Skip to content


Subversion checkout URL

You can clone with
Download ZIP
tree: 1b7684c09b
Fetching contributors…

Cannot retrieve contributors at this time

179 lines (175 sloc) 17.04 kB
=== patentzy yes actually do vaguely ordered ===
blu-ray OCR
more blu-ray stuff: can ID them, can playback (through...mkv? VLC? powerdvd OCr/scripted?)
some blu-ray player descriptor somehow
accomodate splits vs. blanks and work
it can "resync" based on lipsync'ing or some user input (back, forward buttons to subtly resync edits with a currently playing back video) or
with user input "tell me where this subtitle is, for you, and also this one" to synchronize it (+- auto-playing to "around" that level so they can check more easily),
or you can ask them manually for time entry for the same.
OCR'ing subtitles on the fly (or, cache'd OCR'ing them on the fly, to give it time to process) to synchronize or check for profanity or other "questionable" content
post a demo of it playing back blu-ray edited local computer, +- projector, +- playon (?)
can infer a rating based on (looking it upon imdb, letting people vote, etc.)
I can forsee some browser window that pulls up some sweet player [mod vlc plugin?] somehow something quality anyway, to play local media, or remote
or same thing but for upconverting...but how?
bounty NACL VLC :P
could add dragging since we already control the mouse up and down, and movements etc., for example to skip forward etc.
a little robot demo thingy that can "click" your remote (or hardware) to fast forward through certain scenes, et al
EDL can say "this exact text was at this exact time" to allow for matching with other videos of same content that may have slightly different timestamps/speeds/start time etc.
with several, or start one, end one, etc.
allow for extra start time offset for videos, if netflix instant has any
a way to integrate with cable TV, ex: "this show we know started at 1:01, and its offsets should be this after that point"
make it so an EDL can be "split" (or subtitles split, same way) so that it can span across multiple youtube videos that have been divided in some way.
it can query the currently open browser window, check to see which movie it's playing, and auotmatically spawn off a correct player instance with EDL for that movie
based on title or url.
when playing a movie, it identifies by colors/tones the movie or the scene, to edit accordingly
when there's a profanity mute, to display some extra text on the screen like "and I really want you to be happy!" I guess it could just be like a transparent overlay on the movie
netflix "try this srt, does it match?" no "what about this one, now?"
it can "infer" where the thumb slider would be, and "drag" it to where it infers the new timestamp location should likely be, for skipping et al.
people can "vote" on certain sections, if they're profanity or not, should be skipped or not
overlay one video on another
I can imagine a system that scans incoming audio, does voice recognition sofware analysis on it, search for (angry tones, profanities, "questionable" tones/text like making love), and excludes them from output [search forward, skip forward, just display nothing, either one, mute audio, decrease audio et al].
make an API web service that can search for EDL's
or with youtube, use the api to fast forwarrd there
and if it isn't already loaded, display an extra notice "forwarding" or what not.
and/or display some other overlay and or audio in place of the video or audio, or supplementing it.
a video of "recording to broadcast it" (or save it one by one)
can parse a screenplay et al to look for violence/profanity/sex
can parse/detect "innuendo" based on natural language semantics, from subtitles et al.
can "suck in" a large list of EDL's from imdb at build time,
and/or a "resync" button, to re-download just the list of updated EDL's/IMDB wiki settings [?]
can handle multiple dvdid's et al (same EDL) (mostly done already)
can "obscure" full frames [sections of frames] and/or "partial" frames or sections of frames (ex: lips, nudity)
can add extra overlays of text (personalized subtitles, like)
some tie in with blog posts
some tie in with google +/facebook for community viewing/editing/realtime collaboration/editing/vieweing.
track the audio noise level, use that to auto-scan for violence
use the descriptive captioing audio to discover violent scenes (she says "he hits her" or what not), or for other questionable scenes.
ratings system, where users can rate based on various criteria/statistics, each EDL, and it aggregates them to display/obtain/use a sum rating metric
can 'remove' or 'obscure' CC or subtitles deemed offensive
it can pull in a DVD based just on the dvd-id "query id, not local? go get it from online to see if the list has been updated since"
control VLC RC to overlay a logo dynamically
can control volume *and* play other audio
can permanently black out certain coords to keep things al
remove black when they really move the mouse?
a transport schema talking from a sensible cinema database "to" a player which uses it, and/or has selection options for levels of (content category) or can pick options they want, stored in sensible cinema database or in the "player" side, with transfer/communication between the two. Ex: XML as a transport, oauth for authentication, etc. etc.
