Rob:
FPGAs are not too expensive if you don't have engineering costs. That is why I keep fishing for someone to hand me a working design. Bayer needs a couple of lines of internal buffering, depending on the algorithm. Many compression algorithms require tables - also in internal RAM. Noise reduction usually requires a big slower external RAM.

I wouldn't worry about the gates. I know of products that use Virtex II Xilinx FPGAs that sell for $500 in volume. That is a gob of gates. And yes, 'gob' is my WAG.

Obin:
The Altasens does interlaced out of the box. All of our other cameras are progressive only, as far as I know. That data rate was at 150Mpix/sec. Even in 8 bit mode that is a 64 bit system.

As Steve said, the AltaSens can produce interlaced if you want it, though since you can produce 1080i60 out of 1080p30 or 1080p24, I'm not sure why you'd want to.So you're saying that a currently available FPGA should have enough gates for such an implementation? But that you might need external RAM to implement some things? Give me six months and I may be really interested in this.

Not so Polific Steve ;), there getting smaller and smaller all the time as things work out (except for the vacume energy stuff).

Yes that is what worries me about FPGA compared to Memory Stick PIM's, Clearspeed, and DSP. Some of that DSP stuff is cheap, but I think the reason that there are custom compression chips (probably hardwired DSP's) is that normal DSP might have a struggle.

My markup example was on the highend of the ratio of cost to retail, that manufactures desire (10:1) to have enough profit to share around the distrbution chain. Apparently at less than 5:1 they stop talking to you.

Vissually lossless is good (20-50:1 better for non cinematic work), but that stream of codec options again, for any cinematic blow up I would like lossless to play up with very detailed upscaling. Also I think you are much more likely to do lossless than 6:1 on a low powered PC, but each to their own method.

I understood that Altsens chip is a lot cheaper than $700, pretty dissapionting, that explains why nobody wants to do a 3 chip version. I suspect that next year single chip 8MP will be a lot cheaper, call it a feeling of the SLRD market.


<<<-- Originally posted by Rob Scott : For a lossy-but-visually-lossless codec -- as I hope Dirac would be -- it seems like it wouldn't really matter what exact bit depth you started with. Even if the current Dirac is 8-bit, we should be able to create a 16-bit version; then you store everything as 16-bit, whether you start with 10, 12, 14, or 16 bits. Does that make sense?
-->>>
Pack, Pack Pack ;)


Thanks

Wayne.

Sure, I understand the need to pack. But I am just wondering aloud whether it makes any sense if you're going to compress it losslessly anyway?

<<<-- Originally posted by Obin Olson : Steve I jsut spoke with a broadcaster and they will NOT take any 720p stuff... -->>>

Not good, over here, locally, they only want 3 chip. Have you showed them processed footage, are they likely to change there mind next year?

Maybe they want want more temporal information and less blur, pitty.

OK, I think we were both thinking about different things.

why 1080i60? because they want the fluid motion of 60fps but the format they use is interlaced..SO they need 1080i

our 32bit grabvber board will not wrork for that will it?

OK, I guess my question was, why do they want the fluid motion of 60 fps?

And no, the 32-bit/33 MHz board will not work for that. At least 64-bit, possibly 64bit/66MHz.

Shoot, the 32-bit board I have doesn't even seem to be able to handle the 60 fps on the SI1300. I get a PCI buffer overflow when I crank up the pixel clock over about 55 MHz. Steve, does that sound right? What clock rate should I be able to get?

Rob:
That would be in 10/12 bit mode. PCI-32 is rated at a max transfer rate of 132MB/sec. Sustained is between 80-110MB/sec (windows likes to mumble to itself using bus bandwidth). The frame grabber output is unpacked so in 10 bit mode is two bytes, 55MHz is 110MB/sec peak.

FIFO overflow means that the PCI bus FIFO on the frame grabber lost data because the bus fell too far behind the camera output.

OK, that makes sense -- I didn't realize it was sending unpacked. So if I really want 48-60 fps with this camera, it sounds like I may need the 64-bit grabber. 24-30 is fine for now, but I was hoping to be able to capture 48 fps and skip every other frame to get the 24-fps @ 1/48 sec "cinema" shutter effect.

What I'm hoping to do is to translate the pixel clock frequencies, blanking, etc. and provide a simple frame rate/shutter speed interface which will be more understandable to the typical user. I'm still reading and trying stuff to understand all those relationships. I won't bother you any more about it until I've done more homework :-)

why can't we get 1080 60i? I thought that interlaced was 1/2 the datarate of progressive? it should be the same datarate as 30fps progressive right?

