Temporal denoiser concept - "hush"

[Katie Boundary]

An ideal temporal denoiser is one that yields the greatest ratio of denoising to ghosting. If you have the same mental illnesses that I do, then you'll probably also want it to be fast, have no arbitrary modes or thresholds to configure, and not try to be "too smart" in ways that have the potential to blow up in your face (motion-compensation and crap like that). After many hours of thought, I've developed a concept that I'd like to submit for consideration. The first name that I wanted to pick for it was "Glass", but I've since settled on "Hush". It's optimized for deflickering of bob-deinterlaced material but should work fairly well on everything.

Let's use the common convention of C being the current pixel being denoised, P being the pixel in the previous frame and same location, and N being the pixel in the next frame and same location. First, the filter calculates the differences in the values of these pixels. Then, it makes a decision based on how different or similar they are:

If P and N are more similar to each other than either one is to C, the filter will assume that the pixel is "noise", or that there is a flicker or strobe effect happening. It will then replace the pixel's current value with (P+2C+N)/4.

If P and C are the most similar to each other, and C and N are the most different, the filter will assume that C is in the last pixel before a sharp change, and replace it with (P+C)/2. If C and N are the most similar to each other, and P and C are the most different, then the filter will assume that C is in the first pixel after a sharp change, and replace it with (C+N)/2

If P and N are farther away from each other than either one is from C, the filter could be facing a sharp change, OR a fade or other smooth motion. To figure out which is which, it compares the differences in pixel values. If the difference between P and C, and the difference between C and N, are more similar to each other than either one is to the difference between P and N, then the filter assumes a fade or other smooth motion, and does nothing. However, if C is much closer to P than it is to N or vice versa, the filter assumes a scene change, and replaces C with either (P+C)/2 or (C+N)/2, whichever is appropriate.

As you might guess, this filter would have a pretty merciless effect on any strobe or flicker effects that are actually supposed to be in the video, or any effect that cycles between the same 2 frames over and over again (like the semi-famous animated gif of the subway train that's going either forward or backward). Otherwise, I expect its harsh denoising to have little or no ghosting. I'd love to see it in action, if anyone is interested in doing the real work of programming it. There's probably some special situation in which it will fail catastrophically...

[Groucho2004]

Hm, the only way you can find out if your concept works as expected is to write a filter for Avisynth and try it out.

[Katie Boundary]

Or a plugin for Virtualdub, or any other program. Unfortunately, PyScenario scripting for Civilization IV was about as close as I ever got to learning a programming language.

The situation in which I most strongly suspect ghosting to occur would be if P, C, and N all had extremely different values, which in turn I expect would only occur in very fast-motion scenes or those with shakeycam. The situation in which I expect its denoising properties to be weakest would be when combating "splotch" noise in bob-deinterlaced content that hasn't been decimated yet, as the spatial size of the splotches means that interpolation will actually reconstruct them, and the doubling of the frame rate will result in the splotches being preserved across consecutive frames, and therefore not being considered noise.

I'm thinking that the Andromeda DVDs just need to be remastered, period...

[wonkey_monkey]

It's optimized for deflickering of bob-deinterlaced material

Just use a decent bobber. You've been told about several, several times, including some fast ones.

If P and N are more similar to each other than either one is to C, the filter will assume that the pixel is "noise", or that there is a flicker or strobe effect happening. It will then replace the pixel's current value with (P+2C+N)/4.

Great. So anything that moves by fast enough is now going to be semi-transparent.

If P and C are the most similar to each other, and C and N are the most different, the filter will assume that C is in the last pixel before a sharp change, and replace it with (P+C)/2. If C and N are the most similar to each other, and P and C are the most different, then the filter will assume that C is in the first pixel after a sharp change, and replace it with (C+N)/2

Terrible idea. A plain average of pixels from two frames? Based on what will be a common, naturally occuring situation? That's going to look awful.

If P and N are more similar to each other than either one is to C...

If P and C are the most similar to each other, and C and N are the most different...

If P and N are farther away from each other than either one is from C...

