Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
Evolution at its best. The coyote has adapted to the harsh environment with superior intelligence and regenerative ability. What I wonder about the coyote's regeneration, if he is cut into two pieces, will both regenerate into a coyote? It could possibly explain his apparent ability to follow the bird around, a seemingly impossible task, and to appear wherever the bird goes. The more he is hacked into pieces, the more the surroundings are swarming with individual coyotes.
I liked the part about the fat lady singing.
I did not like the nondescript music. It reminded me a bit about the Tartarus theme in Battle of Olympus. Is it automatically generated or something?
The game physics are too bouncy. It needs more Mario in it.
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
Initially, I would probably use it for mundane things, like undoing mistakes and predicting what people are going to say, or how they are going to respond to different things I might say, in order to get the most favorable outcome from a discussion for instance.
Over time though, I might get more reckless and start actually doing some stunts that require agility / reaction times that are simply out of question normally for all but the most experienced parkour freaks. I do not know why I would do that, but the concept alone fascinates me. I am not very agile or physically oriented person, but I occassionally like to surprise people with stealth and unconventionality. Perhaps being lacking in one aspect ignites a will to overcome that shortcoming by other means. Might not make me any happier though. EDIT: Might do it to save dear people from accidents.
I might also use it to cheat in some non-notable manner of acquiring wealth, to add some security to the future. By non-notable, I mean not outright winning at lottery; that would have too high chances at making me a target.
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
I think I got 53 or something like that. Screwy physics. Mario should accelerate slower and deaccelerate faster, and he should jump taller when running, and he should stomp enemies by default if they collide whenever Mario is at a higher location.
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
Heh, I just had a look at Elastomania, and turns out that people have submitted quite incredible videos at Elastomania to Youtube.
Here's a set of per-level speedruns (unassisted): http://www.zeropaid.com/bbs/showthread.php/48559-Elastomania-Speedrun-%28All-Levels%29
I don't know whether the times achieved in these apply to the DOS version, Action Supercross, but the techniques certainly do.
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
I think it befalls to the same category as by-carefully-timed-resets game-save-manipulations do. In both cases, you do not abuse the game itself; you abuse its data. Here, the data is the debug menu, or rather, the RAM data that is modified through the debug menu.
It is a hazy area. Yes, it is theoretically possible on the real console...
But then again, if I extrapolate more, the Cutman's-magic-scissor glitch (Rockman 1 in eighth stage) also can be considered as belonging to the same category. There, too, RAM is overwritten by something that was supposed to be a different kind of a write.
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
Dwedit wrote:
PHPBB's search feature is worse than worthless. All it tells you is which megathread to go to for your search terms, nothing more specific than that. And too bad if you need to search for a three-letter word.
Which is why I developed and use http://tasvideos.org/Search.html .
Speaking of which, someone has apparently removed the "forums" checkbox and made it always-enabled.
EDIT: Oh. It appeared when I logged in. So the extra two options are only available for logged-in users.
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
Patashu wrote:
boct1584 wrote:
That doesn't count because all the ROMs the movie file completes are just different dumps of the same game.
What, and mega man 1 through 6 aren't the same game? :p
Megaman 1 and 2 are actually, mostly, the same game. A different memory mapper and some enhancements, but mostly the very same code.
Megaman 3 had a different engine, and Megaman 6 had a different engine.
I don't know about Megaman 4-5 versus the others.
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
Nach wrote:
Allowing Google to index the forum is a bad idea. phpBB is swiss cheese and constantly being exploited, there's malicious robots running around the internet destroying phpBB sites all the time. We do not need to advertise our existence
How about this: Create special indexed pages for phpBB threads on the mainsite for the purpose of being a readonly archive for search engines. With links to the actual threads on forums (àla links to discussions on submissions).
That said, I agree with the idea of unconstrained, overridable posts_per_page parameter.
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
GuanoBowl wrote:
Is it physically impossible to go through the door in the sky at the very beginning of the game?
Oh, that must have made people mad. I remember trying it for ages as well.
And now I have learned that game programmers usually do not put any easter eggs in situations like that that are expected to be honestly absolutely impossible.
Games are truly just a big coulisse; peek behind the coulisses and there is literally nothing there. In 3D games, those scenes that are always observed from one viewpoint only often don't even have a backside; the polygons are viewable from one side only, and the backside is see-through. Doors that are permanently locked have literally just void on their other side if somehow walked through.
In other words, find a way to go break the established bounds of the scene, and you find that you did not end up in a previously undiscovered but fully designed and decorated place. Except when it is really meant to happen.
bisqwit@hariyu:~$ mplayer earthbound-tas-pirohikomugg.mkv
MPlayer SVN-r31235-4.4.4 (C) 2000-2010 MPlayer Team
Playing earthbound-tas-pirohikomugg.mkv.
