Capturing at Higher Res for Film-to-Digital Transfer....

[StopMoWorks]

.... or Animation!

This was from another topic that VideoFred commented on:

I can only speak for myself, with my 1024x768 machine camera. I never capture uncompressed, but I use the lossles Huffyuv codec.
And even then 30 minutes of film takes 20GB!Fred.

Quote:

do you have the irez camera?

No, I have this one:
http://www.1394imaging.com/products/cam ... ccec397c1e

And this kind of lens:
http://www.1394imaging.com/products/optics/std_pentax/

But my lens is the C3516-M 35mm and its not on this list anymore.

I see that the DFK 31F03 camera comes with software and is this what one uses to capture the single images, as for example, in film-to-digital single frame transfer applications? I assume, you can set the size of the image to be captured. What is the actual image format that it captures to .... .bmp, .jpg, tiff? Does the software that comes with the camera have settings to capture uncompressed? Video Fred or anyone, please feel free to comment!
LIO 8)

EDIT / 5 hours later: Okay .... nevermind. I can hear the wind whistling and the sound of crickets on this topic I just searched deeper at the 1394imaging.com about the software that comes with DFK 31F03 firewire cam. It says:
--------------------------------------
IC Capture allows the end user to acquire single images (BMP, JPG), image streams (AVI), and auto-enumerated image sequences (BMP, JPG) under Windows 2000 and Windows XP. IC Capture may be used with any video source as long as it is compatible to DirectX.
--------------------------------------
It does not say anything or I could not find info about capturing "uncompressed", unless it is the BMP format that is considered uncompressed? It is my understanding that BMP is suppose to be a "loss-less" format.

[LewisASellers]

.... or Animation!

I see that the DFK 31F03 camera comes with software and is this what one uses to capture the single images, as for example, in film-to-digital single frame transfer applications? I assume, you can set the size of the image to be captured. What is the actual image format that it captures to .... .bmp, .jpg, tiff? Does the software that comes with the camera have settings to capture uncompressed? Video Fred or anyone, please feel free to comment!
LIO 8)

EDIT / 5 hours later: Okay .... nevermind. I can hear the wind whistling and the sound of crickets on this topic I just searched deeper at the 1394imaging.com about the software that comes with DFK 31F03 firewire cam. It says:
--------------------------------------
IC Capture allows the end user to acquire single images (BMP, JPG), image streams (AVI), and auto-enumerated image sequences (BMP, JPG) under Windows 2000 and Windows XP. IC Capture may be used with any video source as long as it is compatible to DirectX.
--------------------------------------
It does not say anything or I could not find info about capturing "uncompressed", unless it is the BMP format that is considered uncompressed? It is my understanding that BMP is suppose to be a "loss-less" format.

I've written BMP encoder/decoders in C++ on the Windows platform before. The format is a somewhat slippery one that could change at any time, and there _might_ be a variant of BMP that is lossy, but YES, all the variants I know of are lossless.

BMP in the wild typically comes in two flavors: RLE compressed (which is the weakest kind of simple, lossless compression there is) and uncompressed. Very uncompressed. BMP's tend to waste space doing alignment padding and the file will typically be larger than the actual size of the raw data.

In short, BMP is the worst possible format to email a photo to someone with.

(PNG or TIFF are _generally_ the best to store lossless processed photos in.)

--min

[VideoFred]

Yes, the stills are in bitmap format. The avi's can by with any codec, or uncompressed.

But please take a look here:
viewtopic.php?t=10377

These bitmaps are taken from Huffyuv Avi files, with virtualDubMod.


Yes, I do my best to keep my topic on top :mrgreen:
It goes fast, this morning.

Fred.

