newman]

<<<-- Originally posted by Wayne Morellini : I asked David (from Prospect) about his codec, bulk purchase, oem for cameras etc, but got no reply. I think you can arrange special pricing from them, rather than goign through Boxx.-->>>

Wayne, sorry I must have misplaced your correspondence. I have received many inquiries to using the CineForm HD (CFHD) codec for applications like this. Although CFHD is an element of the Prospect HD editing system that BOXX is shipping (soon), we are considering the codec's licensing separately.

[Wayne Morellini]

Rob, have a look at all my previouse posts and ideas. The problem with compression at the head is that none of the manufacturers (except maybe prosilica??) have indicated they will do it, nor multi-link interfaces, so we are knee capped straight away. That is why I have suggested an itermediatory caspture card (or ITX ref with Camera link) that receives cameralink from any camera (not just this camera) and compresses it to gigabit ethernet or USB2.0. There are better codecs too, look at the ones mentioned in the threads, even Prospect HD. You could even use).

Forgot to mention, the Russian guy (mentioned on the first page of the Home made thread) has openware comrpession circuits he used with his camera. Honestly if we had a PCI-Express/AGP camera link capture board/mainbaord, or 10Gb ethernet, we would not need comrpessed transmission at all then it could all be done in the computer (maybe next year).

[Wayne Morellini]

Hi David, it was the second post after your last post, I was meaning to follow it up with personal correspondance shortly. I am interested in some of these questions before we get our chioces locked in. As long as the solution is cheap enough to not break the $5K budget. Actually we only need the features for seriouse film/TV work, maybe aZ lite version ;) The big worry I see with Prospect is that the comrpession might take so much CPU time that it will blow the budget for a portable capture device.

Rob, that is what I was trying to ask in the other thread, about estimated CPU performance rather than disc perforamance.

Rob the FCP cousin, good idea, I hope you do well.

Taking the rest of the week off guys see you latter.

Thanks Guys.

Wayne.

[Rob Scott]

<<<-- Originally posted by Wayne Morellini : The problem with compression at the head is that none of the manufacturers -->>>

If we do compression at the head, there is no point unless it's some form of lossy compression. I really don't think even 2:1 lossless (probably doable with a standard CPU) will buy us anything because it will have to be *after* the Bayer filtering, which will require even more CPU, and immediately expands the data by 3x.

<<<-- [Prospect HD codec] ... the compression might take so much CPU time -->>>

Yes, that's a big concern, plus the licensing fees. The project would no longer be fully free/open at that point either. (Assuming that's an important goal.)

<<<-- Russian ... has openware comrpession circuits -->>>

Yeah, that looks cool, but a FPGA board like that is way beyond what what I can manage. And that's probably what it would take to do any sort of lossy (especially "near lossless") compression.

<<<--- about estimated CPU performance -->>>

Ah, missed that, sorry. I'm not sure how to estimate CPU performance.

Have a great week!

[Steve Nordhauser]

Rob Scott: You nailed it:
<<If we do compression at the head, there is no point unless it's some form of lossy compression. I really don't think even 2:1 lossless (probably doable with a standard CPU) will buy us anything because it will have to be *after* the Bayer filtering, which will require even more CPU, and immediately expands the data by 3x.>>

I did a lot of research work on lossless compression and that is a good number. Absolutely right about Bayer. The only advantage to doing Bayer before storage is for a preview window. I think that you always need two Bayer algorithms: quick and dirty for preview and as good as you can get for post process. As much as I don't want to deal with 6-12 disk drives on the high end, why capture 12 bits of HD and use a nasty Bayer algorithm just because it is faster?

By the way, was my title "lossy compression" or "lousy comprehension"? That was penned by one of my group - Steve Aceto, during a discussion of compression quality. I like it.

[Rob Scott]

<<<-- Originally posted by Steve Nordhauser :
why capture 12 bits of HD and use a nasty Bayer algorithm just because it is faster? -->>>

Exactly so. I had another idea recently -- a "delayed processing" feature. The camera head captures raw direct to disk; and then during pauses in recording, it executes a background task to do high-quality Bayer filtering and compression. (This would be optional, of course.) Problem is, it would still expand the data unless we used a lossy codec.