[mkv] Track 1 has been compressed with an unknown/unsupported compression
[mkv] algorithm (3). Skipping track.
[mkv] Track ID 1: video (V_MPEG4/ISO/AVC), -vid 0
[mkv] Track ID 2: audio (A_VORBIS), -aid 0, -alang und
[mkv] Will play video track 1.
Matroska file format detected.
VIDEO: [avc1] 256x224 24bpp 60.000 fps 0.0 kbps ( 0.0 kbyte/s)
==========================================================================
Opening video decoder: [ffmpeg] FFmpeg's libavcodec codec family
Selected video codec: [ffh264] vfm: ffmpeg (FFmpeg H.264)
==========================================================================
==========================================================================
Opening audio decoder: [ffmpeg] FFmpeg/libavcodec audio decoders
AUDIO: 48000 Hz, 2 ch, s16le, 0.0 kbit/0.00% (ratio: 0->192000)
Selected audio codec: [ffvorbis] afm: ffmpeg (FFmpeg Vorbis)
==========================================================================
AO: [oss] 48000Hz 2ch s16le (2 bytes per sample)
Starting playback...
[h264 @ 0x89ecc80]AVC: nal size 2115429
[h264 @ 0x89ecc80]no frame!
Error while decoding frame!
[h264 @ 0x89ecc80]AVC: nal size 4673
[h264 @ 0x89ecc80]no frame!
Error while decoding frame!
[h264 @ 0x89ecc80]AVC: nal size 5185
[h264 @ 0x89ecc80]no frame!
Error while decoding frame!
[h264 @ 0x89ecc80]AVC: nal size 4417
[h264 @ 0x89ecc80]no frame!
Error while decoding frame!
[h264 @ 0x89ecc80]AVC: nal size 4417
bisqwit@hariyu:~$ vlc earthbound-tas-pirohikomugg.mkv
VLC media player 1.1.2 The Luggage (revision exported)
<...>
swScaler: pal8 is not supported as output pixel format
[0x829bc4c] swscale scale error: could not init SwScaler and/or allocate memory
[0x81efb54] avcodec decoder error: don't know how to convert chroma 12
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
Also, correct me if I am wrong, but the point of ramdrive is kind of the same as the point of smartdrv. Neither make much sense under an emulator that does not simulate disk access delays.
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
Naturally it changes the number of frames. The game may or may not run at unchanged speed, but it will render more frames per second and may accept input more rapidly.
As for standard point of reference, the standard point of reference should be the amount of time (measured in seconds) it takes to complete the game starting from PC poweron.
In my opinion, framecounts are completely irrelevant. Why would you like to count the number of times the screen refreshes? Why don't you count disk accesses or something equally irrelevant for a change?
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
Rules question.
Some games accept input only once per a frame. However, contrary to legacy platforms like the NES, on the VGA (and especially on SuperVGA display cards), the framerate can be altered programmatically without altering the actual hardware. For example, instead of the default 70.0 fps, the 320x200 display mode (mode 13) could be changed to be rendered at 120.0 fps, and it will still be 320x200. This is all possible.
The question is, is it theoretically allowed on TASVideos to load up a little TSR prior to the game's loading, that when installed, will alter the mode 13 after it is set such that the screen refresh rate is changed to 120 (or even higher) from the default 70*?
There is no precedent on TASVideos for a question like this, so no law is set on it either. What do you people think?
Let me emphasize that this is not an emulator hack. It all happens within the emulation, which means it is possible on actual hardware as well (provided that the CRT (monitor) can keep up with the signal produced by the display card).
Also I do realize also that this is a slippery slope: If a TSR that alters BIOS functions is allowed, how about an ActionReplay style TSR that pokes stuff into the game's RAM?
*) Yes, mode 13h is 70 fps, though JPC versions 10 and earlier force all screen modes to be refreshed at 60.000 fps contrary to how VGA actually works.
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
Derakon wrote:
you'd be continually interrupted by probes (which IIRC spawn at a significantly increased rate so long as you haven't visited the starbase)
Well, that is absolutely true. Their encounter rate is fixed at 100 % when you haven't visited the starbase. In normal play, the rate begins at 2 %, and increases by 2 % automatically every 182 days until it caps at a meager 8 %. When the annihilation signal is sent, it begins decreasing 2 % at 23 day intervals.
EDIT: Yeah, it seems that the support for starbase-less game completion has been added in UQM, possibly after that very bug report. The original code does not have it. So scratch this idea then. Bleh...