[StopMoWorks]

Woo-Hooo! I got some replies

Whoaaa .... LewisASellers .... you reeally know your stuff :o Only "code" I sort-of know is Html The camera VideoFred uses also comes with software for programmers. Just a hypothetical .... Would it be possible to modify the software to capture in TIFF format? For Stop Motion there are a variety of specialty frame grabbing software for animation and I have a special page devoted to them where I give a general overview .....
http://www.stopmotionworks.com/stopmosoftwr.htm
.... most of the frame capturing software works with DirectX and they all have as their highest resolution, 640 x 480 which is the standard of consumer DVcamcorders. I know animators are itching for higher res capture and their only choice thus far seems to be the digital still cams with their inherit bugs & longevity issues. Something like VideoFred's cam choice with good ol' manual C mount lens plus specialty higher resolution animation frame grabbing software (that includes onion skinning, frame scrolling, etc) at the higher res of 1024 x 768, for animators,would probably be a welcomed alternative/option to the digital still cams. I am not sure if higher resolution frame capture for film to digital transfer would also be desired for Super8? Maybe for 16mm to digital transfers? I assume the eventual progression from DVD will be High Definition format and just thinking about the future. Speaking of HD .... check this out at the same website .... 1280 x 960 resolution! That's already High Definition quality!? .....

DFK 41F02 firewire cam @ $1190. (USD)
http://www.1394imaging.com/products/cameras/dfk41f02/
and another one....
DFK 41BF02 firewire cam @ $1290. (USD)
http://www.1394imaging.com/products/cameras/dfk41bf02/
The above cams cost much less than the pricier Sony DFWSX910 firewire wire cam @ $2400. (USD)
http://www.1394imaging.com/products/cameras/dfwsx910/

All the cams listed here .....
http://www.1394imaging.com/products/cam ... ccec397c1e
Great resource/link you provided VideoFred!

LIO 8)

[VideoFred]

Woo-Hooo! I got some replies

OK, Lionel, then just to do you a favor, here is:
*ANOTHER REPLY!*

The camera VideoFred uses also comes with software for programmers. Just a hypothetical .... Would it be possible to modify the software to capture in TIFF format?

You need to download their developer kit, and some knowledge about Visual Basic or C++. I do not know about the TIFF format, maybe you could ask on their Forum, have you seen it?

However, someone wrote an application for me, to do the double capturing, and it works very fine. (but it creates and saves 2GB bitmaps)
It runs at 1/2 frames/sec. Recenty, I changed my backlight and now I can take frames at higher shutter speeds, mabe it runs a little faster, now.

Could be interesting for you: sequence capturing of stills is already available in their standard capturing program. And you do know it is so easy to import stills in virtualDub? Just clic on the first one, all the others are following. It takes about 5 minutes to make an AVI this way, with the desired frame rate. But I guess you StopMotion guys have more sophisticated tool for this. Is there a chance you have software to make new frames from two different stills to create smooth motion?


Fred.

[LewisASellers]

Woo-Hooo! I got some replies

Whoaaa .... LewisASellers .... you reeally know your stuff :o Only "code" I sort-of know is Html The camera VideoFred uses also comes with software for programmers. Just a hypothetical .... Would it be possible to modify the software to capture in TIFF format? For Stop Motion there are a variety of specialty frame grabbing software for animation and I have a special page devoted to them where I give a general overview .....

I have, for other projects, on the Windows platform written encoder/decoders in C++ for a variety of image formats, including TIFF. I believe roughly that TIFF took something like two weeks thereabouts for the _base_ functionality. TIFF is one of those formats which tries to be THE image format. And what I mean by that is that it incorporates new technologies into it's specification as they come available. The latest spec version of TIFF for instance can imbed a raw JFIF (ie, JPEG) image stream.

For these reasons TIFF is among one of the hardest image formats to support -- it has so many sub-formats, official and otherwise.

As for your question, it depends on whether they have the source code available or not. And what platform it is on.

Theoretically if the source code was available and it wasn't too badly mangled, decently commented etc it might not take too long to integrate TIFF format in (since I already working source code for TIFF that I know intimately). But there are a lot of variables so it's hard to say off hand. (On the Mac, btw, it comes with support for many of the subformats built in, which is nice.)

BTW... since we're chatting... I have onion skinning in the app Eric and I are working on. Currently I have it set up so that you can move a slider and choose between -12 to +12. +1 meaning you see the current image plus one frame ahead. -1 meaning you see the current image plus one behind. And there's is an alpha slider to adjust how much the other frames blend through.

