Instrumental Anime Project

NeoQuixotic · Post by **NeoQuixotic** » Mon Oct 11, 2004 9:21 pm

Just posting to say that I put my video screencap and the picture of myself on the FTP.

rose- If you don't think my screencap is unique enough or the picture of myself is bad, just post here or e-mail. The current picture I have of myself is sort of redish. I tried like 20 pictures with my digital camera and that was the best I could do with low lighting indoors for now. I'll probably try some more.

rose4emily · Post by **rose4emily** » Mon Oct 11, 2004 10:17 pm

downwithpants wrote:hey song, before you redo mine, it's bergamasque, not bergmanasque. i think it's pronounced bare-zhe-MASK where "zh" is the soft z sound, like in "vision". according to dictionary.com debussy is pronounced duh-byoo-SEE (like: dub you see). If our canadians who know french know how to pronounce bergamasque, they'd probably have a better guess than me.

Looks like maybe I, too, have been mispronouncing two of those words for a while now (and misspelling one of them as well - my bad). Actually, I have heard "Debussy" pronounced quite frequently in both the manner I suggested and the one DownWithPants looked up. Maybe it's just an emphasis thing - first syllable, second syllable, that varies between French and Canadian French, or English and American English, or something like that. Maybe, though, it's like the debate between "nuclear" and "nucular" - where one version is definately the correct one, but another still maintains popularity. Supposably.

I think Bush's speechwriters should tell him to just say "atomic" instead. It's pretty hard to screw up a word like "atomic".

---

I've just picked up Jasper-Isis's two video pictures that I missed earlier, and have added borders to the ones submitted by Anubisx00, so I now have all of the end credit images I need from:

anubisx00
jasper-isis
otohiko
rose4emily

I will send PMs/emails to all of the members who I haven't seen posts from in the last few pages to give them a heads-up.

---

Anubis - The red balance wasn't actually too high in your portrait, but the color saturation was far too strong. I'm guessing you shot it indoors, without a flash, with (realtive to sunlight) dim illumination from an incandescent bulb. Those conditions are common among photographs, and tend to produce reddish-looking results. No problem, though, as I just had to drop the color saturation to a level proportionate to the lighting saturation and contrast, and it now looks fine.

I also upped the saturation on your video capture by just a little bit to compensate for the washed-out look of the R1 Evangelion DVDs (which were appearently encoded off of the television stock, rather than directly form the cells - hence the Platinum edition now being released). This helped to make the crimsony goodness of the LCL sea just a little more vivid, without ruining the appearence of the slightly greenish crosses.

rose4emily · Post by **rose4emily** » Tue Oct 12, 2004 6:57 am

I've finished my "missing picture" head. I think you all will like it in all its "Serial Experiments Azumanga" glory. I figure I'd need one even if I do get all of your pictures, to continue the "picture-wiping" transition for the "other credits" portion of the end sequence, so I put a little time into making it a really nice "missing picture" head.

rose4emily · Post by **rose4emily** » Tue Oct 12, 2004 9:12 am

Okay... so, like, I'm an idiot, and, umm, I accidentally deleted the original narrative screencaps for several of the fullscreen videos that I made "set" images for with the first "set" design, and I think there are a couple of videos for which I never got images in the first place, so I'll just list the ones I'm missing and haven't been able to find in my old e-mails and scattered "temp" folders:

Tomoe in the Moonlight (I have little versions, inside the set TV, but not the originals)
A Boy I Know (I have little versions, inside the set TV, but not the originals)
Simplicity
Ararat

Orotario of Ghibli
Requiem for a Nightmare

I'll keep looking for the originals, as I still have two external hard drives to search, but I thought I should say something as soon as I realized this had happened to give you a chance to look for them on your computers. If anyone still has their own images - or anyone else's images, please upload them on the FTP so I can fill in the empty spaces.

---

That was the bad news, now the good news:

I have settled upon a narrative "set" for Yukino's fullscreen section. It's on the simplistic end of things, but I think it looks much better than everything I tried in the way of including her in a photographed or simulated full news set with desks and chairs and the like.

http://www.thewired.info/instrumentality/images/01.png

and, with her mouth open (actually her mouth from a different picture, rotated and repainted to make the skin tones match):

http://www.thewired.into/instrumentality/images/02.png

Here's the Osaka head:

http://www.thewired.info/instrumentality/images/03.png

The new Instrumentality Logo (a sort of backward spin on a common theme use by A Capella groups).

http://www.thewired.info/instrumentalit ... tality.png

The new logo in it's natural habitat, Yukino's screen

http://www.thewired.info/instrumentality/images/04.png

I know, the text isn't exatcly readable in the shrunken version, but the full-size one will have already been shown in the intro.

Org Profile · Post by **Otohiko** » Tue Oct 12, 2004 10:42 am

Hmm, pretty crazy - again, I could only see the folder structure of the FTP by uploading a file into the root folder (which, again, is not supposed to happen).

If you see an empty text file there, that's my handiwork.

Uploading images for Ararat... (luckily, unlike the HuffYUV of the video itself, I actually archived those)

Also, checking out a few of the example pictures and such on the FTP - looks quite interesting.

rose4emily · Post by **rose4emily** » Tue Oct 12, 2004 4:08 pm

Thank you. I really am trying to do something special with this to make it look like a finished product that was almost a year in the making, rather than a haphazard conglomeration of disparate parts that were almost a year in the making.