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
One thing that would possible be interesting to see would be to beat Star Control II without ever visiting the starbase. I do not know whether it is possible, but according to the source code (at least UQM, but probably original SC2 as well), at least the game is prepared for the event that it happens. For instance, once you equip the Sa-Matra detonator (the Chmmr-empowered Utwig bomb), you will be teleported back to Earth rather than to the Starbase.
Minerals will be useless. You will have to collect biodata and rely on the Melnorme traders. Fwiffo (the lone Spathi captain on Pluto) will become invaluable in fending off the rogue Slylandro probes, which you cannot evade with your slow SIS ship. (The Melnorme do take your turning/thrusting jets if you cannot pay for fuel, but they will not sell you even one.)
Your fuel tank will only store 60 units of fuel (you cannot build more tanks or even the bigger tanks disclosed by the Melnorme without visiting the starbase), so you will need to plan your excursions very carefully and to rely on the Melnorme traders on tight intervals. You can also buy fuel from the Druuge, but it will cost you crew, which is not easy to come up with.
Also, the only way to recover lost crew is by transferring them from ships you receive as a gift. This way, it is theoretically possible to purchase Druuge's expensive gifts. Your SIS ship can only store 50 crew, since you cannot build new crew pods. Note that once your fleet is full, you cannot receive more gift ships until you intentionally destroy some of your fleet. Note that Pkunk ships can regenerate crew if they are destroyed in a battle.
This will easily be at least a 2-hour or 3-hour movie, but it will include many things that are interesting to see that are completely omitted in the current TAS.
For reiteration, here is the list of things the Melnorme do sell:
-- Fuel. You can buy fuel any time. 1 credit per unit of fuel. Even if you run out of fuel in deep space, they will come and sell you fuel; it will just cost you an arm and a leg; and if you don't have those, they will donate invest some for you. One of those "arms and legs" is your planet lander, which you cannot get a replacement for without visiting the Starbase. So you do need credits. You do have a very miniscule chance (about 0.25^256) of keeping your stuff though.
-- Information. Information is sold about current events, about alien races, and about history. Since no information advances the plot, they will never be seen in a TAS. All information pieces cost 75 credits per item.
-- Technology. All technology cost 150 credits per piece, and you cannot purchase tech N without purchasing tech N-1 first. Technically none is required in a TAS, but Tech 2 and Tech 7 have potential of making the TAS faster.
---- Tech 1: Blaster weapon SCHEMA.
---- Tech 2: Faster landers. Which they will not sell unless you bought Tech 1.
---- Tech 3: Antimissile defense SCHEMA. Will not sell unless you bought Tech 2.
---- Tech 4: Lander shields against biology. Will not sell unless you bought Tech 3.
---- Tech 5: Improved lander cargo capacity. Will not sell unless you bought Tech 4.
---- Tech 6: Larger fuel tank SCHEMA. Will not sell unless you bought Tech 5.
---- Tech 7: Improved lander stunners. Will not sell unless you bought Tech 6.
---- Tech 8: Lander shields against earthquakes. Will not sell unless you bought Tech 7.
---- Tech 9: Tracking system SCHEMA. Will not sell unless you bought Tech 8.
---- Tech 10: Lander shields against lightning. Will not sell unless you bought Tech 9.
---- Tech 11: Lander shields against fire. Will not sell unless you bought 10.
---- Tech 12: Cannon weapon SCHEMA. Will not sell unless you bought 11.
---- Tech 13: Shiva furnace SCHEMA. Will not sell unless you bought 12.
All SCHEMAs are completely useless: Credits will be paid, but you cannot build any of them without visiting Starbase.
Your space battles will be all done with ships that you acquire on-visit; you can never build new ships. That means no Chmmr, no Orz, no Arilou. You do get:
-- One Spathi ship with just one crew member (Fwiffo on Pluto)
-- Four Pkunk ships (repeated as many times as you want, but require at least 1-gameday intervals)
-- Druuge ships (one per every 100 crew you sell to slaverySince your SIS ship can only store 50 crew, this is impossible. Six ships if you pick up and sell the Shofixti maidens)
-- Four Umgah ships (after de-zombification) Impossible, because you're given a battle you cannot escape nor win.
-- Four Supox ships (repeat as often as you want, but only until you provide repaired Ultron)
-- Four Utwig ships (repeat as often as you want, but only until you provide repaired Ultron)
-- Twelve Yehat ships (your fleet can be refilled with rebel Yehat ships as many times as you want, but require at least 1-gamemonth intervals.) However, to accomplish this, you must first start the Yehat civil war, which can only be done if you have a Shofixti ship with you. And you cannot build them... (And without the escape pod, you cannot convince Tanaka.)