These three situations are not all mutually exclusive of each other. Your filter is not well-defined. Otherwise I could knock it out in ten minutes and show you how bad it is.

If you mean for the rules to be evaluated in that order on a first-match/discard-the-rest basis, you must specify this.

[filler56789]

http://forum.doom9.org/member.php?u=221865 says:

Join Date 27th January 2015
Total Posts 233

17 months later, ¿does she have at least one post that is not trollish?

[feisty2]

http://forum.doom9.org/member.php?u=221865 says:



17 months later, ¿does she have at least one post that is not trollish?

Dunno if she does or not, BUT, the following works,
1. Don't waste your time trying to reply any of her comments, since you already know she's a troll
2. If you do have the need for posting something, do not reply anything technical or related to her trolling statements, just warn others that she's a troll and no one should waste any time to reply and fall for her trolls.

[johnmeyer]

An ideal temporal denoiser is one that yields the greatest ratio of denoising to ghosting. Ive never used a temporal denoiser that created ghosting. I've seen plenty of other unwanted artifacts, such as noise correlation and clumping, but never any ghosting.... have the potential to blow up in your face (motion-compensation and c*** like that).I don't appreciate foul language. It sounds like a teenage punk boy, not an adult woman.

And, to your point, motion compensation is at the heart of any decent denoiser. In fact, in order to do what you describe, you must use motion compensation. Without it, "P" and "N" will have no relation to "C."

Temporal denoising without motion compensation won't work very well, and may not work at all. So, if you don't like motion compensation, for whatever reason, you'd better stick to spatial denoising.
If P and N are more similar to each other than either one is to C, the filter will assume that the pixel is "noise", ...What you describe is too basic to be useful, even as a general idea. The basic problem is that most noise is not confined to one pixel. Life sure would be easy if that were so. Also, there are lots of reasons, besides noise, that will lead to [ P=N & (C<>P & C<>N) ].As you might guess, this filter would have a pretty merciless effect on any strobe or flicker ...No it wouldn't. I know of no instance of intentional strobing, or flickering, that have a duration of only one frame. In fact, if you strobe at 30 Hz (for NTSC), you won't see strobing because your persistence of vision will average out the on and off.

[burfadel]

If the filter looks at each pixel, then the changes of the pixels around it will indicate intentional flicker. Maybe if the luminance is detected as changing in a 9x9 square spatially, then maybe a more consistent luminance can be applied to the pixel.

[Katie Boundary]

17 months later, ¿does she have at least one post that is not trollish?

All but 2 of them.

The real question is, why are you trolling in what is obviously a serious thread?

Ive never used a temporal denoiser that created ghosting. I've seen plenty of other unwanted artifacts, such as noise correlation and clumping, but never any ghosting.

Try TemporalCleaner(ythresh=10,cthresh=20)

Or use even bigger values. You'll get your ghosting.

I don't appreciate foul language. It sounds like a teenage punk boy, not an adult woman.

"Crap" is not foul language. It's one of those lame words like "darn" and "heck" that people use in place of actual foul language. If you want to see what foul language looks like, visit my Facebook page.

In fact, in order to do what you describe, you must use motion compensation.

Really? Because I'm pretty sure that not a single one of the functions that I specified had anything to do with motion at all.

So, if you don't like motion compensation, for whatever reason, you'd better stick to spatial denoising.

This thread was inspired by a very specific type of noise that I found on the Andromeda DVDs that is basically unkillable by any kind of spatial denoising method, and which exists even in scenes with no motion.

What you describe is too basic to be useful, even as a general idea. The basic problem is that most noise is not confined to one pixel.

That's a problem for spatial noise, not temporal noise.

No it wouldn't. I know of no instance of intentional strobing, or flickering, that have a duration of only one frame.

Then you obviously haven't gone frame-by-frame through many animated shows.

In fact, if you strobe at 30 Hz (for NTSC), you won't see strobing because your persistence of vision will average out the on and off.

LOL. I've seen proof to the contrary. You can see such proof too by running literally any DVD footage through a dumb bob-deinterlacer.