Basically what I'm asking is, personally, if you could have the onion skinning set up anyway you liked, what would that be like? I played around with stop motion on camcorders and still cams of course, but not the extent you have obviously. Seems if anyone here might have a good opinion on the matter it might be you.

I was thinking of adding an option that would create an inbetween image that you could toggle on and off to work as a reference. The 'tween image would just be a subtraction of the current image and the previous one (or ones) visually showing the change. A "Log" option would boost the more subtle intensity changes so it would be easier to see. Of course I can make it do pretty much anything....

(I would continue, but at the moment, as usual, I'm about to fall over out of the chair. Tired.)

http://www.stopmotionworks.com/stopmosoftwr.htm
... I assume the eventual progression from DVD will be High Definition format and just thinking about the future. Speaking of HD .... check this out at the same website .... 1280 x 960 resolution! That's already High Definition quality!? .....

480P/I, 720P/I and 1080P/I are the standard HDTV resolutions. I'm not an expert, but it seems much of the HD content imoving to the 720P or 1080I resolutions.

720P (ie, 1280x720) seems to be the resolution to aim for for HD.

So... I have a question about the aspect ratio of a 1280x960 still being converted to 720P ... but otherwise, yes....

Note: My calculator informs me 1280x960 is in fact 4:3 as I suspected. HD is normally 16:9.

--min

[Justin Lovell]

HD may normally be 16:9, but super 8 is always 4:3... so what do WE do now..?

so if i were to make a move in an HD digital camera direction.. still using cinecap...

1. Which camera is the camera of choice to go with? Again, i'd still be using cinecap, capturing at 6fps w/ WPXP.

2. And of the camera the you can recommend, what are the manual controls that are available with it/ does it also have an analog output for video preview...
3. i'd need that to be able to help with my negative transfers (i have a video equalizer hooked up for preview to see how the image may render as a positive image, then i capture it as a neg and flip it in Aftereffects/Avid or premier...(unless someone has a way to invert the image on the fly)?

cheers,

justin

[chachi]

Justan, I don't know if this is what your asking, But I invert my neg films on the fly using the negitive effect feature built into my mini DV cam. Mind you I have only used it in B&W, but it does create a nice sepia tone since it really thinks its inverting colour anyway. Plus, you can see the positive image right in the view finder..

[Justin Lovell]

no no invert feature on the cameras that i'm using.. seems to be more of a consumer feature.
For some reason, they don't put cool things like negative and mosaic, or mirror image on broadcast cameras

ps i would run my image throug the analog video equalizer (to invert the image) than i would have to get an analog to digital convertor box, but the image gets degraded too much going through all the extra analog inputs/outputs.

better to just do it in the computer, non?

jusetan

[MovieStuff]

no no invert feature on the cameras that i'm using.. seems to be more of a consumer feature.
For some reason, they don't put cool things like negative and mosaic, or mirror image on broadcast cameras

ps i would run my image throug the analog video equalizer (to invert the image) than i would have to get an analog to digital convertor box, but the image gets degraded too much going through all the extra analog inputs/outputs.

better to just do it in the computer, non?

jusetan

But that isn't how the video equalizer should be used in this application. The image quality is too low. What you do is take the firewire output of the camera and run that directly to the computer. You take the analog output of the camera and run THAT to the video equalizer and then from the video equalizer to a standard tv monitor. You are still capturing a negative image on the computer via firewire but you are able to see a positive image on the tv monitor to make judgements about exposure during capture. You are STILL using the computer to invert the neg into a pos but, this way, your exposure is guaranteed to be correct because you previewed a positive image during capture via the video equalizer.

Roger
http://www.moviestuff.tv

[chachi]

I don't understand why that feature isn't on pro-sumer models? I have had several old higher end vhs tube cameras in the past and pointing them at the tv while fiddling with that feature (and others) created some wicked feedback, which is actually how you create the phycadelic background that the Wolfman danced to on Hilarious House of Frankenstein.

[StopMoWorks]

I've written BMP encoder/decoders in C++ on the Windows platform before. The format is a somewhat slippery one that could change at any time, and there _might_ be a variant of BMP that is lossy, but YES, all the variants I know of are lossless.

