vid commentary: vidding process
ghost_lingering had a couple of general questions about my vidding process, including the arcane and mysterious "clip databases" to which I occasionally refer.
When you make clip databases for a show, how do you organize the clips? By episode? By character?
When I make a database, I don't yet have any clips; all I'm doing is collecting information about potential clips. I rewatch the show, usually with the sound turned off, and I enter notes about viddable shots into an Access file with the following fields:
episode number
episode title
timestamp
characters
scene (location)
external motion
internal motion
vid (this is where I mark any shots I'm sure I want to clip)
description (my notes about expressions, mood, context, etc.)
Once I've rewatched the show and made the database, I go back to my printout of the lyrics and do some further brainstorming. So, for example, if there's a line where I know I want a shot of Mal alone and sad, and I can think of a couple of obvious ones but I know there are more and I have no idea what episode they're in, I filter the database to clips of Mal and scan through the descriptions of those shots, marking additional options for that line in the database (so I'll remember to clip them) and on the printout (so I'll remember what I thought I wanted to use that clip for).
After I've done the initial clipping, I also use the database to fill in gaps on the timeline: If I need a clip with camera motion circling around Fraser, or a clip of Claire smiling, or a shot of Mal moving from left to right across the frame, I sort and/or filter the database (by character, by external motion, by internal motion, whatever) in order to find the clip I want (or to ascertain that the clip I think I remember doesn't actually exist).
I've uploaded a few screenshots of the Heroes clip database if you're curious: the first 33 entries for the pilot, the database filtered to Claire and sorted first by external and then by internal motion, and the whole database sorted by external motion to find all shots with the camera circling left to right.
When you make the clips do you dump, say, the entire episode into Premiere and make subclips or do you literally make a new file with the bit of the episode you wanted clipped? ...mostly I just want to know how you organize them/label them/etc.
For the last few vids, I've made AviSynth "wrappers" for groups of episodes and dumped those groups of episodes into Premiere. So, for example, I'd rip the due South DVDs, index an entire DVD's worth of eps, and have one .avs file for five or six episodes. (I also stick a bunch of filters in the .avs file to clean up the source; some shows obviously require more filters than others.) Then, once I have the big multi-episode groups in Premiere, I load each .avs into the Premiere monitor window, one file at a time, and scroll through to find the moments I need (as marked in the database) and set the in and out points to make "virtual clips," which I pull into bins and label in various ways depending on how I'm thinking about the vid.
The Heroes vid project file has more than a dozen bins with character names, because I sorted primarily by character; each bin would then have anywhere from three to several dozen clips, each one labeled with the character name and something to remind me of the content: "Bennet rubs eyes, tired" or whatever. That same project file also has a few bins that are more thematic or emotional and that contain a mix of characters; I know there's one called "all the love in the world," and another one called "chaos" where I was collecting clips that felt like they might go in the second instrumental bridge, and there's one just for clips of Isaac's paintings. The due South vids, on the other hand, focused on just two characters, so the bins are more specific: "Fraser alone," "Fraser sad," "Fraser walking," "RayK cranky," "RayK dorky," "F/K tense," "F/K connecting"... you get the idea.
For my first due South vid (Out Here), I actually used a mix of virtual clips and "real" new-file clips. The reason for this was that I knew I wanted to have access to all of the third and fourth seasons (because the vid focused on Fraser's relationship with RayK rather than with RayV), but I also wanted to use a few isolated bits from the first two seasons - establishing shots of Chicago, a few clips of Fraser alone in the city - without having to have the entire seasons, or even entire eps, taking up room on my hard drive. So I dumped S3-4 into Premieres via .avs and did virtual clips; and then I ripped the few relevant eps from S1-2, opened them in VirtualDub, saved a few tiny snippets as new separate files (uncompressed .avi format), and deleted the original .vob files. When I imported the snippet files into Premiere, I sorted them into bins like "Chicago est. shots" and "Fraser alone," just like the virtual clips.
I'm sure that this entire system - both the database stuff and the .avs tech - sounds positively Byzantine to anybody who hasn't done it or doesn't think this way. (I used to find the .avs system impossible to wrap my head around until I quit trying to understand it in the abstract and started messing around with it in practice.) All I can say is that it works for me, and that for anything longer than a movie I really need the clip database in order to vid both well and relatively efficiently; my visual memory is simply too intermittent to be relied upon. Having an external repository of metadata about visuals frees me up to worry about other things, such as the eternal "WHY IS THIS MY HOBBY" question.
