Veyrdite

I've just finished watching the download. It's night and day difference compared to the youtube reel. So many birds! I didn't even realise there were aliasing issues before! On that note: wet road aliasing. It's interesting that dry roads don't flicker in the distance, but wet roads do. Perhaps they forgot to implement mipmap filtering on that layer? Anyway, onto more serious notes. Video hosting and backups Services like Mega are brilliant short-term hosting for large amounts of data. But they're useless beyond that. I don't know what their business model is, and I don't know of any good alternatives, but in my experience all free file-hosting sites are transient. I used to upload mods I made to such file hosting sites. Everyone did. This was a big mistake. My mods are gone forever. As are the modding histories of entire communities. Have you ever visited a site that indexes mods and then tried to download one, only to find that it takes to you an external host that's now gone? Or that now demands money for an account? You have a few options, all in various shades of grey: (1) Make torrents and trust the community to keep the videos alive. Whilst this might work and it's almost free, it's not a very accessible option for you or the users. Not many people will go through this just to watch one of your videos, and each video you release would require an individual torrent as you can't edit existing torrents to include new data. Messy and only for the 1% of viewers. It might be worth doing this anyway (or encouraging someone to do this for you) just as a backup plan. (2) Host them yourself. At the end of the day: you are the only person you can trust to maintain motivation to keep them hosted. (3) Hybrid HTTP + torrent website Throw your hat into the experimental arena with something like peertube, which is a hybrid of options (1) and (2). Here's the Blender Foundation's instance which they setup after Youtube removed all of their videos because they didn't agree to enable advertising on them. No idea how well this works in practice. Peertube is a trendy javascript-framework based thing that's sure to break in less than two years, so it's probably worth waiting for something more reliable to come out. This is a bit concerning. I'm not privvy to your plans and who you have organised them with, but what's going on here? New host too distant to move the collection to? Happy to chat via PM, email or voice if you want a sounding board, whilst I don't sell any hosting services myself I might at least be able to listen and perhaps give a different perspective. Stutter You mention stutter issues in the video. I could not see any evidence of this in the video segment you use as an example, but with this higher quality copy I'm noticing it everywhere. Shortlist of mm:ss codes: 8:14 Car rotation during cutscene. Could just be pre-animated and not lerped. 14:19 Bus judder 22:35 Rushmore judder 23:53 Camera rotation judder at "Cloud Gate" The bus judder is a good example. I dumped all the frames and tracked a single spot (rear mini-blinker) on the bus as it traveled across the screen: Those spacings should not be so uneven. Several possible causes I can think of: Frame skipping. This explains the 1/3 then 2/3 spacings above. Could be caused either by the game, graphics drivers or the video capture process itself. Asynchronous (threaded) or similar code (scheduled jobs in one thread) that's providing uneven lock delays or job loads each frame, so each frame legitimately takes a different amount of time to complete. Requires time optimising the code to fix. Bad timer resolution. Many lazy devs (*looks sideways*) use millisecond time sources because they're easier to get and don't always overflow as often as the higher-res sources. On its own it's impossible to notice 1ms jitter, but if you code things badly a few dozen or more of these errors can pile up each frame. "The age old tradition of fixing software problems by throwing more hardware at them" can actually make this last problem worse.

From my own perspective: Time disparity. It takes much, much more than two hours to make something; even ignoring the things that would go wrong with foreign equipment in someone else's setup. Non-linear. Only certain formats/styles of production can be made from start to finish in a single go. Others involve multiple stages of correction and idea changes. Risk. If it's not your equipment, then you don't have control over it. This will bite you every time something is changed, and you have to wait for other people to "fix" it (if at all). Heads up: "Why doesn't he just..." style questions can be taken offensively depending on how they are read.

