True, but most gamers don't play drums and aren't interested in learning. A game where the pedals are played automatically would allow you to concentrate on the hands, and play far more impressive rhythms without having to be an actual drummer.
Accompany the work with the complete corresponding machine-readable source code for the Library including whatever changes were used in the work (which must be distributed under Sections 1 and 2 above); and, if the work is an executable linked with the Library, with the complete machine-readable "work that uses the Library", as object code and/or source code, so that the user can modify the Library and then relink to produce a modified executable containing the modified Library. (It is understood that the user who changes the contents of definitions files in the Library will not necessarily be able to recompile the application to use the modified definitions.)
Static linking is allowed as long as you provide the object code so users can relink with a modified library.
It's possible to distribute statically linked LGPLed binaries, as long as you include in the distribution all the.o files and library source needed to make it possible to relink with a modified library.
Clearly you've never watched higher frame rate movies. The human eye can measurably see improvement in smoothness at up to *at least* 72fps, even on video with motion blur. Movies at 24fps look jerky, it's only because we see so many of them that it's considered acceptable.
Re:Most ridiculous piece of hardware ever concepte
on
PS3 To Run At 120 FPS?
·
· Score: 1
Yes, we can only discern that many colors, but sRGB color space is not perceptually uniform, so when mapped onto a color space that better matches our vision parts of it do not have enough resolution. 24 bit color would be enough if it was in a better color space, eg. CIEL*a*b*.
You test isn't very good because it is possible to subconsciously cheat by predicting where the discontinuities should be. I think it is possible to detect discontinuities in a 24 bit sRGB color gradient, but this is a better way to test this:
Construct a cardboard or thick paper covering for the monitor, so only a small window in the middle is visible. Generate a gradient like you described, and offset it by a random amount. Make software to allow you to click on it, and detect whether you clicked on a discontinuity or not. Test multiple different gradients, as some will be much easier than others.
Re:I agree, but think you disproved your own metap
on
PS3 To Run At 120 FPS?
·
· Score: 1
It wouldn't flicker, but it would not be perfectly smooth. We accept that a 60Hz CRT flickers. We also accept that this flicker is constant. Therefore we are seeing at least 2 states (light and dark) for every screen refresh. Therefore the human eye can see equivalent to at least 120 frames per second (remember that the eye is an analog device and doesn't actually scan in frames, we would see beat patterns in the CRT flicker if it did).
120 fps animation may not *always* be distinguishable from 120fps, but when declaring a lossy codec transparent (and video is always a lossy version of real life), we must consider the worst case scenario. Hence mp3 is called transparent at 256kbps not 192kbps, even though 192kbps sounds transparent for a great many sounds. This is why Douglas Trumbull's research is potentially misleading, as while I do not know the exact details of his work, I would not be at all surprised if he only tested real video, and he certainly didn't test modern FPSs. A FPS computer game features much faster motion and camera movement than would be possible in real life.
Re:I agree, but think you disproved your own metap
on
PS3 To Run At 120 FPS?
·
· Score: 1
Yes, that's what I've been saying all along. But the argument is about whether 60Hz animation is smooth enough, not whether a 60Hz CRT flickers.
Re:I agree, but think you disproved your own metap
on
PS3 To Run At 120 FPS?
·
· Score: 1
To clarify how CRTs not drawing the whole frame at the same time is a straw man, consider three waveforms: =| ___ ___ A| | | | | =|___| |___| |__ = ___ ___ B | | | | | = |___| |___| | = _ _ _ _ C| | | | | | | | | | =|_| |_| |_| |_| |_|
Just because a sensor cannot distinguish between A and B, does not mean it cannot distinguish between A and C.
Re:I agree, but think you disproved your own metap
on
PS3 To Run At 120 FPS?
·
· Score: 1
A CRT does display frames, only they are sloping past to future from top to bottom. Because the eye does not scan in frames the time gradient is not visible, so this is irrelevant.
60Hz CRT flicker is a separate problem from 60Hz animation not being smooth enough, I am only using it as an example to show that 60Hz animation is not smooth enough without you having to test it.
Re:I agree, but think you disproved your own metap
on
PS3 To Run At 120 FPS?
·
· Score: 1
You're agreeing with me without realizing it. Yes, the progressive scanning it what is responsible for the changing brightness, which is equivalent to at least 120 updates per second for each pixel on a 60Hz CRT monitor. An LCD does not flicker at 60Hz because there is no change in brightness.
Re:I agree, but think you disproved your own metap
on
PS3 To Run At 120 FPS?
·
· Score: 1
24fps looks terrible. It's tolerable because of the motion blur, but it looks very far from real. 24fps looks ok for non-realistic animation, because then you are not comparing it to real life.