[manono]

Try TemporalCleaner(ythresh=10,cthresh=20)

Or use even bigger values. You'll get your ghosting.

Yeah, but he said, "Ive never used a temporal denoiser that created ghosting." So, if he's never used TemporalCleaner then his statement could very well be correct. And you, why use inferior cleaners (as you've so often used other inferior filters) when there are much better temporal (and 3D) cleaners available for AviSynth? Is it because the good ones are slow? If so, let them work overnight. Good results for this kind of thing take time. I used to use TemporalCleaner myself, but that was a long time ago. Much better ones have come along since then.

And I agree with you that 'crap' is minor league.

[Groucho2004]

And you, why use inferior cleaners (as you've so often used other inferior filters) when there are much better temporal (and 3D) cleaners available for AviSynth? Is it because the good ones are slow?
His signature provides a hint.
I prefer my filters to be simple, fast, and dumb...

[colours]

Or a plugin for Virtualdub, or any other program. Unfortunately, PyScenario scripting for Civilization IV was about as close as I ever got to learning a programming language.

Then learn something.

Merely having a bunch of ideas is useless if you can't even test them. Do you expect other Doom9 posters to write filters for you pro bono?

[wonkey_monkey]

Then learn something.

Merely having a bunch of ideas is useless if you can't even test them. Do you expect other Doom9 posters to write filters for you pro bono?

I already offered (though perhaps not with an entirely altruistic motive), subject to clarification of the idea. Maybe I'm still on KB's curiously selective ignore.

[filler56789]

1. Don't waste your time trying to reply any of her comments, since you already know she's a troll
2. If you do have the need for posting something, do not reply anything technical or related to her trolling statements, just warn others that she's a troll and no one should waste any time to reply and fall for her trolls.

I agree with you. Her name is on my IgnoreList (again). But IgnoreLists don't do anything when someone else quotes a troll... Killfiles are better than IgnoreLists though (I.I.R.C.), because they can "erase" an entire thread started by a troll.

[StainlessS]

"Crap" is not foul language. It's one of those lame words like "darn" and "heck" that people use in place of actual foul language. If you want to see what foul language looks like, visit my Facebook page.

Oh I see, you annoy the hell out of your FaceBook visitors too, still no reason for them to swear at you though :)
(sorry could not resist)

"Crap" is not foul language

Well actually have to agree a bit there:- https://en.wikipedia.org/wiki/Thomas_Crapper

Thomas Crapper was a plumber/ water closet manufacturer.

If his name was Smith, we would probably be saying, "I'm just gonna have a quick smith".

EDIT: By the way, Bollocks are little wooden balls (castors on a metal ring) that facilitate running a sail up and down a mast.

And the word Wank comes from the name of a German engineering company that made the Wankel Engine
and Wankel speed governor used on maritime engines (I saw one once, instead of a spinning centrifugal force based governor
using a pair of heavy-ish balls, had instead a little metal finger that waggled back and forth [hence the term]).

EDIT: Seems that Crap may not be based on Thomas Crapper, from Wikipedia link above
Origin of the word "crap"

It has often been claimed in popular culture that the slang term for human bodily waste, crap, originated with Thomas Crapper because of his association with lavatories. A common version of this story is that American servicemen stationed in England during World War I saw his name on cisterns and used it as army slang, i.e. "I'm going to the crapper".[10]

The word crap is actually of Middle English origin and predates its application to bodily waste. Its most likely etymological origin is a combination of two older words, the Dutch krappen: to pluck off, cut off, or separate; and the Old French crappe: siftings, waste or rejected matter (from the medieval Latin crappa, chaff).[10] In English, it was used to refer to chaff, and also to weeds or other rubbish. Its first application to bodily waste, according to the Oxford English Dictionary, appeared in 1846 under a reference to a crapping ken, or a privy, where ken means a house.[10]

[Groucho2004]

instead of a spinning centrifugal force based governor using a pair of heavy-ish balls, had instead a little metal finger that waggled back and forth.
This could be straight from a freaky sex toy manual.

