In a small corner of the South Hall at NAB there was this Swedish company that has years experience doing analysis of film for auto crash testing and defence. Those films where you see cars with dummy humans crashed into walls. And slomo films of missiles leaving their silos.
These films need to be analyzed and data used to re-design things for safety or destruction. Whatever. And this company does just that.
Now they've used that expertise in designing a scanner and film analysis system that scans using capstan transport but providing the steadiness of pin-registered scanning. At least that's how I understood it. Its called Golden Eye.
The scanner is quite small, almost like a Cintel diTTo. Table-top. And there's a Win PC to analyze the film and save it as .DPX or a variety of formats. Scans can be 4k or 2k and there's even a swap option. Its a line scanner so many in-between resolutions are also supported.
The light source is a halogen lamp, what we call a 'cup halogen' lamp in India. Its a halogen bulb with a reflector and its an ordinary light source used in showrooms. My German friend even remarked that its a 5 dollar lamp. Light is conveyed to the film via a 'light pipe' so film doesn't get the heat. And a complex system of current control and filters ensures that the colours and intensity stays constant across the life of the lamp. Very ingenious and simple.
The stated purpose of this scanner is for archiving. And should be too. This transport is best suited for old film that has sprocket damage. In fact I think you probably don't even need sprockets.
Other nifty features. There are two 'cameras'. One scans the frame and the other, the entire film. So keycodes are read and need no complex alignment. And these two cameras help the software stabilize the film and ensure a steady frame. The other surprising feature is that the scanner is natively 35mm or 16mm ready with no complex replacement of wheels or lamp house.
Speed is 12fps at 2k and 3fps at 4k. And maybe they could get faster at HD or SD. So its a good healthy speed without the need for very high speed storage that 2k real time needs. 2k at 12 fps should need 180 MB/sec capable meaning HD-Ready storage should do. That's for 2k full aperture. If you're doing 3-perf, 1:1.85 or 'scope, you'll have smaller files so even lower data rates. Even 4k at 3fps amounts to about that much data rate.
The software can ingest EDLs and even ALEs as a source for scan lists. And they told me they could even work out a software feature that pre-reads the film so sorted negs can be read against cut lists and scanned by keycode. That would be a killer feature.
I think this scanner needs a look-at if one is into film restoration, even VFX. And since they can output Quicktime, I'd even try it out for dailies.
There's even a smaller version that's a bit larger than a Digi Beta deck! But its slower and does 1000 ft loads only.
Showing posts with label restoration. Show all posts
Showing posts with label restoration. Show all posts
Thursday, 8 May 2008
Saturday, 15 March 2008
Media content repurposing - A view from India
These days I've been getting a lot of queries on this - media repurposing. Sounds like a new buzz word, but this can become serious business in the next few years.
Media repurposing consists of taking existing media - 35mm print/neg, DigiBeta/BetaSP/U-matic/VHS/DV tapes, an other media - and converting them to a format more friendly to new media vehicles. And playable on mobile phones, portable media players, or sent over a wire or wireless to homes and offices.
IPTV and VOD are two such new vehicles. IPTV is Internet Protocol Television and VOD is Video on Demand.
IPTV consists of sending video to viewers over the normal Internet. In India MTNL (and other services) show many cable TV channels over their broadband internet line - sent to homes over a telephone line.
VOD is when a viewer can go through a menu or index, and choose what he/she wants to see. That content is then played immediately, or downloaded over the Internet or wireless to view later.
Where IPTV and VOD differ from normal TV is that in normal TV many TV channels are being broadcast simultaneously all the time over a cable. In IPTV, only the channel the viewer wants to see is transmitted.
In VOD, only the actual program the viewer wants to see is beamed for as long as the viewer wants to see it. So VOD is sort of like watching a movie on a flight after selecting it from a menu. Like movies downloaded over the iTunes store for AppleTV. Or, in India Movies on Demand from Eros.
But Eros's Movie on Demand only work with PCs as it needs to detect Windows Media Player, which no Mac has. So if you have a MacBook, or iMac or any other Mac, forget Eros.
Other VOD examples are, of course, YouTube, or for really good-looking movies, Vimeo.
While repurposing media, it is also possible to break the programme down and classify it intelligently - like chapters in a DVD so viewers can see a portion of the entire programme. Like songs in a Hindi movie, or important scenes. Or like innings or overs in a cricket test match, ODI or T20 match.
How about going through a list of Hindi film songs and narrowing that down to Shahrukh Khan-Rani Mukherjee songs and selecting one you'd like to see. Or through a list of cricket matches and selecting one of Sachin's heart-breaking nineties collection. Or centuries by Rahul Dravid. Reliving that World Cup semi-final 2003. Whatever. And whenever. Without having to see a TV schedule.
To get to this level of choice to a viewer, one would have to convert tapes and films of entire movies or match coverages to digital files. And along the way, or after digitization, intelligently break down these files or 'mark' them with interesting 'chapter points'.
In India there are quite a few parties who own or have access to many thousands of hours of programmes - entire feature films, TV serials, cricket matches. And now want to convert them and make them available to users in the future via new media vehicles like Internet, Video over mobile, DTH TV (Tata Sky and DishTV), cable set-top boxes and other future 'pipelines'.