<<<-- By the way, was my title "lossy compression" or "lousy comprehension"? -->>>

I like it! I had an idea once for a humorous essay that involved a new technology for business communications known as "idea-based compression". Instead of compressing individual characters or sentences, it would replace entire chunks of text with standard "ideas" from its library. You know, I think it would work on 99% of the memos I get ... :-)

[Wayne Morellini]

Ahhg, that Bayer filter again (no offense intended), maybe we should look beyond Bayer because we will get options like 3 chip shortly, and Foveon x3 like sometime. Lossless 2:1 compression makes Gigabit Ethernet like 2 Gigabit Ethernet (which is more than 1080, and more than 8mp bayer SHD), and thats all we need. I'll say this again, as far as going RAW for Thearactical work, Bayer is like shooting yourself in the foot, at 1080 and lower. Also lossless compression should require much less computing power than lossy compression. If somebody does direct to disk Raw off of the Sony 3chip HDV1 or Canon, then it would be a bit piontless to do raw 1080 bayer instead.

Prospect HD is only one (commonly talked about) option, if it canbe had cheap enough we shall be doing good. Where is David?

The Russian board, I am not really suggesting that you do it, but if somebody here could, or maybe you could ask him direct. He's a bit busy at the moment (emailed him weeks ago) but maybe he would be interested latter. Seriously unless there is somebody that can do it we might as well just drop the idea. Neither Summix or SI has volunteered to do it though a capture compressor camerlink to Ethernet card may offer price advantages for variouse customers.

Delayed compression Rob, nice idea (Tick symbol).


OK Rob's, I have a suggestion, sponsorship. Why not get somebody like Apple (or one of the editor makers, like that Linux one), or VIA to help develope the hardware or software as an Opensource technology initiative, or maybe fund you to develope it? I think it would be a great coue for Apple's multimedia production initiatives, for them to advertise, press releases etc, or for VIA a modified ITX reference platform with camera link or GB/10GB Ethernet, 2-4 top end processors, and raid, would be a great promotoional platform for them (they have so many reference platforms it isn't funny).

[Wayne Morellini]

OK

My promissed discussion on my future research plans, I shouldn't really list these, as you don't know what I can do.

I want to look at server Tape backup for cheap storage of HDD matarials and maybe as prime storage item using buffered data.

Find out more about Prospect HD.

Look into special event software and applications, to see if the company involved might like to build support into their systems. Probably worth 10K sales accross the states. This might also form a option for our systems.

Serach out cameras with compression in the head (please somebody take this one from me).

Look for high class sensors, maybe find somebody willing to explain what to look for , and how to read the data sheets correctly.

Research into small PC platfforms, contact manufacturers see about 10 GB Ethernet. Maybe see if there is a form of multiplatter drive that can break the 50MB/s barrier (anybody).

Cameralink boards/itx version

That new consumer HD serial cable standard I saw 5GB/s, beats GB Ethernet or Camera link, something even VIA, and the camera manufacturers could be persuaded to include.

I know of a guy interested in our cameras and the hardware capture side.

Well that upto a months worth of research there allready, if anybody wants to help thanks. There was more but I don't remember.

[Wayne Morellini]

Well David Newman here is that post again, I was wondering if you could give us some verification.

Thanks

Wayne.

<<<-- Originally posted by Wayne Morellini : Thanks David, I just thought to keep everything tight.

As you can tell there is much camera and computer hardware out there, and everybody has their personal likes and dislikes for editing aswell. I have suggested a general capture board on the hardware side (but nobodies biting on that idea) and on the software side plugins for capture and editing which you maybe able to help. What can your software offer in this respect, or will we have to buy a whole editing/capture package (I have not read the information on your site, I have been snowed under recently).

How much comrpession can you get with vissually near loosless, and how much true lossless?

The processing requirements you quoted, how much more improvement can you expect (I guess they are twin 2.8Ghz chips)? Are you using highly optimised C, or have you gone the final speed step to optimised Machine code, as these can give big gains?

I have not really told people too much about this, but there is a potential cheap technology out there that I want to research in the future that could reduce the pocessig requirement greatly.

As you can tell if people take to what we are expecting (and I might aswell go on the record as saying I will only take to a custom camera myself with specific price and visual performance advantages) then there will be a new market using a variety of datarates, resolutions and bit depths. Could this sort of flexibility be offered in your product for that market?

If people like your package what are the options and purchase options, and are there any attractive bulk purchase options that we could get into in private? Some people here would spend thousands, others would spend much much less.

I have allready located a Linux editing system that captures in HD, but it will be too much trouble for most, and a lot of power users probably would prefer something like your system.

Also there have been a few people wanting to find out more on your product here, maybe they can ask questions.

I would like to say I'm a bit of a novice to the issues in videography, but good on the technical side, so I might not understand all the issues.


Well I have to go to bed, so I hope I can pick up your reply tommorrow.

Thanks for turning up.