Thankyou Ross. Sorry this has been an extended saga. Thou shalt not insult my Seamonkey. To the commandline! .===========================================================================. | __ __ _____ ____ _ _ | | | \/ | ___|/ ___| / \ ___ _ __ ___ __| | | | | |\/| | \ / | _ / _ \ / __| '_ ` _ \ / _` | | | | | | | /__\ |_| |/ ___ \ (__| | | | | | (_| | | | |_| |_|____|\____/_/ \_\___|_| |_| |_|\__,_| | | | | SERVER | `===========================================================================´ [info: 01:38:42] Listening to petitions ... TRANSFERING ||#..........................................||(43/9669 MB: 0.44 %)

Holey moley that 30sec clip is a nightmare Really, really high frequency detail is everywhere in the 2D plane (tiny leaves) and the lighting patchwork flickers every few frames in the temporal domain (unsmoothed low-res shadows). I tried several encoding runs yesterday using ffmpeg & x264. The codec is hopeless at trying to surrepticiously discard content, so I tried some preprocessing. Nothing but downscaling or blurring the video (to get rid of all the HF content) works. Maybe trying to coax the motion predictor a bit more could help, I might try that.

I'm in the same boat. Many baked lighting games look way better than dynamic or cyclic ones. Games like HL1 might have blurry textures, but the edges and the lighting are sharp and crisp. Most of all: the game has a theme and character without needing 2D filters or gray-zoning of everything. If you are noticing the lighting constantly, then it's probably bad lighting. Dynamic day/night lighting vs pre-baked lighting is similar to a review I once watched (read?) of the remake of Abe's Oddysee into 3D. The 2d version allowed artists to do a lot more with each panel/scene. The 3D version made everything much more generic, likely because of the greater amount of time and effort involved to make the scenes. Pre-baked lighting lets a dev control everything and fix areas that look bad. Dynamic lighting teaches a dev just to tolerate the bad bits, because there's no global fix that does not cause lighting issues elsewhere.

Rain and roads There's something really ill-looking about the wet-state roads in this game. They're such perfect mirrors of "not the sky", looking completely stark and alien against the road. Perhaps if they had some diffusion or imperfections to their surface, but as it is they look like glitches. Are we looking into hell? An alternative dimension? Ross: you may have an uncanny valley with the road geometry, but my problem with them is purely non-euclidean.

Love drives the hate race Love/hate relationships with games leave... strong impressions. Often you can never shake them, even years later, and the thoughts they arouse don't make sense. Not being able to go back to some of these games would feel like a big loss to me. I try to archive them, but there's only so much I can do. GTA:San Andreas has some nice driving moments. Turn the HUD off and go out on your own during the night. Meander up mount chiliad in a ute, police trying to follow you but instead tumbling off the road as you listen to on the country radio station. The best thing I have discovered to do in any video game with driving: turn off the damned minimap. I end up playing these games "through the minimap" and find it impossible to appreciate the driving. I don't ever learn to navigate on my own either. Even if turning off the minimap completely breaks the game, it's worth it. Encoding issues with this video On the topic of love/hate relationships: this video. The content is great, but the video itself is difficult to watch for extended periods. I can tolerate low quality video fine. But the oscillating quality levels popping in and out every few seconds on loop gave me a headache. All of the footage of The Crew in this video is like this. Here are another representative pair of consecutive frames, minus the gif paletting: I was wondering if this is what youtube is normally like and I only just started noticing today. Then the video switched to carma 2 and it felt like a smoky haze had lifted. Youtube's current codec config just seems completely incapable of handling The Crew. Bitrate exhaustion causing the motion prediction to give up? The x264 (AVC) version of the video has this problem, the vp9 version does not. If you have the option I recommend watching the vp9 (webm) version: ... 137 mp4 1920x1080 1080p 5136k , avc1.640028, 30fps, video only, 1.28GiB 248 webm 1920x1080 1080p 7765k , vp9, 30fps, video only, 898.04MiB ... This second variant is lower quality overall but a tleast it stays consistent. Much easier to watch. The DCT blocks occasionally do a dance, mostly visible as road lines doing a jiggle wiggle, but that's not too bad compared to the dubstep keyframing of the mp4 version.

Why do the combine never recognise unsuited Freeman? He's no idiot, he knows how to blend in when entering alien environments. There's a reason he doesn't wear that helmet. Source: http://www.eeggs.com/items/34942.html Probably medicated too. Optical morphine.

Dammit, we're going to need to find a "Gordon goes train-hopping" game. I now have yet another pet project I'm unlikely to complete.

