Instrumental Anime Project

Org Profile · Post by **Otohiko** » Sun Oct 17, 2004 9:46 pm

rose4emily wrote: If I have everything in before next Sunday, I should be able to produce a finished compilation for the upcoming Monday - and wouldn't that be sweet.

Wow, I think it's the first time we actually had a concrete final date. That indeed sounds great; hopefully we can get everyone still needing to submit stuff to pull together and get it all in by that time.

I also hope I get a decent connection by that time - but even if I don't, that FTP will probably come in handy (only problem is, it will probably take me half my current monthly limit and mayhaps a good two days to get the whole thing).

As for trackers... what alternatives are there? I'm not very familiar with how BitTorrent trackers work, actually - but I take it setting up our own isn't an option?

rose4emily · Post by **rose4emily** » Mon Oct 18, 2004 6:17 am

Setting up our own tracker might be an option, I'm not sure. I do know, however, that my server's been having issues lately (due mostly to the fact that it's also used as a workstation), and that those will need to be resolved before I can offer such continuous services as tracking a BitTorrent distribution.

---

I've completed the Intermission video - and it's now rendering on my laptop. I ended up doing a composited image thing, somewhat like the Instrumentality logo I made with the mic, textured background, and drawn "No" sign. This looks much better than the drawn concept did, thanks to the fact that the drawing looked either fuzzy or sketchy no matter how I tried to process it. I think I need to get back in touch with my pencil skills.

Anyhow, I took the reference image (Asuka and Rei in the elevator), scaled it to 512x384, added a transparent layer onto which I painted the pattern of the elevator walls, made screencaps of Miazawa and Kawashima in the most similar poses I could find, cleared off everything in those two screencaps other than the characters being put in the elevator, scaled and pasted the characters over the "walls" image, and then did a few alpha-channel tricks to get all of this to look good when layered over the background texture I've been using.

Then I came up with some "thought bubbles" to give the viewers something slightly more interesting to watch for two and half minutes than two people standing still in an elevator. The text is pretty lame, but at least it breaks up the monotony. Actually, given the monotony, the text seems pretty funny - kinda like anything said on the ninth hour of a road trip.

I've also recorded the widescreen narratives. The new recordings aren't entirely clean, but they're cleaner than the first set - and much more intelligable. I still have to trim off the dead air on the ends of each narrative track, and try a few FFT tricks to see if I can downplay the noise that is present a bit, but I think these are the ones I'm going with. Actually, they pretty much have to be - as I have to return the mics this morning and it seems my soundcard is the limiting factor on the SNR issue at this point anyhow.

I'm pretty tied up for the rest of today, but I'll probably get back to the Intro segment on Tuesday - so you can expect that to be showing up next.

rose4emily · Post by **rose4emily** » Mon Oct 18, 2004 3:45 pm

Okay, I've remastered my six narrative tracks as best as I can, and have actually found a fairly pleasent auditory space somewhere between thin, noisy, and boomy. I've also taken care to give each a short fade in and fade out to ease the transition from the professionally studio-engineered music tracks to our home-recording audio tracks.

I rendered the Intermission this morning and have just now posted it on the FTP sever under the "intermission" folder in the "outbox". It's labelled "intermission.avi" - to state the painfully obvious. The filesize is a bit bloated at 7 MiB for about ten images, all of which are very similar to one another, but that's actually about 80% audio, as the mp3 track is at 320 kbps, and it seems I have to re-install libmp3lame to re-encode it at the 128 kpbs being used for all audio in the final render.

If anyone has any better ideas for what to put in the text bubbles, it'd be easy enough for me to change what they say by laying new text over a copy of the "blank" thought bubble. I think I've said this already, but: the text is lame. It's almost supposed to be, as it's both a stupid in-joke and something inserted primarily for the sake of breaking up a two-and-a-half-minute still. Still, I imagine at least one of you is better at coming up with comic "thought dialogue" than I am.

downwithpants · Post by **downwithpants** » Mon Oct 18, 2004 5:32 pm

hmm for some reason the audio doesn't play unless i click on the trackbar at least 2 seconds after the start. i tried it in windows media player and media player classic.

mm, i thought you were going to show the credits of the fullscreen section during the intermission. or are they going to come before or after? well if we're not doing credits during the intermission, we'll try to think of more text for the bubbles. and about the bubbles, it just looked a bit weird how the bubble retracted after the the text was presented - i think it'd look more "natural" if the bubble just disappeared all at once after the text was presented.

as for the audio, if you want to make it sound like elevator music, try applying a high pass filter with a cutoff frequency 1.5-3 kHz to it, or using a filter to kill the bass. or if you want to keep it as standard full-sounding intermission music, it's fine as it is.

well, good to see such progress being made.

rose4emily · Post by **rose4emily** » Mon Oct 18, 2004 8:18 pm

Yeah - I tried putting the widescreen credits in the Intermission, but they looked pretty stupid there - especially when I started thinking about the fact that nearly everything they said was covered in the narratives, and that a most people would be seeing them again in the next file.