Wayne. -->>>

[Rob Lohman]

Wayne: I would certainly be interested in some sort of sponsor-
ship. If we develop software / release source code everyone
benefits.

I'm just not sure how to make this happen soon. I personally
do no have such contacts. I'll see if I can get a message out to
some people and see what happens. If someone wants to send
a chip in return for software do let us know! <g>

p.s. this is for the chip / solution we are discussing in this thread!

[Rob Scott]

Ahhg, that Bayer filter again (no offense intended), maybe we should look beyond Bayer because we will get options like 3 chip shortly, and Foveon x3 like sometime.

The software architecture won't rule out any of those options, but version 1.0 must run on a system that is affordable and easy to build today. As an individual, I can't build or afford a 3-chip system, and who knows when Foveon will have a 30+ fps chip? (I've been watching Foveon for a while, hoping they would come out with one.)

There are going to be several performance/price points that people fall into. Some need nothing less than full 1080p x 3 chip for best possible performance when uprezzing to 35mm. Some of us want better performance than DV/HDV but are more concerned about getting great DVD output and perhaps HD-DVD.

I think we'll end up with several camera systems to meet different people's goals. If the software is modular enough, people can pick and choose the components they need to support their specific needs. That's the goal anyway ...

[David Newman]

<<<-- Originally posted by Wayne Morellini : Well David Newman here is that post again, I was wondering if you could give us some verification.-->>>

Sorry I guess I missed this the first time around.

> What can your software offer in this respect, or will we have to buy a whole editing/capture package (I have not read the information on your site, I have been snowed under recently).

We do software only. We have "yet" to bundle our software with hardware, although OEMs like BOXX are bundling our software on their hardware. We are focused on being a software provider.

> How much comrpession can you get with vissually near loosless, and how much true lossless?

Visually lossless with CFHD is achieved between 9 and 10:1 CFHD on 1080p content, we target 8 and 6:1 for extra headroom. Mathematically lossless is around 2.5:1 with 1080p content, although aren't shipping a mathematically lossless codec yet.

> The processing requirements you quoted, how much more improvement can you expect (I guess they are twin 2.8Ghz chips)?

I don't know which quote you are speak of. Aspect HD is always quoted with a single processor system. Prospect HD is typically quoted with a dual Opteron at 2.2Ghz.

> Are you using highly optimised C, or have you gone the final speed step to optimised Machine code, as these can give big gains?

We are primarily SSE2 and MMX optimized.

> As you can tell if people take to what we are expecting ... then there will be a new market using a variety of datarates, resolutions and bit depths. Could this sort of flexibility be offered in your product for that market?

I think so, yes. Although I believe any of our current three products (Connect, Aspect and Prospect HD) could be used for this market, it would be interesting to design a product for a closer fit. This might mean CameraLink support with real-time Bayer compression and/or a lower priced software only 10bit processing solution for Premiere (i.e. Prospect Lite.) These are just random ideas, for concert and available products go to cineform.com.

[Wayne Morellini]

<<<-- Originally posted by Rob Scott :
I think we'll end up with several camera systems to meet different people's goals. If the software is modular enough, people can pick and choose the components they need to support their specific needs. That's the goal anyway ... -->>>

Your right, and well see if Sumix delivers a three chip camera at 3K is a few months. It is also preferable (business sense) to have a couple of suppliers, and SI and Sumix seem to be just fine for uncompressed cameras.

Obin I forgot, that Lossless codec referenced in the pluginz link on your thread is doing 110mb/s on a 1 ghz G3, so maybe it is possible for something like that (I referenced a free one in the Viper thread some time back) to serve your need on a 1 ghz nano-itx board?

David, sorry. I will have to leave your reply to my message until tommorrow as it is allready 5am and I have to get some sleep ;). I did read through the Russian camera thread and managed to get some of the info from that.

Thanks

Wayne.

[Steve Ipp]

Dear Steve,

We have decided to use Rockwell/Altasens ProCamHD 3560
http://www.altasens.com/products.html#2 It is excellent in many
aspects. It has QE of 70%, almost two times better than other industrial
CMOS sensors and even better than CCD sensors. It has frame rate of 60
progressive (pixel rate of 150 MHz) which would be useful for special
effect needs. The first camera will be a single sensor, later we decide
about 3 sensors version if we can solve the interface problem. Only the
development of digital section of camera, supporting 150MHz, is going to
take another three months. The progress with sensor section depends on
how fast we get the evaluation board from Altasens and other issues. So,
my best estimate is by November of 2004 to have a working prototype.

Best regards
Farhad