[Groucho2004]

And the word Wank comes from the name of a German engineering company that made the Wankel Engine
Hercules once made a motorcycle with a Wankel engine:
https://upload.wikimedia.org/wikipedia/de/e/e5/Herculesw2000inj.jpg


We're still on topic since this motorcycle produced very little noise. :D

[hello_hello]

As you might guess, this filter would have a pretty merciless effect on any strobe or flicker effects that are actually supposed to be in the video, or any effect that cycles between the same 2 frames over and over again (like the semi-famous animated gif of the subway train that's going either forward or backward). Otherwise, I expect its harsh denoising to have little or no ghosting. I'd love to see it in action, if anyone is interested in doing the real work of programming it. There's probably some special situation in which it will fail catastrophically...

Continuing what must now be a record for the number of consecutive posts without being correct, I'm fairly sure the image Katie is referring to has four unique frames and the train was moving forward. The effect..... it probably has a name I can't recall.... is where the motion of an object is too fast to be captured accurately at a particular frame rate. Probably the most common example is a turning wheel. It can appear to be turning more slowly than it is, stationary, or turning in the wrong direction.

A "'change direction" gif.
https://s31.postimg.org/6okuoxqvv/Tnfzr_DD.gif

The individual frames:
Frame 1:
https://s32.postimg.org/6qlhd6n85/Tnfzr_DD_frame_0001.gif

Frame 2:
https://s31.postimg.org/eqkizobrf/Tnfzr_DD_frame_0002.gif

Frame 3:
https://s32.postimg.org/nh7x6lgn9/Tnfzr_DD_frame_0003.gif

Frame 4:
https://s32.postimg.org/n9b7z4ur9/Tnfzr_DD_frame_0004.gif

iLOL. I've seen proof to the contrary. You can see such proof too by running literally any DVD footage through a dumb bob-deinterlacer.

If shimmering and strobing are the same thing, why do they have two different names?

[filler56789]

The real question is, why are you trolling in what is obviously a serious thread?

The real question is, ¿why no moderator has banned you yet?

[Sharc]

..... The effect..... it probably has a name I can't recall.... is where the motion of an object is too fast to be captured accurately at a particular frame rate. Probably the most common example is a turning wheel. It can appear to be turning more slowly than it is, stationary, or turning in the wrong direction.....

It's called Aliasing. Happens when the sampling rate (in this case the frame rate) is too low (Nyquist-Shannon-Kotelnikov Sampling Theorem).
https://en.wikipedia.org/wiki/Nyquist%E2%80%93Shannon_sampling_theorem

[filler56789]

And the word Wank comes from the name of a German engineering company that made the Wankel Engine
and Wankel speed governor used on maritime engines (I saw one once, instead of a spinning centrifugal force based governor
using a pair of heavy-ish balls, had instead a little metal finger that waggled back and forth [hence the term]).

Thanks for the info. The first time I read about the Wankel engine was in 1972, when I was 9 years old, through a magazine printed by the French consulate in São Paulo.

To whom this might interest :)

https://www.youtube.com/watch?v=B4DxCX3XLpc

[hello_hello]

It's called Aliasing. Happens when the sampling rate (in this case the frame rate) is too low (Nyquist-Shannon-Kotelnikov Sampling Theorem).
https://en.wikipedia.org/wiki/Nyquist%E2%80%93Shannon_sampling_theorem

Ahh.... of course.... although following your link led me to discover the example of aliasing I offered is referred to as the wagon-wheel effect, or a stroboscopic effect.... according to Wikipedia.

Wikipedia refers to persistence of vision causing strobing to appear continuous at around 60 "flashes" per second, but thinking about it, if the NTSC equivalent of a "flash" is a single screen refresh (60Hz) then I'll confess it caused the comment below to rattle around my brain and eventually pop up for review:

No it wouldn't. I know of no instance of intentional strobing, or flickering, that have a duration of only one frame. In fact, if you strobe at 30 Hz (for NTSC), you won't see strobing because your persistence of vision will average out the on and off.

My chain of thought.....

60 flashes per second = a continuous, non-strobng image.
The NTSC equivalent of 60 flashes per second is 60 screen refreshes per second ie 60Hz.