Re:I agree, but think you disproved your own metap
on
PS3 To Run At 120 FPS?
·
· Score: 1
I never said full frames *at once*, but it is full frames all the same. It doesn't matter much that the top of the frame is drawn a little bit earlier than the bottom because the eye is an analogue device, it does not scan in frames. Each pixel on the screen is updated at 60fps or 144fps, and that difference is clearly visible.
Re:I agree, but think you disproved your own metap
on
PS3 To Run At 120 FPS?
·
· Score: 1
Did you even read the parent post? It's talking about *synced* to vertical refresh framerates, so the CRT most certainly is displaying full frames.
Re:Most ridiculous piece of hardware ever concepte
on
PS3 To Run At 120 FPS?
·
· Score: 3, Interesting
Are you one of those people who thinks humans can't see more than 24 bit color? Display a smooth 24 bit color gradient on a good monitor (ie. a high end CRT, not a television or LCD), and look at all the steps. We need at least 30 bit color. A pity only Matrox realized that and all the other graphics card manufacturers ignore it.
Re:I agree, but think you disproved your own metap
on
PS3 To Run At 120 FPS?
·
· Score: 1
Electron gun scans left to right, top to bottom. Phosphors are excited, fade. The phosphor is excited, it fades, the beam hits it again.... To see a flicker there has to be a change in brightness. Therefore there are at least 2 states, bright and less bright. Therefore if you are seeing flicker on a 60Hz CRT your vision is capable of at least 120fps.
Get a fast CRT and test an old game (eg. Quake 3) at 60Hz and 120Hz, with details on low and vsync turned on so the frame rate stays rock solid. Then you will see that 60Hz is in fact *not* "silky smooth", and we can see faster than that. 60Hz maximum frame rate of the eye is a myth.
There is a difference between displaying both odd and even fields as half of a 30 fps frame, and alternating displaying an odd and even field each as half a 60fps frame. The former will look jerky, the latter will display artifacts on motion. The latter is what is called "60fps" on consoles, and while both look much worse than what you'd get on a progressive scan monitor it is the best you can do unless you want to halve vertical resolution.
Re:I agree, but think you disproved your own metap
on
PS3 To Run At 120 FPS?
·
· Score: 0, Flamebait
Absolute bullshit, and you obviously haven't tested it. Proof: a 60Hz CRT looks flickery, each flicker has to be light *and* dark, therefore equivilant to at least 2 frames, therefore human eyes can see at least 120fps. Get a fast CRT and *test* fps perception before you keep repeating this stupid myth.
I can tell the difference between 75Hz and 100Hz animation on my CRT. This is *not* the same as "flicker-free" stills, which is not relevant on LCDs. It is generally accepted that a CRT flickers at 60Hz, and for a flicker to be seen there has to be a change from light to dark. This means that each flicker is actually equivilant to *two* frames, so the maximum animation frame rate for which increasing does not improve perceptible smoothness must be around 120Hz. In practice I think 100Hz is enough.
Interlaced telivision looks terrible, and 50Hz update hurts my eyes. If you think television is acceptable quality then you must not be the kind of person who cares about fast graphics, and would be better off with an LCD.
Still maximum displayed frame rate of 75fps, you can definitely tell the difference between 75fps and 100fps. Q4 is also frame rate capped, and BF2 is not the type of twitch gaming I'm talking about, so those are not good examples. Try playing Q3 1v1 instagib, and see how every frame counts.
Hz does apply to an LCD in that displayed FPS can never be higher than display update Hz. A LCD won't flicker at 60Hz, but it won't give you perfectly smooth movement either.
If you play fast action games then CRTs are still the best, and until LCDs can display at refresh rates of at least 100Hz without blurring they always will be. People saying you can't notice the difference with frame rates higher than 60Hz obviously haven't tested it. CRTs are also best for playing older or emulated games that require a low resolution.
Ah, you are correct. Probability of 12 or more sucesses is required, and as only presence of difference was tested two-tailed significance is needed here, so probability of 3 or less needs to be added also. Thanks for the correction.
Correct significance (probability of Type I error, falsely rejecting the null hypothesis that it is impossible to tell the difference between 0 and 1 frames of added control latency) is 0.035, which is still considered statistically significant.
Yes, it is that probability of getting 12 correct answers in 15 trials. This is what I got, and this is the correct statistic to consider, as the binomial probability is calculated from the number of successs *and* number of failures.
No, it would be extremely significant if I got 12 out of 1500 trials correct, as it would mean not only could I highly reliably tell the difference, but I was also subconsciously swapping "matched" with "unmatched". It would be a very weird and interesting result.
A lot of people say very small changes in lag don't make a difference, but I think they are confusing it with reaction time. It's true that you're not physically able to react to visual stimulus in less than about 150ms (reaction to sound is slightly faster, reaction to touch faster still), but we are certainly capable of detecting differences in response times at much finer granularity than that.