or they could transmit a time signature "across the wire" somehow (ex: every second), which could be used to react to time events (skip, mute) appropriately. Or allow for programatic callbacks "call me at this time" or "at this time do this action" or the like, and/or accomodate for buffers/freezing etc. [kind of already demo'ed this one]
can add to to supplement EDL + EDL timestamps
can pick specific profanities to avoid (I think I demo'ed this already actually)
it can scan "playing" instant streaming titles, and/or OCR or take the audio and look for words, to check for profanity.
it can examine "output volume level" to auto-detect violence scenes.
similarly for "questionable" content
it can detect from subtitles (and/or OCR, same thing), questionable content, or violence, based on words used
use multiple filters/complex mapping graphs to overlay just the exactly correct overlay: or AviSynth scripts
user preference "how many down clicks on volume for profanity" (already possible)
space -> advance from pause until next edit item?
can jerk mouse only every so often (cheap would be to just arbitrarily have it from x to y, demo that)
can scan several player descriptors to "pick the right one" (kind of already showed/did this one)
post some demo screencast...this should prove all shouldn't it [?]
can "control mouse" to drag it to change location (ex: "move down 10 pixels at beginning of movie" LOL)
beep at them/warn when not tracking [?]
default on...I think.
can jerk mouse more often closer to failure sections...
and/or resize the playing/streaming window to match it when the mouse "has been moved recently" +- by us, or not.
can pull one from imdb at all, video "show it creates a file"
C DVD player: minimize instead of skip :P
don't mute, just almost mute LOL or almost mute and static-ify
or combine the two
can suck it in "in realtime" from online, and parse it +- from a user editable wiki
a UI to select categories, save as preference, apply it each time at DVD selection time
don't use timecodes, use "% of the movie" and "total length" perhaps?
look at patents for other ideas? [+- nissim p's?]
just pause it, tell them how far to fast forward at that point :P [somewhat already done/showed actually, thanks to mplayer]
have the user forced to tell me what each digit is, themselves, at runtime, for OCR :P
the online edl stuff itself (its todo list)
javascript API
phase 1: done
phase 2: I can accomodate for pausing, too.
phase 3: I can re-train based on events of stoppage (buffering et al)
phase 4: I can constantly poll and react appropriately to current timestamp
post video demo of it streaming to a console device
maybe could just "play from x to y" [no exclude info?] instead of "here are the bad parts"
stage one: query, auto-download when they hit a netflix movie of some type
stage two: chrome plugin for the same, auto-detect, auto-user, auto-download, auto-watch edited :)
can insert deleted scenes et al [even external scenes :P] (already showed it able to even do this somewhat)
can overlay with an alpha transparent other video, which video is sped up from its original form.
or which video is itself edited for content, or displayed with realtime content editing/edited movies.
it can overlay with enhanced/modified "other video image" files.
it can overlay audiobooks
it can overlay it with edited audiobooks/sped up/modified somehow audiobooks
it can "enhance" movies, like sharpen them, magnify, zoom, et al
it can jump/insert a pre-recorded"clean" version of a certain scene.
many others were already completed and done from this list and removed
there are some more ideas in the TODO file itself, and/or on the blog sites where I have posted other ideas too
fan editor, that can jump positions within the same DVD/insert audio overlays/insert videos et al
voice recognition for detection of profanity/violence/etc. etc.
detect subtitles for "what appears to be suggestive" content or lewd scenes, etc., based on the text, +- natural language processing
overlay with transparent movies (other movies)
"double check synchronize" button for making sure the edits line up, with or without text saved of subtitles, et al
auto-detect "dark scenes" or "skin tones" or "noises" or the like
can "fast forward" through scary/otherwise questionable scenes
can auto-compute/compare md5's (using any arbitrary hash function)/truncated md5's/streamline (faster) partial md5's/can md5 specific frames to detect times/to lookup correct EDL's on the fly.
can use the same or some other mnemonic to download something or to go to some online url for it/related to it.
can auto-lookup and display things like images/ratings/user reviews, dynamically, about a piece
can "auto-sync" "this EDL maps to this movie in this other player (hulu, for instance) which has synchronization points x,y, so it adjust the
timestamps in the EDL accordingly, for that one .
can work with compressed audio/video, or uncompressed
can work with encryped data/image/audio data, or unencrypted, with or without 3rd party player/codecs etc.
can have its "own" metadata stored off somewhere (like user contributed images/photos/comments/blogs/wikis)
can link videos to adds/adwords/add providers somehow/voice over/overlay adds, et al.
can have overlay (+- transparent) of adds/off-screen/below the video, always, or every so often, based on timestamp (fixed, or specific to a video)
can have overlay text (ex: news ticker, adds, chats, comments from users) on a video.
integrate with google plus (+- api) somehow, and/or broadcast [?]/realtime "chat" while watching, voice/text, etc.
google TV too, and apple TV, as an app., etc. etc. (auto-download EDL's etc)
machine learning for video analysis for synchronization/detection of "questionable scenes/audio" [?]
can "blank out" or "obscure" certain parts of the screen based on skin color, or region, or region with vectors (moving region), based on timestamp, or pixel color, or hue, or YUV, or combination of the two ("this color in this region, within these time beginnings and endingses")
can OCR subtitles from DVD's on the fly, to screen them like tvguardian now does. With or without in an embedded unit.