Initially I assumed a single "flash" would consist of an "on/off" cycle of light, requiring an "on" frame and an "off" frame, but even at 60fps that strobes way too much to be watchable (I tested it), which means the number of screen flashes per second were effectively halved (every second refresh displaying a black frame), therefore for the TV to "flash" 60 times per second, at 60fps every frame must represent an "on/off" cycle, which means realistically every frame must be an "on" frame. At 30fps the same applies, if each frame hangs around for two screen "flashes"
After walking that one around in a circles a couple of times, I still can't find a definition for "strobe at 30Hz" that wouldn't strobe like an 80's disco. Am I missing something?

[Katie Boundary]

By the way, here's an example of what I call "splotch" noise:

https://scontent-sjc2-1.xx.fbcdn.net/v/t1.0-9/13417437_1701546320109823_5536313367009351414_n.jpg?oh=11dea4f8d638521fdbcc4acde170bd9c&oe=57D4F43C

Here it is after applying Blur(1), then downsizing to 512x288, then applying Temporalcleaner at y=8 and c=16:

https://scontent-sjc2-1.xx.fbcdn.net/v/t1.0-9/13432139_1701542826776839_8547045875351206569_n.jpg?oh=af22d0572637021d5e2a19e96c1f8a37&oe=57D3F008

It still doesn't kill the damn noise (just makes it less sharp), which is why I'm looking for a more aggressive approach.

Then learn something.

Merely having a bunch of ideas is useless if you can't even test them.

If I wasn't already juggling 50 other competing demands, ranging from World of Warcraft to filling out these job applications that have been sitting in my backpack for the past 2 months, I'd totally be doing that.

Do you expect other Doom9 posters to write filters for you pro bono?

As opposed to all the people who got paid for their contributions to the AVIsynth project... :rolleyes:

This could be straight from a freaky sex toy manual.

LOL. Do you know what the difference is between you and all the other people who clog up my threads with worthless spam? Your attempts to be funny are actually successful :)

The real question is, ¿why no moderator has banned you yet?

They have no grounds for doing so.

[colours]

If I wasn't already juggling 50 other competing demands, ranging from World of Warcraft to filling out these job applications that have been sitting in my backpack for the past 2 months, I'd totally be doing that.

Cool excuses.

If you really care about quality, the hours you spend learning AviSynth and implementing your idea will pay off almost immediately, and pay off across all the videos you want to filter and transcode.

(This is leaving aside the fact that your description of the transcoding process sounds like all the ways it could be done wrong and then some, but you refuse to fix all that wrongness despite everyone telling you to.)

As opposed to all the people who got paid for their contributions to the AVIsynth project... :rolleyes:

I obviously can't speak for everyone, but the majority of scripts I write are for my own use. If I expect them to be generally useful, I upload them somewhere as a public service. Sure, I don't get paid, but I treat this as a hobby.

I do sometimes write scripts for other people, but that depends a lot on how useful I expect the resulting script to be. If I expect the script to be completely useless (e.g. if someone asks me to write a filter that chains arbitrarily many invocations of sharpen(1)), I won't bother. Not even if it's a request that can be implemented in a minute.

There is an exception to be made for interesting ideas, of course. This YCbCr range limiter (http://forum.doom9.org/showthread.php?p=1711269#post1711269) is slow and practically useless, but I wrote it just as a proof of concept.

Other people have already pointed out potential problems with "hush" and also that your idea isn't even sufficiently well defined to implement as a filter. Perhaps you even have those people on your ignore list for having committed the sin of daring to disagree with you.

[filler56789]

If you want to see what foul language looks like, visit my Facebook page.

Been there, blocked her :devil:
Conservatrolls are annoying anyway.

[Katie Boundary]

(This is leaving aside the fact that your description of the transcoding process sounds like all the ways it could be done wrong and then some, but you refuse to fix all that wrongness despite everyone telling you to.)

As far as I can tell, I'm the only person so far who has mentioned its major weaknesses.