BMP in the wild typically comes in two flavors: RLE compressed (which is the weakest kind of simple, lossless compression there is) and uncompressed. Very uncompressed. BMP's tend to waste space doing alignment padding and the file will typically be larger than the actual size of the raw data. In short, BMP is the worst possible format to email a photo to someone with.

(PNG or TIFF are _generally_ the best to store lossless processed photos in.)

Most of the existing stop motion programs work with .... Firewire, USB, or Analog ..... depending what components/cards you have in your computer. Their max res size is 640 x480 and they mostly capture stills as JPEG, BMP and sometimes a choice of PNG. From what I read/hear and as you said, TIFFs would be the most ideal but if it makes developing a program to format as TIFFs more complex, then perhaps BMPs maybe not as ideal (but more standardized format?), could be 2nd choice. Also, it is my understanding one can later convert BMPs to TIFFS using specialty program or plugins?

Could be interesting for you: sequence capturing of stills is already available in their standard capturing program. And you do know it is so easy to import stills in virtualDub? Just clic on the first one, all the others are following. It takes about 5 minutes to make an AVI this way, with the desired frame rate. But I guess you StopMotion guys have more sophisticated tool for this. Is there a chance you have software to make new frames from two different stills to create smooth motion?Fred.

Okay .... I see that the industrial imaging firewire cam you use can already capture stills with the provided software! So in the program window, you see a live image and take a snapshot? That is very basic frame grabbing and no animation features, like onionskinning, instantly previewing and scrolling through frames ... however .... one can just use that firewire cam for the high res still captures, but one could also setup a USB Webcam (as an assist preview cam) that is connected to stop motion program and use the USB feed to monitor and check the ongoing progress of the animation (through the stop mo capture program), while the other program is controlling the firewire cam and is grabbing the hi-res stills to hard drive. So for every frame you animate ..... you manually trigger the firewire cam to grab the hi-res still, and you shoot one frame in the stop motion program that is connected to USB webcam. This is similar to how animators are using digital still cams using a video assist cam, while DSC captures & saves the high res stills on its memory card. The stills images captured are then processed in post production. This is a workaround .... of course. However, in the progression for future improved Stop Motion tools ... it would be real SWEET if specialized Stop Motion frame capturing/animation software with built-in onionskinning, frame scrolling, instant preview, etc, could be developed specifically for these types of industrial imaging firewire cams that have Higher Resolution!

BTW... since we're chatting... I have onion skinning in the app Eric and I are working on. Currently I have it set up so that you can move a slider and choose between -12 to +12. +1 meaning you see the current image plus one frame ahead. -1 meaning you see the current image plus one behind. And there's is an alpha slider to adjust how much the other frames blend through.

Basically what I'm asking is, personally, if you could have the onion skinning set up anyway you liked, what would that be like? I played around with stop motion on camcorders and still cams of course, but not the extent you have obviously. Seems if anyone here might have a good opinion on the matter it might be you.

As you already know .... for Stop Motion animation purposes, onionskinning is to see the previous frame grabbed so the animator can visually guage the incremental puppet movement for the next frame. Once a frame is shot, the previous frame is automatically onionskinned onto live image. It's a frame by frame process always judging/comparing previous onionskinned stored frame to live image. If it were more than one onionskin layer, IMO, it may not contribute much to helping the animator with the screen filled up with so many layers. Some stop motion programs do offer selectable number of onionskin layers, but most just offer one layer. At the most, IMO, perhaps 3 onionskinning layers for stop motion purposes. Maybe other animators would have different preferences. Also, these stop motion programs do permit adjustment of onionskin layer density (blending). Onionskinning is "static" way to visually judge animation movements. The other way is just scrolling/scrubbing through the previously shot frames and also repeatedly tapping left & right keyboard arrows to see the movement jump of previous frame to live frame .... that is a more "kinetic" way to see and guage the movement by popping through the frames and onionskin is used as a further aid for analyzing increment movement. Don't know if this explanation makes any sense. For film to digital transfer, I am guessing that many onionskinning layers (12 layers) is for frame image alignment purposes?

480P/I, 720P/I and 1080P/I are the standard HDTV resolutions. I'm not an expert, but it seems much of the HD content imoving to the 720P or 1080I resolutions.