When you make clip databases for a show, how do you organize the clips? By episode? By character?
When I make a database, I don't yet have any clips; all I'm doing is collecting information about potential clips. I rewatch the show, usually with the sound turned off, and I enter notes about viddable shots into an Access file with the following fields:
episode number
episode title
timestamp
characters
scene (location)
external motion
internal motion
vid (this is where I mark any shots I'm sure I want to clip)
description (my notes about expressions, mood, context, etc.)
Once I've rewatched the show and made the database, I go back to my printout of the lyrics and do some further brainstorming. So, for example, if there's a line where I know I want a shot of Mal alone and sad, and I can think of a couple of obvious ones but I know there are more and I have no idea what episode they're in, I filter the database to clips of Mal and scan through the descriptions of those shots, marking additional options for that line in the database (so I'll remember to clip them) and on the printout (so I'll remember what I thought I wanted to use that clip for).
After I've done the initial clipping, I also use the database to fill in gaps on the timeline: If I need a clip with camera motion circling around Fraser, or a clip of Claire smiling, or a shot of Mal moving from left to right across the frame, I sort and/or filter the database (by character, by external motion, by internal motion, whatever) in order to find the clip I want (or to ascertain that the clip I think I remember doesn't actually exist).
I've uploaded a few screenshots of the Heroes clip database if you're curious: the first 33 entries for the pilot, the database filtered to Claire and sorted first by external and then by internal motion, and the whole database sorted by external motion to find all shots with the camera circling left to right.
When you make the clips do you dump, say, the entire episode into Premiere and make subclips or do you literally make a new file with the bit of the episode you wanted clipped? ...mostly I just want to know how you organize them/label them/etc.
For the last few vids, I've made AviSynth "wrappers" for groups of episodes and dumped those groups of episodes into Premiere. So, for example, I'd rip the due South DVDs, index an entire DVD's worth of eps, and have one .avs file for five or six episodes. (I also stick a bunch of filters in the .avs file to clean up the source; some shows obviously require more filters than others.) Then, once I have the big multi-episode groups in Premiere, I load each .avs into the Premiere monitor window, one file at a time, and scroll through to find the moments I need (as marked in the database) and set the in and out points to make "virtual clips," which I pull into bins and label in various ways depending on how I'm thinking about the vid.
The Heroes vid project file has more than a dozen bins with character names, because I sorted primarily by character; each bin would then have anywhere from three to several dozen clips, each one labeled with the character name and something to remind me of the content: "Bennet rubs eyes, tired" or whatever. That same project file also has a few bins that are more thematic or emotional and that contain a mix of characters; I know there's one called "all the love in the world," and another one called "chaos" where I was collecting clips that felt like they might go in the second instrumental bridge, and there's one just for clips of Isaac's paintings. The due South vids, on the other hand, focused on just two characters, so the bins are more specific: "Fraser alone," "Fraser sad," "Fraser walking," "RayK cranky," "RayK dorky," "F/K tense," "F/K connecting"... you get the idea.
For my first due South vid (Out Here), I actually used a mix of virtual clips and "real" new-file clips. The reason for this was that I knew I wanted to have access to all of the third and fourth seasons (because the vid focused on Fraser's relationship with RayK rather than with RayV), but I also wanted to use a few isolated bits from the first two seasons - establishing shots of Chicago, a few clips of Fraser alone in the city - without having to have the entire seasons, or even entire eps, taking up room on my hard drive. So I dumped S3-4 into Premieres via .avs and did virtual clips; and then I ripped the few relevant eps from S1-2, opened them in VirtualDub, saved a few tiny snippets as new separate files (uncompressed .avi format), and deleted the original .vob files. When I imported the snippet files into Premiere, I sorted them into bins like "Chicago est. shots" and "Fraser alone," just like the virtual clips.
I'm sure that this entire system - both the database stuff and the .avs tech - sounds positively Byzantine to anybody who hasn't done it or doesn't think this way. (I used to find the .avs system impossible to wrap my head around until I quit trying to understand it in the abstract and started messing around with it in practice.) All I can say is that it works for me, and that for anything longer than a movie I really need the clip database in order to vid both well and relatively efficiently; my visual memory is simply too intermittent to be relied upon. Having an external repository of metadata about visuals frees me up to worry about other things, such as the eternal "WHY IS THIS MY HOBBY" question.