The challenge in repurposing media is to work out systems and infrastructure that can handle an initial load of thousands of hours while still being commercially viable handling the lower ongoing load over time.
Meaning you might need a certain number of systems and infrastructure to convert and classify, say, 10,000 hours of programming in a reasonable time of, say, 3 months. But after that is done, new programing may need to be converted at only 1000 hours or less per month. This earlier infrastructure is too large for that small a load leading to wastage. The trick is in creating an ideal size of setup that can do both.
The other issue is standards. What file format does one convert to? MPEG-2, H.264, VC-1, WMV, Flash...? Many formats exist. And the servers that handle and deliver these files also work with one or more of these standards. So the delivery vehicle and its standards dictates what the encoding is done as.
But typically, encoding to either of these formats is a one-way street. Meaning if one does the conversion, then changes the delivery architecture or provider, and a different standard is needed, then cross-conversion is not lossless and can even ruin the picture.
So, typically, converting to a low loss intermediate standard first then the final compression can permit some standards independence. Or, of course, one can convert to multiple standards and keep all available.
There are other challenges - secure storage, Digital Rights Management (DRM) or anti-piracy, restoration before repurposing, edit-ability of converted material. More on that sometime later.
Media repurposing consists of taking existing media - 35mm print/neg, DigiBeta/BetaSP/U-matic/VHS/DV tapes, an other media - and converting them to a format more friendly to new media vehicles. And playable on mobile phones, portable media players, or sent over a wire or wireless to homes and offices.
IPTV and VOD are two such new vehicles. IPTV is Internet Protocol Television and VOD is Video on Demand.
IPTV consists of sending video to viewers over the normal Internet. In India MTNL (and other services) show many cable TV channels over their broadband internet line - sent to homes over a telephone line.
VOD is when a viewer can go through a menu or index, and choose what he/she wants to see. That content is then played immediately, or downloaded over the Internet or wireless to view later.
Where IPTV and VOD differ from normal TV is that in normal TV many TV channels are being broadcast simultaneously all the time over a cable. In IPTV, only the channel the viewer wants to see is transmitted.
In VOD, only the actual program the viewer wants to see is beamed for as long as the viewer wants to see it. So VOD is sort of like watching a movie on a flight after selecting it from a menu. Like movies downloaded over the iTunes store for AppleTV. Or, in India Movies on Demand from Eros.
But Eros's Movie on Demand only work with PCs as it needs to detect Windows Media Player, which no Mac has. So if you have a MacBook, or iMac or any other Mac, forget Eros.
Other VOD examples are, of course, YouTube, or for really good-looking movies, Vimeo.
While repurposing media, it is also possible to break the programme down and classify it intelligently - like chapters in a DVD so viewers can see a portion of the entire programme. Like songs in a Hindi movie, or important scenes. Or like innings or overs in a cricket test match, ODI or T20 match.
How about going through a list of Hindi film songs and narrowing that down to Shahrukh Khan-Rani Mukherjee songs and selecting one you'd like to see. Or through a list of cricket matches and selecting one of Sachin's heart-breaking nineties collection. Or centuries by Rahul Dravid. Reliving that World Cup semi-final 2003. Whatever. And whenever. Without having to see a TV schedule.
To get to this level of choice to a viewer, one would have to convert tapes and films of entire movies or match coverages to digital files. And along the way, or after digitization, intelligently break down these files or 'mark' them with interesting 'chapter points'.
In India there are quite a few parties who own or have access to many thousands of hours of programmes - entire feature films, TV serials, cricket matches. And now want to convert them and make them available to users in the future via new media vehicles like Internet, Video over mobile, DTH TV (Tata Sky and DishTV), cable set-top boxes and other future 'pipelines'.
The challenge in repurposing media is to work out systems and infrastructure that can handle an initial load of thousands of hours while still being commercially viable handling the lower ongoing load over time.
Meaning you might need a certain number of systems and infrastructure to convert and classify, say, 10,000 hours of programming in a reasonable time of, say, 3 months. But after that is done, new programing may need to be converted at only 1000 hours or less per month. This earlier infrastructure is too large for that small a load leading to wastage. The trick is in creating an ideal size of setup that can do both.
The other issue is standards. What file format does one convert to? MPEG-2, H.264, VC-1, WMV, Flash...? Many formats exist. And the servers that handle and deliver these files also work with one or more of these standards. So the delivery vehicle and its standards dictates what the encoding is done as.
But typically, encoding to either of these formats is a one-way street. Meaning if one does the conversion, then changes the delivery architecture or provider, and a different standard is needed, then cross-conversion is not lossless and can even ruin the picture.
So, typically, converting to a low loss intermediate standard first then the final compression can permit some standards independence. Or, of course, one can convert to multiple standards and keep all available.
There are other challenges - secure storage, Digital Rights Management (DRM) or anti-piracy, restoration before repurposing, edit-ability of converted material. More on that sometime later.
Subscribe to:
Posts (Atom)