If I expect the script to be completely useless (e.g. if someone asks me to write a filter that chains arbitrarily many invocations of sharpen(1)), I won't bother.

That sounds so awesomely bad that I'd actually write it just to see what kind of insanity would result :D

your idea isn't even sufficiently well defined to implement as a filter.

Nah, I defined it pretty damn specifically.

Perhaps you even have those people on your ignore list for having committed the sin of daring to disagree with you.

I have never put anyone on ignore just for that. There are only 3 people on my ignore list, and in all 3 cases, it was for inexcusable levels of language comprehension failure, usually mixed with strawman fallacies (it's sometimes hard to tell when someone else is failing to think rationally or has just given up on trying altogether).

Conservatrolls are annoying anyway.

I'm a libertarian, not a conservative. Just today, I spent a few posts talking trash about Christianity... but let's not drag politics and religion into these forums.

[QBhd]

I have never put anyone on ignore just for that. There are only 3 people on my ignore list, and in all 3 cases, it was for inexcusable levels of language comprehension failure, usually mixed with strawman fallacies (it's sometimes hard to tell when someone else is failing to think rationally or has just given up on trying altogether).


Funniest thing I have read all year!

QB

[wonkey_monkey]

Nah, I defined it pretty damn specifically.

No, you did not, and I told you exactly why. But I guess I'm on your ignore list.

I'm a libertarian, not a conservative. Just today, I spent a few posts talking trash about Christianity...

What the hell do you think a libertarian is?!

[DarkSpace]

If I expect the script to be completely useless (e.g. if someone asks me to write a filter that chains arbitrarily many invocations of sharpen(1)), I won't bother.

That sounds so awesomely bad that I'd actually write it just to see what kind of insanity would result :D


Oh why not, I'll bite:


# This code sample is licensed under WTFPL - see http://www.wtfpl.net/ for details
function ChainSharpen(clip clip, int times, float strength) {
times = default(times, 5)
strength = default(strength, 1.0)

return (times <= 0) ? clip : ChainSharpen(clip.Sharpen(strength), times-1, strength)
}


I hereby offer my services to write your plugin (probably in the form of an avs function like the one above - except that I'll actually test whether it is syntactically correct before I give it to you, if you do choose to pay me).
Be warned, though:
1) There will be no refunds if you find that the function does not have the effect that you want. You have been warned sufficiently that your concept is flawed.
1-a) The resulting code will probably be slow to execute. Likely, unnecessarily much so.
1-b) For future reference, I will quote your exact specifications before I start coding. That way, even if you change your mind, and edit the original post, I can verify, to you and to myself, whether my code follows the instructions that I received. Additionally, I may use this as a building block for a general description of what the result does, or maybe even use it as description directly.
1-b1) You did not specify what to do about the first and last frames. Therefore, I'll probably "invent" something for these two frames...
2) My services are overpriced, and your options to pay me are limited. I propose that, if you unexpectedly choose to accept my offer, you first verify that you can pay me (going by both amount and payment method)
2-a) I will notify you when the code is done, and then wait until I have received payment. Then, I will publicly post the resulting code, without any noteworthy license restrictions. If you have a better idea to insure me against money loss (trust is not one of them), you are free to propose them. I may evaluate these ideas, then, and if I decide that I like it, I may implement it.
2-a1) For example, maybe there is a way to deposit the payment at a neutral third party that only verifies the payment's presence, but does not give me access to it until the third party can verify my code's presence also... However, I very much think that a system like this, while certainly possible, will be too costly to actually implement here without running into the abovementioned trust problems.

I fully expect you to ignore this offer (at least once you know the price), but if you unexpectedly are interested, I'll have to check some legal stuff (such as, are there any restrictions I should be careful about when doing this, or do I have to insure myself against lawsuits and such... if it turns out it would be too much effort [I hope not], I may choose to abort the project. In this case, you don't have to pay me, of course...) before responding, so my response may take some time.

[johnmeyer]

Darkspace,

This guy is a World of Warcraft addict. I didn't realize that until I read his post above (#23, unless he edits it again):

If I wasn't already juggling 50 other competing demands, ranging from World of Warcraft to filling out these job applications that have been sitting in my backpack for the past 2 months, I'd totally be doing that.