Steve tell me about the 1080P chip and what it outputs?

so in 8bit mode it will output the full on 60mhz without FIFO overflow?

Yeah it should, since going to 8 bits would halve the data rate.I don't think the SI1300 supports interlaced. To get 60i we'd have to put it in 8-bit mode, 60 fps and then toss every other line to fabricate interlaced in software.

work-for-hire needs 1080i as that is a standard that alot of the national networks use...lots will accept conversions from 720p and some even 480p...but we NEED 1080i even if it's 8bit..

I think I misinterpreted your question. The AltaSens can do interlaced up to 120 fps (possibly 240fps @ 720i), but the SI1300 can't, unless you simulate it in software.

1080i will certainly be a goal of the project down the line. (Though honestly it's not a personal goal :-)
OK, now I've confused myself again. If I multiple it out: 1280 x 720 x 2 bytes x 48 fps = 105 MB/sec, it looks like I should be able to get this data rate.

Plus the Silicon Imaging specs on the camera indicate that a pixel clock rate of 55 MHz should allow 49 fps. Is this at 8 bits or 10 bits?

it's not my personal goal - I can't stand interlaced images - but people pay money for it - money talks


uploading the wmv HD test file again...hope it works this time

Hmmmm ... yeah, if money talked to me, it would probably be my personal goal too! :-) :-)

Hey Obin,

There's no reason why you can't generate a 1080i signal from a 1080/30p or even 1080/60p signal. Frankly interlace is so . . . ugh. But you can't go from interlace to progressive very well, although going from progressive to interlace is quite easy.

try this:
http://www.dv3productions.com/Video%.../HD%20test.wmv

Re: HDTest.wmv clip
 

Obin,

The images are very clean (such low noise.) This more than proves the capabilities of these camera types. The dancer sequeue seems to show more natural smooth blur. This is very nice.

The only weird thing I see is that half the sequences have a strong motion ghost. I have extracted an example frame the shows the problem:

www.cineform.com/temp/ObinGrab01.jpg

The flower's step appears to be in two places at once. Yet other sequences not have this artifact at all. Was this a filter you have applied?

flower in post filter -yes
dancer(my sister) - 48fps playback at 24fps for slow-motion effect

take a look at a comparison frame!
 

Obin, Interesting footage!

To see where things are related to the consumer JVC HD camera, I made this frame that has a frame of yours on the top and a frame I took with my HD10 on the bottom. No image enhancement was done to either.
I think the HD10 is a good camera to compare to, because it is better looking than DV, but has many other limitations for serious film making. ( only 8 bits, no manual control, etc )
I know the Bayer demosaiker is junk right now, etc, but people should see what the JVC does. Some people are so used to looking at DV, it's sad. I call DV web cam resolution these days, because that's what it is, really.

300KB comparison frame:

http://home.earthlink.net/~lesd/hd/JVC-1300-comparo.jpg

Other images in that dir are my orbital GG mechanism, in progress.

-Les

Well, even the slow motion scene has ghosting and I don't know why, I still think a 24 fps cycle with a mechanical shutter would be better.The first time I used a Cinealta camera, back in 1999, I remember it had the option of a mechanical one( don't know if it was a prototype or what).
I mean, it isn't exactly a natural motion blur but looks like a cheap frame blending (like the one used to add blur to synthetic images)
Obin, did you use some kind of noise reduction filter (temporal) ?

Also, have you tried the dual slope mode?
I'm really interested about how it looks, but nobody says anything about.

It sounds simular to frame blending they use to go from 24/25 to 30FPS and vice versa, is something changing the frame rate along the workflow/viewing cycle. But then it doesn't quiet look like it. Steve?

<<<-- Originally posted by Juan M. M. Fiebelkorn :

Also, have you tried the dual slope mode?
I'm really interested about how it looks, but nobody says anything about. -->>>

Yes. Steve, how does this actually work, is it exactly the same as the Small sensors version? You said the results look un-normal, but is it possible to manage the positioning of the range of both slopes so they naturally run into each other to give a visually continouse range.