-- Four Zoqfot ships (once)
Acquiring Melnorme credits:
-- There are 10 Rainbow worlds in the SC2 universe, and each one is worth 500 credits.
-- The Umgah will also give 500 biounits once they are dezombified. N/A, see above.
-- Life forms (biological scan on uninhabitated planets) can be stunned and captured and sold to the Melnorme. The most life-rich planet is Gamma Geminorum with 341 units, but it is rather far.
Notes on fuel use:
-- The portal spawner will use 10 units of fuel per use.
-- Each planet landing will consume somewhere between 0.1 to 1.0 units of fuel.
-- Escaping a battle with the SIS ship is umpossible until it no longer matters.
-- Hyperspace travel costs fuel approximately 1 unit per 10.0 coordinates. The exact means how this is calculated are somewhat sketchy. Internally, the fuel amount is maintained 100-multiplied so that 5 fuel units is stored as an integer value of 500. The distance calculation uses a 4-bit fraction.
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
Ilari wrote:
<Enter> the boot prompt, HDD only: 661ms
<Enter> the boot prompt, all HDx: 661ms
You're right about this one. But JPC runs faster if all HD slots are filled! For the duration of the discovery of IDE devices, that is.
Which makes no difference for movie timing, but all the difference for the movie author when making the movie.
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
So here are some observations I made while optimizing the boot process for JPC-RR.
CD-ROM discovery takes time. It is fastest if you setup the same HDD image in all four harddisk positions.
If JPC-RR refuses to do that for you, simply save your movie, quit, edit the movie file, find +HDA and duplicate the line four times: +HDA, +HDB, +HDC and +HDD. Then start JPC-RR and load the movie.
Do not hit F12 to explictly select a boot device. If you do, it takes a frame or two longer.
FreeDOS is relatively slow to boot. If you do not need XMM or EMM, consider using DR-DOS (now called OpenDOS), available here: http://drdosprojects.de/index.cgi/download.htm
I found out that in optimal conditions, FreeDOS boots in 14..16 frames but DR-DOS 7.01.01 boots in 1..2 frames, while still having enough RAM to run e.g. Star Control II. OpenDOS also seems to have faster file access, further speeding up the game throughout its runtime.
You can download it here: http://drdosprojects.de/index.cgi/download.htm
Or a pre-prepared floppy image for JPC-RR here: http://bisqwit.iki.fi/kala/drdosjpc.zip *
Note that you may need to insert one dummy key press into the keyboard buffer before DOS loads up to make it most swiftly, and two hits of "Enter" into the datetime prompts.
*) Created by mounting DISK.IMG (provided in opendos) -o loop, copying there the files listed in makeboot.bat (provided in opendos), umounting, and then disk-imagifying it with ImageMaker in JPC.
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
I wonder if someone could contribute similar subtitles for the Star Ocean TAS. It has been largely downvoted because of the language nobody understands. It is a great game!
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
Action Supercross is cool. It has been very popular a shareware game especially in Finland. It is basically like "Excitebike meets Paper Mario". Nowadays, it is known as Elastomania.
It does not seem to work in JPC-RR yet (also, JPC-RR's usability is horrible), so I made this quick demonstration movie with DOSBox to show what the game is like. Note that it does not aim for speed records of any kind.
Link to video
Here are high-scores for various levels of the game:
From http://www.elastomania.com/ascbestt.htm :
+---------------------------------------------------------------------------------------------------------------+
|1. Warm Up | 14:97| StIkKy ||22. Precarious | 25:78| AnTTiM HfK |
|------------------------------+---------+---------------++---------------------------+---------+---------------|
|2. Flat Track | 17:57| SoGe ||23. Circuitous | 28:15| psym0rph FM |
|------------------------------+---------+---------------++---------------------------+---------+---------------|
|3. Twin Peaks | 20:49| DeG ||24. Shelf Life | 40:42| tLDOmni |
|------------------------------+---------+---------------++---------------------------+---------+---------------|
|4. Over and Under | 32:55| Sventurbo ||25. Bounce Back | 54:22| Petri |
|------------------------------+---------+---------------++---------------------------+---------+---------------|
|5. Uphill Battle | 22:30| umiz ||26. Headbanger | 49:69| Peddo |
|------------------------------+---------+---------------++---------------------------+---------+---------------|