Since I don't yet have any narrative audio for the fullscreen tracks I do have images for, and I don't yet have any images for the fullscreen tracks I do have narrative audio for, I'm changing plans a bit and finishing the intro next, rather than an example of the narratives.

They'll be pretty static anyhow - lipsynch a couple of words while panning into the "screen" and pushing our lovely anchor Songbird / Yukino Miazawa out of sight before the static-head pose lipsych gets old (and excessively time-consuming on the production end).

A similar setup will be used for Kawashima's (my) set, except the picture will be letterboxed and it will feature Kawashima rather than Miazawa. The transition between the two will actually be done using a clever little idea I had to start Kawashima's set with his screen showing a 16:9 version of Miazawa in her set (which will be expanded to include the big news camera kept neatly out of sight in the 4:3 version).

downwithpants · Post by **downwithpants** » Tue Oct 12, 2004 6:07 pm

k, just uploaded my set pictures, i think. i'll get my end credit and self-photo images in soon.

the pictures you've created look good. just wondering if you can get three positions of yukinon's and kawashima's mouths: one closed, one halfway open, and one all the way open. lip synch looks much more convincing that way. but if that's too much trouble, don't worry about it.

jasper-isis · Post by **jasper-isis** » Tue Oct 12, 2004 7:53 pm

Rose4: the set pictures seem to be coming along nicely. However, one thing that really startled me was the "missing person" picture, as Osaka looks just a bit on the creepy side. Maybe you can fill in her pupils and lighten her skin a little? Either that, or make her entire body semitransparent too, not just her skin.

I'm thinking that Osaka doesn't exactly have the appearance of a generic human being.

Hopefully we won't be needing a missing persons picture anyway.

rose4emily · Post by **rose4emily** » Tue Oct 12, 2004 10:41 pm

Aw, Osaka's face-only-transparency was deliberate - the first version actually had no face at all, but that didn't look so good. So I added the face, which clashed with the color tone of the background behind it and just looked kind of "bleh". Then I did a first-derivative operation on the face to reduce it to just lines, and left the rest as a slightly lightened version of the underlying background. It does make her look sort of ghost-like, but that was intended.

As to why I used Osaka: I tried a plain silhouette head (the typical bald guy in profile), which was pretty boring (even with the background), and then I thought it might be amusing to do a silhouette of some character that was blatently recognizable in silhouette. So my second idea was to use Chiyo, with her yam-shaped pigtails. Then it occured to me that Chiyo would never forget to submit a picture - but Osaka certainly would. So I used Osaka. Not very generic, but interesting. I am, however, considering a Spike Spiegal version (once again, easily recognizable hair, and I could see him neglecting to provide evidence of his identity) for male missing persons. Hopefully there won't be any, but there are people other than the project members to credit and I need something to fill in for them as well.

I haven't yet found a good half-closed Yukino mouth, but I do plan to come up with one and make a third image for the narrative set featuring the half-closed mouth. The particular set of Kawashima images I plan to use have relatively subtle talking action for an anime character, and really don't look bad with just 'opened' and 'closed'. I might change my mind, and find a scene where he more fully opens his mouth, but I'm not going to worry about that right now.

I am now dusting off my Cinelerra manual and looking into the translation keyframing I'll need for portions of the intro. Shouldn't take me long, but I haven't edited video in a while and need a refresher.

rose4emily · Post by **rose4emily** » Wed Oct 13, 2004 10:28 am

I've created three versions of the "A Boy I Knew" narrative audio where I combined the separate segments uploaded by Songbird. One is a "clean" mix, with only a moderate amount of filtering to downplay the high-end noise and occasional low-frequency transients that inevitably occur when using a non-shockmounted mic. The other two apply two versions of a short-decay reverb that actually does a much better job of masking noise than actual noise removal. The second reverb, however, seems a bit on the heavy end to me - so I'd say my favorite is the "light echo" mix.

All that having been said, it is best to let the mic run for a few seconds before and after the actual audio, so I can cross-mix the "room noise" (the ambient sounds coming off of the mic's circutry, as well as the ambients sounds occuring in the recording space) between segments rather than having to separate them with silence (which stands out like the ATMs at a Rennaisance fair). Oddly enough, noise is most audible at the moment it abruptly ceases to exist.

A couple of the segments are actually clipped so close that the last word is being cut short. This is the case for the first of the three clips forming the "A Boy I Knew" narrative. I originally thought that was just a matter of my media player cutting the stream short on playback, but my editing software does the same thing, indicating that it's a matter of the stream actually ending in the midst of its last syllable.

Songbird: if you have the "raw" recordings of these segments - with noise buffers from the time between switching the mic on, starting to speak, and turning the mic off - I will be able to mix them in a manner much more becoming to the quality of your speach. I can also play with the filtering a bit more if you'd like to hear more of one thing or less of the other - as there are a million different ways to "clean" and "thicken" (or, as my middle-school band teacher used to say, "fatten" (maybe it was "phatten", i'm not sure)) an audio stream without going over the top with a bunch of obnoxious "chorus" effects.