I think this video contrasts nicely against the previous Game Dungeon. Most of this one is made up of stories, memories and interpretation. I like it. The previous GD sourced most of its humour from the game itself. It felt like the game was doing the work, not Ross, and that I could get more of the same by playing the game myself. I didn't enjoy this one as much, apart from suggesting an interesting game to play. (This isn't a critique of the way the videos were made. It's a critique of the final product. There are many ways to shell a spoon, and all take time to improve.)

No one has mentioned Minecraft? I remember falling head over heels during the free weekend many years ago.

@legisland: I was given a copy of this when I was a kid and I could not click the little pictures of vehicles in the 3D world to start the 'build a vehicle' scenes. Recently I had a go at getting the game working again. Apparently this was a very common problem with the game (and I think I had to apply the same solution). I became stuck at the end of the campaign for the skateboarding character. You have to throw pizzas from a helicopter and I always run out well before I can get the police to catch the crim. I think collision detection between the crim and the police is goofed, because they would sometimes travel very closely and not trigger the ending sequence. Let's see if it still runs under WINE: *runs executable* Narrator legoman standing next to a giant spinning CD: "Wooops! You have to put the CD in your computer" Yep still works.

Parallel option: repair "The Claw" Background: I do electronics repair. Do you still have your claw? Hand mouldable products exist that slowly set in air (or can be baked, like some forms of plasticine). This would cut out the 3d scanning and printing steps. Example: sugru.

Suggested hardware option for Ross: pick up a <5 AUD USB two-jack audio dongle off the web for now. Drivers tend to be much more generic for these and they perform better than motherboard inputs/outputs in some situations. I've sent these out to people with mic input problems. When you can hear the other person's interrupts firing off every time they move their mouse it's worth suddenly appearing at their doorstep and forcing it on them. The difference between them and a bad motherboard for the input side of things is amazing, but I'll leave it up to you how you want to compare it's output it with your more expensive card's outputs. Your results will vary. Keep in mind that your speakers tend to be the biggest sources of distortion, not your DACs. Wooden bench: Damn, I should have added more pieces of pine under that monitor. If it's anything like my monitor stand there should be several novels and books called "Phase Transitions in the Early Universe" below it. I'm very careful not to knock my monitor stands. It may cause cataclysms. Minecraft video suggestions: I thought someone would notice this. Google has not worked out how to differentiate people in my house yet, so I don't get the bad minecraft videos suggested. I also get lots of "NASA hiding secret alien life on mars!" videos too. I preferred the video suggestions when they just run by similar videos to what you currently have open on the page, not to what they want to profile you with. I'm looking forward to a day that Youtube no longer has a monopoly. Perhaps archive.org might be the next big thing for video.

The sudden noise symptom suggests that digital audio data is being decoded incorrectly. Anything from losing synchronisation in an audio stream to not noticing the audio stream format has changed can cause this. Indeed it may be a hardware/firmware bug, but let's hope not Ross: have you tried sending an email to ASUS about the issue? The cards are not crazily expensive, but they're not free either, so they may be happy to give some assistance.

Absolutely nasty. Definitely not a hardware issue. The xonar drivers are unique and I don't have access to the hardware, so I can't suggest much more on the driver front than head-banging some of its settings around (eg nuke the sample rate down to 48K or 44.1K). Changing settings in other software may be an easier route (see below). I think I may have had both, I'm not sure, it's practically an assault when it happens, it's distorted my memory somewhat. Have you tried disabling hardware acceleration in firefox? It should be in the options dialog under the advanced tab. This may force the audio to take a different route through different code before it gets to your drivers. Soundcards have different acoustic properties similarly to how headphones do. If you set them up at approximately the same volume they will still both render different pitches at different volumes, which heavily influences what parts of songs you hear. I've exploited this when mastering recordings: often certain instruments/vocal parts can be brought in/out by adjusting the frequency response of parts of the recording (notable victim: double bass, but this applies to everyone). It's not anywhere are good as having a separate mic/channel on the musicians you wanted in the first place but it can get you halfway to ensuring there's good balance between the parts. Similarly I've notched out frequencies to try and mask horrible distortion in recordings, but that's a story best told at night over campfire with haunted marshmallows. The reason I'm ranting on about this: it's often easy to tweak the EQ curves yourself. Some drivers have options that let you do it, other times a box with a 'treble' and 'bass' knob on it is good enough.