|6. Long Haul | 39:97| psym0rph FM ||27. Animal Farm | 1:16:87| Ufo FM |
|------------------------------+---------+---------------++---------------------------+---------+---------------|
|7. Hi-Flyer | 32:61| Nikke ||28. Zig Zag | 1:05:92| ChampiON |
|------------------------------+---------+---------------++---------------------------+---------+---------------|
|8. Tunnel Terror | 1:06:70| Dr_Luni FM ||29. Labyrinth Pro | 2:40:65| tLDloX |
|------------------------------+---------+---------------++---------------------------+---------+---------------|
|9. The Steppes | 14:72| Elbono ||30. Fruit in the Den | 1:04:50| Limpan |
|------------------------------+---------+---------------++---------------------------+---------+---------------|
|10. Islands in the Sky | 22:75| tLDOmni ||31. Jaws | 1:01:17| Ntc |
|------------------------------+---------+---------------++---------------------------+---------+---------------|
|11. Hill Legend | 19:40| TuA ||32. Curvaceous | 32:94| TuA |
|------------------------------+---------+---------------++---------------------------+---------+---------------|
|12. Loop-de-Loop | 20:50| tLDJohnDoe ||33. Double Trouble | 1:38:54| tLDloX |
|------------------------------+---------+---------------++---------------------------+---------+---------------|
|13. Serpents Tale | 1:08:55| tLDOmni ||34. Framework | 41:53| BoBBo |
|------------------------------+---------+---------------++---------------------------+---------+---------------|
|14. New Wave | 1:37:69| psym0rph FM ||35. Enduro | 52:04| Mandel |
|------------------------------+---------+---------------++---------------------------+---------+---------------|
|15. Labyrinth | 1:00:05| umiz ||36. He He | 1:01:76| Sventurbo |
|------------------------------+---------+---------------++---------------------------+---------+---------------|
|16. Spiral | 50:04| Garion ||37. Freefall | 34:91| Tier |
|------------------------------+---------+---------------++---------------------------+---------+---------------|
|17. Turnaround II. | 39:46| tLDzeth ||38. Enigma II. | 29:13| tobak |
|------------------------------+---------+---------------++---------------------------+---------+---------------|
|18. Hangman | 17:13| ChampiON ||39. Expert System | 1:02:07| Mandel |
|------------------------------+---------+---------------++---------------------------+---------+---------------|
|19. Slalom | 45:62| TuA ||40. Tricks Abound | 55:48| Fulgore |
|------------------------------+---------+---------------++---------------------------+---------+---------------|
|20. Quick Round | 48:84| Sventurbo ||41. Hang Tight | 28:01| Ufo FM |
|------------------------------+---------+---------------++---------------------------+---------+---------------|
|21. Ramp Frenzy | 59:36| BoBBo ||42. Apple Harvest | 1:56:40| Sventurbo |
+---------------------------------------------------------------------------------------------------------------+
Records for tracks of the previous version
+--------------------------------------------------------------------------------------------------+
|Fair Trial | 10:70| Karlis FM ||Slippery Slope | 14:16| SoGe |
|-----------------------+--------+---------------++-----------------------+--------+---------------|
|The Turnaround | 22:76| ChampiON ||On Top | 24:19| Karlis FM |
|-----------------------+--------+---------------++-----------------------+--------+---------------|
|The Enigma | 29:58| Tuska || | | |
+--------------------------------------------------------------------------------------------------+
Across world records
Last updated: 16 September 2004, table #418
Level Kuski Nat. Time Updated
1. Warm Up StIkKyBeAr [TNT] USA 14,84 -
2. Flat track mr [FM] FIN 17,53 27 Oct 2002
3. Twin Peaks TuA [TNT] SWE 20,46 -
4. Over and Under Sventurbo [MoA] SWE 32,55 -
5. Uphill Battle umiz [SAT] SWE 22,30 -
6. Long Haul psy [.] FIN 39,97 -
7. Hi-Flyer Nikke [EMT] FIN 32,61 -
8. Tunnel Terror barbapappa [CF] SWE 1:06,65 16 Sep 2004
9. The Steppes StIkKyBeAr [TNT] USA 14,01 26 Jan 2003
10. Islands in the Sky Omni [tLD] NOR 22,75 -
11. Hill Legend TuA [TNT] SWE 19,40 -
12. Loop-de-Loop JohnDoe [tLD] NOR 20,50 -
13. Serpents Tale Omni [tLD] NOR 1:08,55 -
14. New Wave barbapappa [CF] SWE 1:36,33 19 Nov 2003
15. Labyrinth umiz [SAT] SWE 1:00,05 -
16. Spiral Garion [AA] SWE 50,04 -
17. Turnaround II barbapappa [CF] SWE 39,08 16 May 2004
18. Hangman zyntifox [SC] NOR 16,67 -
19. Slalom TuA [TNT] SWE 45,62 -
20. Quick Round barbapappa [CF] SWE 45,02 12 Feb 2004
21. Ramp Frenzy KingKong [PRA] SWE 59,36 -
22. Precarious AnTTiM [HfK] FIN 25,78 -
23. Circuitous StIkKyBeAr [TNT] USA 27,00 26 Jan 2003
24. Shelf Life ciph [EM] CAN 40,31 9 Aug 2002
