OT: Totally P!ssed off at UPN

[Jack Simth]

geoschmo said:
I am hopeful that as the computer technology advances and comes down in price, we can break out of this cycle and some sci-fi series that don't get spectacular ratings right off the bat can last long enough to develop an audience. I'd like to see some variety too. How about a sci-fi comedy series? I think there's an audience for it, if you could keep the costs down.

It isn't the computers that are the problem - it's the artists that use the computers. A recent desktop, purchased for $2000 US, has all the processing power and RAM needed to do some very high-quality rendering in a reasonable period of time (medium-quality renderings in real time). If one doesn't do it for you, get 10 or 100, and link them together in a beowolf cluster; even purchasing them fresh for EVERY EPISODE, you won't hit even 1/4 the per episode budget for Enterprise listed earlier. However, consider: of those who play SEIV, how many can produce quality, rendered shipsets from scratch? Three? Four? How long does it take them to make one? How long would it take them to make one that would allow you to zoom in to any part of the ship's hull to the point where it is STILL realistic and has good lighting, shadowing, and reflection effects off the nuts & bolts? Now consider that every episode where they show damage has to have that section completely redesigned to show the internals and damaged components, and every time they show damage being dealt they need to make several stages of the damage in the same way, and tell the machine how to transition between, for each instance of damage-dealing. It isn't the machine time that's the problem - it's the time of the artists involved that run the machines.

[narf poit chez BOOM]

Which is why we need a computer with the type of capability of a ST holodeck.

See Renegade 13 realise his 'pitchfork' is actually four pythons! See R13 get rushed to the hospital! Thrills! Chills! And at the end, the winner gets ten million dollar bills!

[Jack Simth]

narf poit chez BOOM said:
Which is why we need a computer with the type of capability of a ST holodeck.

There are (at least) two major technical problems and two social problems with that:
1) A programming problem - you can convince a computer to take a library object, run some specified portion of it's vertices and surfaces through an algorythm, and render the result, but the variety of point-manipulation algorythms needed, and the methods of specifying what segment, what manipulation, and what variables in a reasonably intuitive fasion would take ages to program, as ALL of that must come from at least one very intelligent human mind (probably hundreds or thousands).
2) A library problem - sure, the holodec can build highly detailed re-creations of scenes - but, of necessity, it's pulling things out of a library somewhere - the user tells the computer he wants a nature scene, and the computer picks one from a library, or generates one based on a template and componentes from a library (a component may, in turn, be a template with components pulled by some algorythm out of a library - recursive to whatever finite degree is desired) - but fundamentally, it's pulling stuff out of a truly enormous (and incredibly well-indexed) object library. This library must be built before a Star Treck Holodeck ease rendering system can be implemented.
3) A copyright problem - you can take a chunk of data, and manipulate it through some algorythm, but unless you have an agreement with the owner of the original data chunk specifying otherwise, you can't legally distribut the altered work willy-nilly. So, even after you have dealt with 1 and 2, you have a big legal mess to deal with before you can use the easy system in an mass-media fashion.
4) A repetition issue - Have you ever noticed how games of the same basic type from the same company tend to have the same basic feel to their appearence? That's mainly because the game companies tends to re-use their engines; it's even common for a company to liscense it's engine to other companies - so you will sometimes see many games of the same nature from different companies that have the same feel to their appearence. If you don't completely replace the library (step 2) and algorytms (step 1) fairly often, the easy-made shows built from the engine will start to have a repetitive look to them, and your audience will become tired of it, eventually shunning any show made with the easy-create engine and library. A very, VERY big library of algorythms and objects, well-used random factors and style/genre variables will alleviate the problem, but will ultimately just delay the point at which they all start looking alike. With the number of times the holodec is featured in ST, this isn't an issue. With the number of times it would be used to make many different rendered components for shows, it would be - in short order (a few years), most probably. For something used to entertain the crew on a long voyage, this isn't an issue, as THEY are the creators, and can change things to suit their individual preferences, puttting as much work in as needful to make things match; for something used to produce mass entertainment, this is a very big issue.

[Atrocities]

I make bike videos all the time. The resources needed to store that footage digitally is astounding. To render a four minute video takes about 20 minutes. To render a video without lock ups or fall out is a roll of the dice with XP and or any windows based PC.

[Jack Simth]

1)
5->1 rendering time is perfectly acceptable for 45 minutes of professional footage for a 13 episode season one season per year; use low-quality renderings for testing, then let the final run overnight. If it is an issue, you can speed things up by using more computers on a high-quality LAN in a Beowolf cluster (I had a class on it in college - fun stuff - whole new ways to lock up a machine). Run a Linux-based system to help prevent lockups.

2)
I've never rendered video, but I record TV shows on my PC; mpeg-4 Fast Motion works wonders for the file size - at 15 frames/sec, 2700 kbits/second, 48 kh stereo sound, I get an hour long show stored in about a gigabyte, with little danger of errors crashing the capture, and have rather good image quality - and I've only got a little PIII with 128 MB of RAM running Lose98. A little compression can do wonders.
And, ya know, with 200 GB hard drives being fairly common nowadys in the $2000 machine price range, space isn't an issue. If you arrange a filing system that can handle it, it doesn't matter that you burn through 4 GB in 5 minutes (I did this a few times when recording TV on my computer - I had the settings WAY TOO HIGH) - a full 60 minute episode is only 12 5-minute blocks; with the above ratio, that's 48 GB - you can store 4 of those on a 200 GB hard drive, available for purchase at most computer stores.

Yeah, the recourse requirements for video can get pretty high - but modern equipment can get a lot higher fairly readily anymore.

[Sivran]

Youch, 1 gig for only an hour? Chop a couple hundred meg off it with Xvid

[Atrocities]

When I render video it must be absolute quality. Thats 30 frames a second, hi defination. The quality must be there in order for preserve the moment and or if I ever decide to air or cable cast the videos. Reality shows are based upon real life and I have 80 plus hours of hard core dirt bike / quad action. Not to mention all the gun work and such my father has performed on video. Memories worth saving in high quality.

[geoschmo]

Jack Simth said:
It isn't the machine time that's the problem - it's the time of the artists involved that run the machines.

I am aware of this. When I spoke of technology advances I was refering to software as well as hardware. When computers first started getting into business, database systems were only available to the big companaies that could afford to hire programmers or pay third parties to administer them. Now my dentists office can buy an off the shelf program like Access, or something else specifically tailored to the medical field, and the billing person can maintain it. There's no reason to think image technology won't have a similer progression in time. I can imagine 50 years from now the director could input some basic parameters himself and the computer would do the hard work of turning it into the shot he's got in his minds eye. Then all you need is some actors speaking lines in front of a green screen, or maybe not even that. At some point we'll be to where we don't even need actors. That's fantasy today, but it's coming. When the technology advances to where the computer can interface with your typical non-artist type person, the cost will come down precipitously.