I also agree that your motherboard's sound chipset "should" not sound worse than a bigger sound card. It's possible something is wrong with your motherboard/power supply (some are designed very poorly, especially re noise on the mic inputs). Keep in mind that sound quality testing is as evil as room acoustics. Small differences in volume can make you favour one device under test over another DUT. Similarly the Xonar card may be tweaked (in software or hardware) to have a higher bass response (or other EQ shenanigans), which some people prefer. It's good to see you did a comparison between the two. Hopefully you've been happy with your results.

Ross: you mention moving the time slider in youtube videos can sometimes trigger your audio problem. It is necessary to interrogate you further before I can let you leave: Has this problem ever triggered at another time/in a different situation? What were these situations? Once the problem is triggered, how do you stop it? Is it as simple as repositioning the slider again or do you have to reboot your computer? Are you watching youtube videos using flash player or the htm5 player? You can right click a playing youtube video to find out. ] I concur with some of Presence's ideas: A good visual inspection for any leaky/popped parts is worthwhile. While you are there you may as well re-socket the sound card and do the same for your RAM for good measure. Intermittently faulty contacts can cause all manner of annoyingly unique problems. Electrical interference from other cards, nearby fluorescent lights, electric fans and HAM broadcasters next door are all also possible.

Fifty crackheads are eating meth in overgrown seductive chocolate pizza inside your emulator.

Interesting. Thanks for making and sharing. Feature request: ability to disable mouse navigation

How to stop the 'Press F6 to return to the road' message annoying you Or "how I learned a little bit of x86 today". Open up TD3.EXE with a hex editor (make a backup first!). Go to address 00015981 The surrounding bytes should look like: ... D1 7C E2 33 C0 ... Please tell me if they don't. I have an inkling my TD3.exe is different to other people's, but I could be incorrect. Replace the "7C E2" with "90 90" ... D1 90 90 33 C0 ... Save the executable. Run the game Tell me if it works for you (this step cannot be skipped without the game crashing) If you are curious to what this does My interpretation is likely riddled with inaccuracies, so please correct me if I'm wrong Whilst the game is displaying the "Press F6 to return to the road" message it's running a small loop of instructions. Until this loop ends the game won't proceed. Normally a number is decremented (I think) every time the loop is run. At the end of the loop the game checks to see if this number has reached a certain value, if it has not it repeats the loop. By running this loop many times the game "wastes time" keeping the message on-screen instead of running the normal game instructions. Once the number reaches a certain value the game moves on and the loop is not run again (until we next veer off the road). "7C E2" is a 'conditional jump' instruction that tells your processor to jump back to the start of the loop and run it again. "90" or 'NOP' is a no-operation (do nothing) instruction. By replacing the jump with a couple of NOPs we ensure the loop is only ever run once. It's interesting that this mod stops the message appearing at all: it doesn't even flicker for a frame as far as I can see. Perhaps the loop needs to be run twice for the message to be properly rendered on-screen. You will still see the message pop up, but it'll only last a frame and should not interrupt you.

This might be an interesting episode. If you want to reference me my name is [redacted for a thought experiment, pm if this is a problem]. Ripa and Komojo have done much more than make noisy pictures, however

It's interesting how varied these episodes are. I was really hoping for a climactic evil revelation in this episode and was disappointed to discover that their evil influence apparently takes decades. Why bother planning a harvest when other people will beat you to it? Additionally: I swear I've seen that red ball guy somewhere before. Is he featured by any other authors? There is no such thing as a disk image when it comes to floppies... Just grabbing all the files from the disk is all there is to it. (you obviously never dealt with them in person) Floppies were commonly formatted with FAT12. Making images of them is a common way of backing up and distributing floppies these days -- it has a few advantages, some of them unexpected (eg subtle copy protection) and others just convenience (mounting in some emulators).

DATAA.DAT (least interesting) DATAB.DAT (most interesting) DATAC.DAT (small compared to the others) Award for the least interesting pictures in this thread!