25. Bounce Back mr [FM] FIN 54,15 16 Sep 2004
26. Headbanger StIkKyBeAr [TNT] USA 49,51 25 July 2003
27. Animal Farm Ufo SWE 1:16,87 -
28. Zig Zag zeth [tLD] NOR 1:04,16 -
29. Labyrinth Pro barbapappa [CF] SWE 2:40,52 25 July 2003
30. Fruit in the Den Limpan [SAT] SWE 1:04,50 -
31. Jaws barbapappa [CF] SWE 1:00,86 16 Sep 2004
32. Curvaceous TuA [TNT] SWE 32,94 -
33. Double trouble loX [tLD] NOR 1:38,54 -
34. Framework mr [FM] FIN 41,45 25 July 2003
35. Enduro Mandel [PRA] SWE 52,04 -
36. He He ciph [EM] CAN 59,55 4 Feb 2002
37. Freefall Tier [AAF] AUS 34,91 -
38. Enigma II StIkKyBeAr [TNT] USA 25,90 25 July 2003
39. Expert System mr [FM] FIN 1:00,59 9 Aug 2002
40. Tricks Abound mr [FM] FIN 49,28 27 Oct 2002
41. Hang Tight ramone [CF] SWE 27,61 21 May 2002
42. Apple Harvest barbapappa [CF] SWE 1:54,61 16 May 2004
Total time 33:45,57
Across 1.0/1.1
Level Kuski Nat. Time Updated
12. Fair trial mr [FM] FIN 10,53 -
14. Slippery Slope SoGe FIN 14,16 -
15. The Turnaround StIkKyBeAr [TNT] USA 22,48 26 Jan 2003
17. On Top StIkKyBeAr [TNT] USA 23,99 25 July 2003
23. The Enigma mr [FM] FIN 28,83 -
Total time 11:45,68
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
Here is a C++ programming challenge I have been working on lately.
------------------------
There are 13 features. There are thus 213, that is 8192, combinations of these features.
There are 7 classes, each of which implement any number of these features from 1 to 13; there is overlap in these implementations (there may exist more than one class that implements a certain feature). With multiple-inheritance, one can also devise new classes which implement all the ancestors' methods. There are thus 27, that is 128, classes to choose from.
Your task: Write a function, which, given a parameter "bitmask", which is an at least 13-bit integer that denotes the set of features required, returns an instance to the smallest possible class that implements all of the requested features. The class may implement more features, but it must be the smallest implementation to do so. In case of ambiguity, chooses the more fundamental class. You can assume that asking for parameter values outside the range 1..8191 (inclusive) are errors.
Expect the number of features, the number of classes, the size of the classes, and the list of features they each implement to change later. Write your code so that this will not pose a problem.
Design goals: Try to produce optimal code in both speed and size (for example, if possible, no 8192-case switchcases, and no pointers to methods of classes that are never going to be used). And of course, it must be maintainable. As such, no 8192-case hardcoded switchcases.
Example code:
[code c++]enum Feature
{
Feat00 = 0x0001,
Feat01 = 0x0002,
Feat02 = 0x0004,
Feat03 = 0x0008,
Feat04 = 0x0010,
Feat05 = 0x0020,
Feat06 = 0x0040,
Feat07 = 0x0080,
Feat08 = 0x0100,
Feat09 = 0x0200,
Feat10 = 0x0400
};
/* For convenience, here is the base type for all the classes. */
class Base { public: virtual ~Base(); }; /* Base is 0 bytes and implements nothing */
class Class00: public Base
{
char bulk[16];
void DoFeat00();
}; /* Class00 is 16 bytes and implements Feat00 */
class Class01: public Base
{
char bulk[36];
void DoFeat00(); void DoFeat01(); void DoFeat02(); void DoFeat03();
void DoFeat04(); void DoFeat05(); void DoFeat06(); void DoFeat07();
void DoFeat08(); void DoFeat09();
}; /* Class01 is 36 bytes and implements all features EXCEPT Feat10 */
class Class02: public Base
{
char bulk[4];
void DoFeat02();
}; /* Class02 is 4 bytes and implements Feat02 */
class Class03: public Base
{
char bulk[4];
void DoFeat03();
}; /* Class03 is 4 bytes and implements Feat03 */
class Class04: public Base
{
char bulk[8];
void DoFeat04();
}; /* Class04 is 8 bytes and implements Feat04 */
class Class05: public Base
{
char bulk[28];
void DoFeat00(); void DoFeat04(); void DoFeat05(); void DoFeat06(); void DoFeat07();
}; /* Class05 is 28 bytes and implements Feat00, Feat04, Feat05, Feat06 and Feat07 */
class Class06: public Base
{
char bulk[44];
void DoFeat00(); void DoFeat04(); void DoFeat05(); void DoFeat06(); void DoFeat07();
void DoFeat10();
}; /* Class06 is 44 bytes and implements Feat00, Feat04, Feat05, Feat06 and Feat07 AND Feat10 */
/* This is the function that needs an implementation: */
Base* InstantiateSmallestClassWithFeatures(unsigned featureset);[/code]
You are free to add static methods / properties to the classes or even templatize them if required, but do not change their size or their set of features (do not alter the bulk either; it represents the mandatory size). For example, you can add a static const int into each class indicating the set of features implemented.