Thanks

Wayne.

this sensor does not have dual-slope! it's not a crappy FillFactory chip like the IBIS5 with dualslope

I have not used any frame blending at all maybe somthing in the compression...I see NOTHING on the original that has any type of frame blending arifacts...the flowers have 2 layers that fade in and out over each other...everything else is 24fps and 48fps

webcam....HAHHAHAA perfect! I love that! well a really good dvcamera has a good pic but it is low res forsure!!

Les I have the camera and computer at home today do you want me to get the 2 images you have been after for months!? ;) I could do it today...let me know what you want and I will do my best

Help! I just bought a 4mm c-mount lens and when it arrived it works - but with 4 very dark corners...it's like it was made for a 1/3 in camera chip not 1/2.....how do I know what to buy for lenses for this camera...I can see that c mount is not the only thing to look for! (4mm is a NICE wide I wish it was 2/3 in size!)

in looking at the back of the lens I can see that the glass on the back is smaller then the 12.5mm and 75mm I got from Steve...is this a clue?

on the lens the 4mm reads:

1:2.2/4mm

and the 12.5mm reads:

12.5mm 1:1.3


I understand the 12.5mm and the 4mm, what about the 1:1.3 and 1:2.2??



this looks 100% like what I have here but it's 25mm not 4mm:

http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&category=30077&item=3826939852&rd=1

On a comical note, Robs, I just watched this without speakers, and still laughed:

http://www.sun.com/aboutsun/media/fe...nsidejack.html

Not many listings appear to specify the aperture size (is that the right way to say it?). I noticed that there were different sizes when I was trying to figure out the lens I bought. If you look here you'll see that these lenses are specified as having a "format" which can be 1/2", 2/3" or 1". In this line of lenses, the 1/3" format is only for CS-mount lenses.

Sorry, I don't really have any more advice.

Software update: I'm making some progress with writing to multiple disks, but I range into comething strange with my hard drives. I have two 160 GB drives which should be identical, both mostly empty and unfragmented, but one gives me 22 MB/sec and the other ~12 MB/sec. They are on separate IDE channels, one is the master on its channel and the other is the slave on its channel. There are more details in the blog entry for today. Thanks!

Test images
 

Obin,
OK heres what would be great:

Take 3 pictures, all of the same motionless setup ( still life ).

Try to use the 1/24 or 1/48 sec exposure and timing you would use for a motion picture. ( very Important ! )

Store the pics as B&W raw with > 8 bits
Have the image be of something with darks and lights in the same scene.
Have the camera be way out of focus for the images. ( see example below )

Even though it's out of focus, the images should still have areas with large dark and large while areas.

If this pic was motionless, this amount of focus would be perfect:
http://home.earthlink.net/~lesd/hd/test.jpg

So, It's out of focus, but there are still lights and darks.

That's it, let me know if you have questions!
-Les

Ok Les I am doing htat now for you

You need to give me your email address
I have 3 tiff files 16bit ready to send you

Obin,
First of all, the footage looks quite impressive from a color and detail standpoint, but I'm also quite curious about this frame blending. When the video is playing, certain movement seems to have an odd strobe like effect. When I paused it, I realized there was a double image in several of the frames.
What did you use to encode the WMV file? There are alot of options in Windows Media Encoder and Adobe Premiere Pro when it comes to encoding WMV (not sure about Vegas).
I take it you encoded to 24fps, Non-interlaced. Perhaps you had inverse telecine turned on or some other feature that WM9 uses for encoding different kinds of 24fps materials.

Other than that, the artifacts associated with Bayer imaging are only noticeable when I pause the video and get up close to my projection screen (I mean less than two feet away.) Specifically, the edge of your earlobe in the last shot of the wmv, since it's against a black background, shows a kind of dithered anti-aliasing.

And then lastly, I keep thinking that I see a set of thick bars running horizontally across certain shots. It's noticeable in the darker parts of your face when you hold up the watch. It is almost exactly like what was in the low-res quicktime file of the girl waving that shiny fabric around.... but drastically less obvious. Do you notice that as well? It's always in the same place in the image, so it could obviously be corrected with a filter in post if it couldn't be worked out before hand.

Keep up the pioneering.

I'm feeling sad at this moment cause I see the technical part of this thread seems to go nowhere most of the time....
Anyway here I go again.

I'm trying to develop an opensource project for a codec for this project.The main idea is to use a mini-itx motherboard with a P4 or athlon (may be a mobile athlon).Now the codec is performing more or less good, giving around 20 fps for a 1280x720 RGB.
The final idea is to have at least 30 fps for 1920x1080 Bayer.
It is based on a hybrid of DCT and DWT.
When the moment arrives, I will need some expert coders to optimize it.
Anyone here is able for this task?
Thank you very much.