This is something I have actually tested. I set up a white square on a black background, moveable using the arrow keys. It was shown on a CRT running at 60Hz, synced to vertical refresh so each frame was exactly 1/60s. I implemented an ABX test, with one of control methods A and B being lagged by 1 frame, and the other not lagged (chosen at random). X randomly matched either A or B, and I could switch between them at will so long as the square was not moving. I tested myself with 15 trials, with no feedback on score until all the trials were complete.
It turned out to be exceedingly difficult to tell the difference, but it was possible. I successfully matched 12 of the 15 trials, and as guessing would produce 50% success we can use the binomial theorem to calculate the probability that I was just lucky.
p = (15 nCr 12) * 0.5^12 * 0.5^3 = 0.014
A statistically significant (p>0.05) result, so it is overwhelmingly likely that 16.7ms change in latency is perceptible. It's subtle, but the "feel" is noticeably different. However, I think the oldness and softness of my keyboard made things much harder, and I'm certain I would not have been able to tell the difference after drinking even a small amount of alcohol. Room for further testing here!
Slashdot Mirror
True, but most gamers don't play drums and aren't interested in learning. A game where the pedals are played automatically would allow you to concentrate on the hands, and play far more impressive rhythms without having to be an actual drummer.
Grafting is possible, but you'll get perfectly ordinary THC-free hops so there's no reason to.
It's possible to distribute statically linked LGPLed binaries, as long as you include in the distribution all the .o files and library source needed to make it possible to relink with a modified library.
Clearly you've never watched higher frame rate movies. The human eye can measurably see improvement in smoothness at up to *at least* 72fps, even on video with motion blur. Movies at 24fps look jerky, it's only because we see so many of them that it's considered acceptable.
Yes, we can only discern that many colors, but sRGB color space is not perceptually uniform, so when mapped onto a color space that better matches our vision parts of it do not have enough resolution. 24 bit color would be enough if it was in a better color space, eg. CIEL*a*b*. You test isn't very good because it is possible to subconsciously cheat by predicting where the discontinuities should be. I think it is possible to detect discontinuities in a 24 bit sRGB color gradient, but this is a better way to test this: Construct a cardboard or thick paper covering for the monitor, so only a small window in the middle is visible. Generate a gradient like you described, and offset it by a random amount. Make software to allow you to click on it, and detect whether you clicked on a discontinuity or not. Test multiple different gradients, as some will be much easier than others.
It wouldn't flicker, but it would not be perfectly smooth. We accept that a 60Hz CRT flickers. We also accept that this flicker is constant. Therefore we are seeing at least 2 states (light and dark) for every screen refresh. Therefore the human eye can see equivalent to at least 120 frames per second (remember that the eye is an analog device and doesn't actually scan in frames, we would see beat patterns in the CRT flicker if it did).
120 fps animation may not *always* be distinguishable from 120fps, but when declaring a lossy codec transparent (and video is always a lossy version of real life), we must consider the worst case scenario. Hence mp3 is called transparent at 256kbps not 192kbps, even though 192kbps sounds transparent for a great many sounds. This is why Douglas Trumbull's research is potentially misleading, as while I do not know the exact details of his work, I would not be at all surprised if he only tested real video, and he certainly didn't test modern FPSs. A FPS computer game features much faster motion and camera movement than would be possible in real life.
Yes, that's what I've been saying all along. But the argument is about whether 60Hz animation is smooth enough, not whether a 60Hz CRT flickers.
To clarify how CRTs not drawing the whole frame at the same time is a straw man, consider three waveforms:
=| ___ ___
A| | | | |
=|___| |___| |__
= ___ ___
B | | | | |
= |___| |___| |
= _ _ _ _
C| | | | | | | | | |
=|_| |_| |_| |_| |_|
Just because a sensor cannot distinguish between A and B, does not mean it cannot distinguish between A and C.
A CRT does display frames, only they are sloping past to future from top to bottom. Because the eye does not scan in frames the time gradient is not visible, so this is irrelevant.
60Hz CRT flicker is a separate problem from 60Hz animation not being smooth enough, I am only using it as an example to show that 60Hz animation is not smooth enough without you having to test it.
You're agreeing with me without realizing it. Yes, the progressive scanning it what is responsible for the changing brightness, which is equivalent to at least 120 updates per second for each pixel on a 60Hz CRT monitor. An LCD does not flicker at 60Hz because there is no change in brightness.
24fps looks terrible. It's tolerable because of the motion blur, but it looks very far from real. 24fps looks ok for non-realistic animation, because then you are not comparing it to real life.