WoW addicts live in an alternate reality, and have a very difficult time dealing with real situations and actual people. It is pretty scary, actually:

World of Warcraft Addiction (https://www.google.com/search?q=world+of+warcraft+addiction&ie=utf-8&oe=utf-8)

Note that he includes WoW as a "demand." My nephew also felt this way about his comrades in WoW.

I have known a few people who got caught up in this, and they get very, very strange. In fact, someone to whom I am related by marriage actually tried to commit suicide after his mother managed to cut him off from WoW by removing all computers from the house. He has now come of age, but has failed at everything. Like our friend, he fills out job applications, but nothing ever happens.

So, I don't know if you were sincere in your offer, but my strong advice is to not spend any time actually doing any coding. This guy is not a troll, in the usual sense (although that is the effect he has on the forum), but someone who can't hold down a job because of WoW. I actually feel quite sorry for him, and am now going to completely leave him alone. He needs help, but not the kind we can give him here, and certainly not help with his various simple-minded schemes.

[Groucho2004]

I have known a few people who got caught up in this, and they get very, very strange.
Quod erat demonstrandum (https://www.youtube.com/watch?v=YersIyzsOpc). :D

[filler56789]

Darkspace,

This guy is a World of Warcraft addict.

Conservatrolls are annoying anyway.

As I always say, Keep Calm Because I'm Always Right :)

http://www.bit-tech.net/news/gaming/2009/07/30/wow-to-get-in-game-addiction-therapists/1

&&&

https://twitter.com/auntoona/status/531428333238513665

Nope, coincidences do not exist :D

[Katie Boundary]

stuff

Although I must decline your offer, mostly because I don't really feel like paying you to do something that's normally done for free, I do thank you for making the only post, out of the six posts since my last, that was actually on-topic and not trolling/spam.

Also, thanks for pointing out the issue with first and last frames. I'd just replace C with (C+N)/2 for the first and with (P+C)/2 for the last.

[wonkey_monkey]

mostly because I don't really feel like paying you to do something that's normally done for free

How arrogant. Writing filters just because other people request them is not normally done at all, and you should not expect it from anyone. We're not your personal code monkeys.

That said, I have already offered to create your filter for you to prove to you that it's a terrible idea.

Perhaps someone who isn't on KB's ignore list could tell her this, although I assume she must have read my previous post to asceratin whether or not it met her criteria for being "on-topic."

[WorBry]

This guy is a World of Warcraft addict. I didn't realize that until I read his post above (#23, unless he edits it again):


Either that or simply a compulsive attention seeker, which is what comes across to me:

http://bullyonline.org/index.php/bullying/bullies/21-the-need-for-attention

Quote: "....The troll gets a kick from single-handedly moving others to spend their time and emotional energy on him or her, and even on each other. The best way to deal with Internet trolling, and other adult behaviour which is primarily intended to elicit attention, is to not respond, to not engage and to thus deny the person the attention they seek. If the perpetrator is denied the attention, there's a chance that they will think twice before taking this approach again. If they get attention, they might well do it again."

Look how many threads she/he has going at the top of this General Discussion section alone.....five as of now. And as one fizzles out and drops down the table, no doubt she/he will start another to keep the attention coming.

I'm just sad to see the forum deteriorating with tolerance for this sort of thing.

[Katie Boundary]

Well, I used to have a policy of telling people to post something relevant or shut the hell up when they polluted my threads with personal, off-topic nonsense like you're doing now. I can see how a schizophrenic on LSD could mistake this for "attention-seeking". I've dropped this policy in favor of "report, ignore, and move on", but I'm making an exception for you, because your post is the most blatant, egregious, spammy, worthless, hypocritical attempt at thread-derailment that I've seen since coming off my 30-day suspension. It's so bad that I'm actually goignto try reporting it twice, just to see if that's possible. Also, welcome to my ignore list.

[Wilbert]

Closed thread. Feel free to come back when you implemented such a filter.