Examples:
-- When one requests Feat04, gets Class04.
-- When one requests Feat02+Feat04, gets a combination of Class02 and Class04.
-- When one requests Feat04+Feat05, gets Class05.
-- When one requests Feat00+Feat02+Feat03+Feat07, gets Class01 ( the combination of Class05, Class02 and Class03 is the same size, but Class01 is more fundamental ).
------------------------
This programming challenge arised to me in the implementation of animmerger. I solved it. But I think it is an interesting challenge! I would like to see other solutions/attempts.
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
adelikat wrote:
The built in weight was actually implemented by Bisqwit, but I have no idea when as it was never mentioned or documented. I found out about it after taking over the site. I personally would have never approved such a weight, thanks to grunt for taking the time to remove it from the site code.
Very well.
I admit I was not particularly vocal about the rating weighting system, but I never kept it secret from anyone who asked. It's just that I was completely content with people never asking. I think I explained it a few times on the IRC channel; that is how you knew the bit about Neofix's special case, for instance. And of course, people who had to do work with it (such as Warp) had disclosure.
Basically, the weighting system existed since the beginning, that is in 2006 when the rating system was implemented.
To recap, here is how it worked when I handed things over:
To calculate authority coefficient, i.e. rating weight for the user:
Anyone with a player entry has a post_count_threshold of 9 and scale of 1.1. Anyone else has a post_count_threshold of 22 and a scale of 1.0.
Entrusted users (judges, publishers, admins, and any combination thereof) have the coefficient between 1.1 to 2.0, exact value whereof is lower or higher depending on the entrust level of the role.
For anyone else, the number of posts is calculated. The coefficient is a value between 0 to scale, directly proportional to postcount, that maxes out at post_count_threshold. It is always higher than 0, though.
Consequently, the highest coefficient was on admin users, who had 2.0. The lowest coefficient was on lurkers who had never posted or contributed in a movie, who had 1/22, that is 0.045. One had to get at least 22 posts to get a coefficient of 1.0. If the user has a player entry, their coefficient is always at least 1.1/9, that is, 0.12, even if they have never posted. If they've posted more than 8 posts, their coefficient is 1.1.
Ratings you provide now will retroactively grow more significant as your current coefficient grows. That means you don't need to refrain from rating until your post count has increased. Your already-cast ratings will retroactively benefit from your posting / publications.
Personal exceptions can be made. (Only one has been done so far, and it was a demoting type.)
Rationale:
The system of rating was similar to that of Eurovision song contest. In the Eurovision song contest, anyone can "televote" by sending a SMS/phonecall, but there is also a panel of judges (jury), consisting of selected professionals in the area of music and entertainment. The weight in 2010 contest was 50% televote and 50% for panel judges. Thus, the average juror had the same rating power as some 1000 to 100000 "regular" people combined.
On TASVideos, anyone with a player entry likely has a sense of entertainment matching the site's goals, since a movie they authored according to their own ideals, was published on the site. Therefore they can be said to belong into a "panel of judges" as far as rating is concerned. Similar rationale goes for entrusted contributors, but with stronger confidence.
The post count provides an additional confidence factor, and especially with the case of lurkers it shields against rating abuse. It is a high threshold for anyone to register 22 dummy accounts to match some seasoned member's vote.
Oh, and I am not a fan of pure democracy either. It is unbiblical. No dictatorship nor communism either, but the fact stands that "majority" is always stupid. In leadership positions you need people who have the right vision and wisdom. And they must never be outranked by a mob. Democracy is why Europe and USA are falling to islamization. Own dog bites.
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
rhebus wrote:
My guess is that some hackery maps this "fake" RGB value into the real RGB value that results in the image (190,0,50) in some way which doesn't preserve luma -- and at least one of luma or chroma must be diddled to get back to real RGB values.