Everything else mentioned here, but in an embedded unit/side by side with other things/as an app in a console device.
people can send you a file and you send it back edited to them LOL (file meaning a rip of some DVD they own, +- requiring them to send in an original DVD too LOL.
see also business ideas.txt file
realtime chat while viewing/editing/use google wave
being able to coordinate/synchronize playing 2 dvd's when you're away
can display "user contributed" subtitles (edited, fan edit style, etc.)
can overlay them "transparently" above various players, arbitrarily, by tracking timestamps on screen, etc.
while synchronized, they can chat, like in a video game LOL
can overlay an srt file to an "online" OCR tracked player
right click on DVD "play edit with ..."
bookmarks within edited playback/remember where you last left off
you can record the screen of an "edited playback" in order to create an edited version. If legal.
each EDL can have a "type" specified, like "type a starts at time x, ends at time y" "type b starts at time z, ends at time mu" so that you can re-use the same EDL
cross several media types.
EDL action "zoom" to a certain part/portion of a movie, for a specified amount of time.
during a "tense" scene it can "darken" the scene more, or blur it, or shrink it to a fraction of its original size, or make it faster,
or turn down volume, or mute. User preferences to "respond to EDl action type x with reaction y" (where y is blur, mute, skip)
have jump/jump to file to be able to insert alternate endings.
an avisynth script to "cut" the file to be able to cut out/mute out sections
an avisynth script to "cut" the file to be able to watch it playback edited without modifying the original. +- DVD input too.
split EDL's based on DVD "title"
have some menu system for "choose, now, which path you'd like to take" (ex: which alternate ending)
when parsing a subtitle file, can specify an "extra buffer" to put around all of the subsequent mutes that it calculates for you, based on timestamp.
scan subtitles for text that "accompanies" violent or sex scenes, edit around those automatically.s
when playing they are presented with a list of "groups" of items that can be added to their EDL.
One for profanity, one for boredome, etc. and can pick and choose, mix and match the ones that they will apply for that viewing.
It then adds them together and uses them. lists some useful ideas
Create a system that "receives" for instance netflix instant streams, or DVD streams, or screen casts
1) edits them, "sends" them back to a client receiver. with or without OCR'ing elements of the screen.
it can "automatically synchronize" with a list of EDL files contained on some "other server somewhere"
basically, when it starts, it runs a git update, which brings its local list of EDL's up to date, so it can see any changes.
This also allows for p2p EDL sharing, since you can have git repositories anywhere, and pull a list of them.
Another option would be to "always" pull from a/the server when you lookup EDL's for a certain media object (DVD, file, online, etc.)
then you're always getting the latest/greatest.
it can have multiple "synchronization points" for various different medias, for example specifying that "in netflix instant, this particular subtitle starts here, where as with dvd's, it starts here, and this ending subtitle is here with netflix instant, here with other media"
it can have multiple "synchronization points" for coordinating subtitles with the movie, like a beginning, middle, ending point, for better incoordination.
it can "save" a movie using upscaling/upconversion techniques, so that when it is played back later it can be played back looking good, without requiring more cpu to upscale it "on the fly" at that time.
"user prompt each time" for sections, like "would you like to skip this section this time around?" per EDL
upconvert the audio somehow, like adding bass on its way out :)
it can "insert" video/videos
it can "loop" just once or always, when it hits a certain timestamp (ex: just loop once to insert a video)
you can specify an image to "overlay" at "exactly this frame" then a different one for "exactly this one..."
allow for interaction among users watching a video together (but on separate devices), for instance chat, synchronization,
they can "circle" things in the playback/insert their own overlays/their own dynamic overlays/their own programmable
filters that do something to the video before displaying it/interact with ads presented during video playback or
"while" video is playing back. Playing games together while the movie is playing. Being able to pause and then "color over"
parts of the frame, etc.
it can display a screen/wiki editable something that accompanies a film while it is being viewed, a la:
including online chats with others, etc.
Jump to Line
Something went wrong with that request. Please try again.