The retracting bubble thing is supposed to look weird and unnatural. I can get rid of it, though, and replace it with a fade if I get a consensus that more of you hate it than like it.

I don't know about more text. Adding more text would ruin the intentional understimulation. I do think different text might be in order. I've been playing with a couple of ideas on that myslef, but haven't come up with anything spectacular yet. I'm also not so sure about the empty bubble at the end. It was supposed to be something on the idea of starting to say something to break an awkward moment of silence - only to realize you've nothing to say. I'm not sure if it really gets that across, though.

I like the high-pass idea. I'll try it to see how it looks and sounds.

There's about a second of silence at the beinning of the song, but not two. I do have to encode it at the 128 kbps with LAME to include it in the final render, though, so it will be ecoded in exactly the same manner as the audio for all of the other segments at that point - which should eliminate any playback problems it might be causing on the Windows machines.

---

BTW - Only at an engineering school:

Just before recording the narratives, my rommate and I were on the floor laughing for about ten minutes - after he tried to "drag and drop" a file onto a Unix console running the text-based FTP client.

The sad part is that he actually started using Linux before I did, but has been using a Mac for the past few years.

Appearently the OS X user interface has the strange side effect of making its users believe everything will respond to anything else being dragged onto it in some interesting way.

Songbird21 · Post by **Songbird21** » Mon Oct 18, 2004 10:50 pm

Rose4- Which narratives do you want me to redo?

downwithpants · Post by **downwithpants** » Tue Oct 19, 2004 12:55 am

hm k, here are suggestions i came up with for the intermission

1. fade the video in after the guitar struts so that the bass enters by the time the word "intermission" fades in.
2. The total length could some trimming, to at least under 2 minutes.
3. Should there be dialogue if they're both thinking? also, dialogue has timing constraints when you're trying to make it appear natural- like a response has to come quickly after a prompt

here's my stab at a script:

Pause
Miyazawa: "How long are we supposed to stand still like this?"
Pause
Kawashima: "Something about all of this seems strangely familiar"
Pause
Miyazawa: "My legs are starting to cramp."
Pause
Kawashima: "Where have I heard this music?"
Pause
Miyazawa: "..." or perhaps if you can find/draw a sweatdrop image, place it on her head.
Pause
Miyazawa: "Oh good, it's fading out"
Fade Out

pen-pen2002 · Post by **pen-pen2002** » Tue Oct 19, 2004 12:56 pm

downwithpants wrote: Pause
Miyazawa: "How long are we supposed to stand still like this?"
Pause
Kawashima: "Something about all of this seems strangely familiar"
Pause
Miyazawa: "My legs are starting to cramp."
Pause
Kawashima: "Where have I heard this music?"
Pause
Miyazawa: "..." or perhaps if you can find/draw a sweatdrop image, place it on her head.
Pause
Miyazawa: "Oh good, it's fading out"
Fade Out

I love it.

How long is it supposed to run for?

Bakadeshi [AuN Studios] · Post by **Bakadeshi [AuN Studios]** » Wed Oct 20, 2004 12:43 pm

Just updating you guys to tell Rose that I stoped procrastinating and finally uploaded my picture and screencaps to the ftp.

let me know if theres any problems or if you need anything else.

rose4emily · Post by **rose4emily** » Wed Oct 20, 2004 5:31 pm

I'll try out all the Intermission modifications this weekend. They sound good, I just need one or two free hours to figure out how/whether all of these changes will work together and render the new and improved version.

Songbird - I need all of the fullscreen narratives I don't have already, and could use a re-do of the Debussy narrative. If you have versions of the others with a little dead-air padding on either end of the audio, that could be helpful too - as I need a little room to blend the recordings into the silence on either end of them to make sure the change in ambient sound quality isn't too abrupt.

Sorry, but the intro still isn't much closer to being done than it was at the beginning of the week. You can thank my classwork for that. I have fully settled on using a shortened version of the original audio track (a two-minute intro is too much), and I've pulled together a much smaller subset of all of the images I originally collected - selecting the most relevant ones and giving me a little more room to do something other than just flash a million pictures at the viewer. I am, after all, trying to make a good simple intro - not a really crappy imitation of one of Ermac's videos.

I'm afraid I'm also going to have to scrap the idea of a hand-animated widescreen intro. As it turns out, my next animation assignment (for which I still need to make about a few dozen more drawings), has to have prominent color, concrete characters, and a plot (and all in 20 seconds). Consequently, I can't do an abstract peice on music - and am instead animating a solid black, white, and red depiction of Macbeth freaking out over the blood on his hands and how it'd turn the whole ocean red should he try to wash it out. Interesting, but hardly what this project needs. The one after that has to be done in either sand or wet paint being pushed around a glass pane, so I don't think that'd be any good either. Unfortunately, I don't have time to make a hand animation independent of those assignements without turning them into absolute crap.

Well, I have to produce some drawings now, and then scan them (the animation camera's have some sort of automatic lighting compensation in the capture software that inevitably produces poorly and inconsistently saturated frames), and then edit them to threshold everything into my three desired colors, so I've got to break off now. I just thought I should check in to say what's going on.