720P (ie, 1280x720) seems to be the resolution to aim for for HD.

So... I have a question about the aspect ratio of a 1280x960 still being converted to 720P ... but otherwise, yes....

Note: My calculator informs me 1280x960 is in fact 4:3 as I suspected. HD is normally 16:9. --min

So even though 1280x960 aspect ratio from the industrial imaging firewire cams are not at the HDTV ratio, then they can be cropped/resized to 720P!! Good to know that!
LIO 8)

[LewisASellers]

Note: My calculator informs me 1280x960 is in fact 4:3 as I suspected. HD is normally 16:9. --min

So even though 1280x960 aspect ratio from the industrial imaging firewire cams are not at the HDTV ratio, then they can be cropped/resized to 720P!! Good to know that!
LIO 8)

HD may normally be 16:9, but super 8 is always 4:3... so what do WE do now..?

As LIO suggests, if we assume the pixels are _square_, all you have to do is crop off ... um... 120 pixels from the top and 120 from the bottom. That would be one way to do it.

720P/I and 1080P/I have what are called "square pixels" -- just means the length horizontally and vertically is the same -- as do most computers. The older NTSC 4:3 TV standards as well as 420P/I HDTV are non-square -- the pixels are slightly rectangular (which is a headache -- the images have to be converted to 1:1 square pixels to get most video effects such as convolution filters to work properly).

As long as both your source device (the camera) and your destination format (720P HDTV) are both square you can just crop like this and be done with it. That is, as long as both source and destination both use square-pixels then such operations do not threaten to cause your output to look "squashed" horizontally or vertically.

If it bothers you that you're dropping 25% of your vertical subject matter, you could scale the image to ... 960x720 (I think) then add black side bars.

If you just resize the 1280x960 image to 1280x720 everyone in the film will look was though they've been squashed severely.

(This subject can be rather complicated though if we're talked about any other type of non-square output than 720P/I or 1080P/I or if the input was non-square).

3. i'd need that to be able to help with my negative transfers (i have a video equalizer hooked up for preview to see how the image may render as a positive image, then i capture it as a neg and flip it in Aftereffects/Avid or premier...(unless someone has a way to invert the image on the fly)?

cheers,

Technically the "negative" film image is simply in transparent CMY colorspace and we're doing a CMY to RGB conversion, but it's the same as a logical "negative" in most photoediting programs (-- unless you have the photographic orange mask to deal with, in which case it gets more complicated as you have to spectrally extract the mask.)

Etc etc.

Will reply more later. Gotta go right now...

[Justin Lovell]

Technically the "negative" film image is simply in transparent CMY colorspace and we're doing a CMY to RGB conversion, but it's the same as a logical "negative" in most photoediting programs (-- unless you have the photographic orange mask to deal with, in which case it gets more complicated as you have to spectrally extract the mask.)

Well, seeing as most kodak negative motion picture stocks inherently have this orange masking to deal with.. it does become an issue. Reversal film, no orange, no problem. Negative film... oi.

So what is it that you are getting at when you say "spectrally" you can extract the mask? You've found a method that won't affect the other orange tones by keying it out somehow? There must be a simple solution as EVERY single transfer suite that i've ever dealt with for super16 has been able to just pop a switch and voila, no orange masking.

jusetan

[paulcotto]

You can get the Veo Velocity USB 2.0 1280X1024 X 24 Bit webcam and modify it for a C-Mount. This will save to big $$$ and deliver very similar results to the Industrial cameras. It works with Dodcap - Cinecap since its VFW compatible.

http://www.veo.com/Velocity_Connect/default.asp

You can have a C-mount adapter made

This guy makes the adapters>>>

http://webcaddy.com.au/astro/ProductLis ... at=adapter

I had one done and he does nice work. I had planned on building these myself and selling them but I dont have the time with my new job. You can get c-mount extension tubes and use your favorite lens to macro focus on the film gate. You can also go the big bucks rout and get a proper industrial macro lens from Edmunds optics.

http://www.edmundoptics.com/onlinecatal ... uctid=1518


Regards,
Paul Cotto