That is pretty much my theory too.
YUV can express colors that are not possible to express in RGB.
For example, a fully-saturated red that is brighter than 30% of the maximum brightness of brightest white.
In RGB, if you try to increase the brightness
of #FF0000 (assuming 8 bits per channel), you lose saturation.
Wich means, if you've got a fully saturated green
(#00FF00, brightness 59% of #FFFFFF), i.e. luma = 59%, and you change the chroma into fully saturated red, you've got a YUV that cannot be converted losslessly into RGB.
The best you can get
is #FF7817 (rgb leak with Y weighting; preserves luma (59%), but becomes noticeably yellow),
or #FF4BFF (rgb leak with Y weighting; also preserves luma (59%), but becomes noticeably purple),
or #FF6969 (non-discriminating rgb leak with Y weighting; luma=59%; some saturation lost),
or #FF2424 (simple rgb leak calculation; luma=only 40%; some saturation lost too),
or #FF0000 (preserves chroma, but luma is only 30%. This is clamping conversion, which is what most software do).
EDIT:
Here's an image I created to illustrate this better. It has the basic SMPTE color bars*. Vertical axis: Luma; Horizontal axis: Saturation (used to multiply chroma). Eight different saturation values are shown.
A pair of black circles is used to mark the interval on each color bar, outside which the RGB conversion is lossy. Between values are accurate representations; outside values have had their saturation reduced in order to make it fit into the RGB gamut.
The grayscale conversion shows that luma is preserved in all values. Gamma=1.6.
EDIT 2:
In the above imageset, I went to painstaking efforts to preserve the luma in the unrepresentable colors. Here is what most software does instead, i.e. clamping:
To my surprise, the same also happens in the upper region. Once the luma becomes small enough, too stark chroma values will cause RGB components to appear negative. When only simple clamping is done, these colors will appear anything but black. I am not sure what to make of this. On the other hand, color distribution in these graph correlates with the color of noise patterns seen on badly tuned tv receivers…
*) Not exactly right. The colors are the same, but the real SMPTE color bars are actually 100% saturated color bars in which RGB components are always either 0 or 75%. Consecutively, the luma varies. They are sorted in order of increasing luma.
Editor, Experienced Forum User, Published Author, Active player
(296)
Joined: 3/8/2004
Posts: 7469
Location: Arzareth
I decided to go and do independent tests with the same basic idea.
The test results are here, as is the shell script that I used to run ffmpeg to create all these images.
http://bisqwit.iki.fi/kala/yv12scaledemo/
Further comments I wrote in IRC while nobody seemed to be listening:
For some reason, I'm not getting the artifacts you do on Sonic.
http://bisqwit.iki.fi/kala/yv12scaledemo/
I have tried offsetting the Sonic image by a pixel or two in any direction, but no matter what I try, I cannot seem to get the horrible red-distorted version you have, Aktan, regardless of the choice of filters.
I think there's something wrong in your setup instead. Maybe you converted the images a few times back & forth in succession? It would explain the way heavier ringing artifacts than in my results, too...
I used ffmpeg for both conversion phases. The input (RGB) is a PNG file; it is converted into a single-frame AVI with FFV1 codec, which uses YUV420p (4:2:0) by default while being lossless in all other aspects, and then back into PNG file, which requires a RGB conversion. YUV420p is what H.264 (and thus x264) also uses, thus relevant here.*
*) MEncoder does output different messages when encoding into FFV1 and when encoding into X264:
SwScale: scaling 162x37 RGB 24-bit to 162x38 Planar YV12 (FFV1)
SwScale: scaling 162x37 RGB 24-bit to 162x38 Planar I420 (X264)
However, further study reveals that the only practical difference between these two is how the data is stored; the pixel transformations are exactly the same. I420 planes are numbered 0,1,2; YV12 planes are numbered 0,2,1.
EDIT: Found the way to reintroduce those artifacts. So I did, and I fixed bugs, and updated the score table. The winner is: gauss.
EDIT 2:
In the above imageset, I went to painstaking efforts to preserve the luma in the unrepresentable colors. Here is what most software does instead, i.e. clamping:
To my surprise, the same also happens in the upper region. Once the luma becomes small enough, too stark chroma values will cause RGB components to appear negative. When only simple clamping is done, these colors will appear anything but black. I am not sure what to make of this. On the other hand, color distribution in these graph correlates with the color of noise patterns seen on badly tuned tv receivers…
*) Not exactly right. The colors are the same, but the real SMPTE color bars are actually 100% saturated color bars in which RGB components are always either 0 or 75%. Consecutively, the luma varies. They are sorted in order of increasing luma.