P.S. Everybody seems to ask for complete off-the-shelf FPGA projects.The Russian Net-Camera is a working FPGA project.I know it gives 15 fps, but in the worst case we could duplicate its compression system to get 30 fps(I mean a duplicated FPGA board.)
There are also some huffman compression working FPGA projects, which I posted before but nobody seems to notice them.

Ok, send it to my gmail account, it has that nice free 1 gig of storage that they want to snoop into for sending me junk popups :)

it's lesdit at gmail
com

-Thanks
-Les


<<<-- Originally posted by Obin Olson : Ok Les I am doing htat now for you

You need to give me your email address
I have 3 tiff files 16bit ready to send you -->>>

matt how does it compare to the HDTV signals you have seen from broadcast? does it look clear? if you sit back would you say ya this could be a movie...or nope this is pure cheap looking video even though it's HDTV resolution??

Stirring up the soup
 

OK, I was hinting that I might have something interesting brewing. Micron has a 3.2Mpix sensor that is part of the same family as the 1.3Mpix that Obin and Scott are using. The down side is that it is the same size sensor so the pixels are smaller (3.2 micron). The interesting thing is that we have tested it so far at 1920x1080, 10 bit, 24fps and 1280 x 720, 10 bit, 48fps. We will see if they make it to 30fps/60fps. The other thing is that this sensor is the one suggested by Micron when I discussed the oversaturated smearing of the SI-1300. Next nice thing - single piece pricing is $2195 alone or $2695 bundled with the Epix 32 bit FG. We will be doing a 64 bit bundle when the Epix 64 bit FG is available.

This is not a replacement for an Altasens camera. It is a cost effective solution for some applications. I'll try to get some test images posted when I get a camera - they are out getting integrated with Epix and the gigabit ethernet interface right now. I don't have a manual but the sensor data is available from Micron.

Obin, the numbers reffer to aperture opening, not lens coverage. Try to find Cmount lenses that cover 2/3, 16mm or ask some questions before you buy.
Apart from vignetting, how does it look?

<<<--
I understand the 12.5mm and the 4mm, what about the 1:1.3 and 1:2.2??
-->>>

Re: Stirring up the soup
 

Steve, this sounds good. Make sure that if you get to do some test images, they include some like what I just described to Obin. I know they are very ugly to look at visually, but they say a TON about how good the camera is for movie camera use. True DP's ignore the artistic content of test footage, DOF, and lighting, they look behind the 'scenes' to see image quality.

Maybe the smaller pixels will be OK ? Let's see!

-Les


<<<-- Originally posted by Steve Nordhauser : OK, I was hinting that I might have something interesting brewing. Micron has a 3.2Mpix sensor that is part of the same family as the 1.3Mpix that Obin and Scott are using. The down side is that it is the same size sensor so the pixels are smaller (3.2 micron). The interesting thing is that we have tested it so far at 1920x1080, 10 bit, 24fps and 1280 x 720, 10 bit, 48fps. We will see if they make it to 30fps/60fps. The other thing is that this sensor is the one suggested by Micron when I discussed the oversaturated smearing of the SI-1300. Next nice thing - single piece pricing is $2195 alone or $2695 bundled with the Epix 32 bit FG. We will be doing a 64 bit bundle when the Epix 64 bit FG is available.

This is not a replacement for an Altasens camera. It is a cost effective solution for some applications. I'll try to get some test images posted when I get a camera - they are out getting integrated with Epix and the gigabit ethernet interface right now. I don't have a manual but the sensor data is available from Micron. -->>>

like a really wide lens ;) it's a cheapo lens but it would have worked if it was full coverage....oh well ...the ebay guy is nice enough to take it back - the shipping ;)


Steve keep me in the loop on the 3.2megapixel...the smear upgrade is worth it alone from everytying else...what will the cost be on the Altasens camera anyway?

when do you think you will have a Micron 3.2 for testing and image shooting?

A Xilinx Spartan3 FPGA 1 million gates costs $15
And here is a developmente board

http://www.nuhorizons.com/products/x...ent-board.html

Free IPs

SDRAM Controller:

http://www.cmosexod.com/sdram.html

Open Core compression system

http://www.opencores.org/projects.cg...stems/overview

Diagrams of the Elphel camera:

http://www.elphel.com/3fhlo/index.html