I never said full frames *at once*, but it is full frames all the same. It doesn't matter much that the top of the frame is drawn a little bit earlier than the bottom because the eye is an analogue device, it does not scan in frames. Each pixel on the screen is updated at 60fps or 144fps, and that difference is clearly visible.
Did you even read the parent post? It's talking about *synced* to vertical refresh framerates, so the CRT most certainly is displaying full frames.
Are you one of those people who thinks humans can't see more than 24 bit color? Display a smooth 24 bit color gradient on a good monitor (ie. a high end CRT, not a television or LCD), and look at all the steps. We need at least 30 bit color. A pity only Matrox realized that and all the other graphics card manufacturers ignore it.
Electron gun scans left to right, top to bottom. Phosphors are excited, fade. The phosphor is excited, it fades, the beam hits it again.... To see a flicker there has to be a change in brightness. Therefore there are at least 2 states, bright and less bright. Therefore if you are seeing flicker on a 60Hz CRT your vision is capable of at least 120fps. Get a fast CRT and test an old game (eg. Quake 3) at 60Hz and 120Hz, with details on low and vsync turned on so the frame rate stays rock solid. Then you will see that 60Hz is in fact *not* "silky smooth", and we can see faster than that. 60Hz maximum frame rate of the eye is a myth.
There is a difference between displaying both odd and even fields as half of a 30 fps frame, and alternating displaying an odd and even field each as half a 60fps frame. The former will look jerky, the latter will display artifacts on motion. The latter is what is called "60fps" on consoles, and while both look much worse than what you'd get on a progressive scan monitor it is the best you can do unless you want to halve vertical resolution.
Absolute bullshit, and you obviously haven't tested it. Proof: a 60Hz CRT looks flickery, each flicker has to be light *and* dark, therefore equivilant to at least 2 frames, therefore human eyes can see at least 120fps. Get a fast CRT and *test* fps perception before you keep repeating this stupid myth.
I can tell the difference between 75Hz and 100Hz animation on my CRT. This is *not* the same as "flicker-free" stills, which is not relevant on LCDs. It is generally accepted that a CRT flickers at 60Hz, and for a flicker to be seen there has to be a change from light to dark. This means that each flicker is actually equivilant to *two* frames, so the maximum animation frame rate for which increasing does not improve perceptible smoothness must be around 120Hz. In practice I think 100Hz is enough. Interlaced telivision looks terrible, and 50Hz update hurts my eyes. If you think television is acceptable quality then you must not be the kind of person who cares about fast graphics, and would be better off with an LCD.
Still maximum displayed frame rate of 75fps, you can definitely tell the difference between 75fps and 100fps. Q4 is also frame rate capped, and BF2 is not the type of twitch gaming I'm talking about, so those are not good examples. Try playing Q3 1v1 instagib, and see how every frame counts.
Hz does apply to an LCD in that displayed FPS can never be higher than display update Hz. A LCD won't flicker at 60Hz, but it won't give you perfectly smooth movement either.
If you play fast action games then CRTs are still the best, and until LCDs can display at refresh rates of at least 100Hz without blurring they always will be. People saying you can't notice the difference with frame rates higher than 60Hz obviously haven't tested it. CRTs are also best for playing older or emulated games that require a low resolution.
Ah, you are correct. Probability of 12 or more sucesses is required, and as only presence of difference was tested two-tailed significance is needed here, so probability of 3 or less needs to be added also. Thanks for the correction.
Correct significance (probability of Type I error, falsely rejecting the null hypothesis that it is impossible to tell the difference between 0 and 1 frames of added control latency) is 0.035, which is still considered statistically significant.
Yes, it is that probability of getting 12 correct answers in 15 trials. This is what I got, and this is the correct statistic to consider, as the binomial probability is calculated from the number of successs *and* number of failures. No, it would be extremely significant if I got 12 out of 1500 trials correct, as it would mean not only could I highly reliably tell the difference, but I was also subconsciously swapping "matched" with "unmatched". It would be a very weird and interesting result.
Dumb typo. Should have been 0.05
A lot of people say very small changes in lag don't make a difference, but I think they are confusing it with reaction time. It's true that you're not physically able to react to visual stimulus in less than about 150ms (reaction to sound is slightly faster, reaction to touch faster still), but we are certainly capable of detecting differences in response times at much finer granularity than that.
This is something I have actually tested. I set up a white square on a black background, moveable using the arrow keys. It was shown on a CRT running at 60Hz, synced to vertical refresh so each frame was exactly 1/60s. I implemented an ABX test, with one of control methods A and B being lagged by 1 frame, and the other not lagged (chosen at random). X randomly matched either A or B, and I could switch between them at will so long as the square was not moving. I tested myself with 15 trials, with no feedback on score until all the trials were complete.
It turned out to be exceedingly difficult to tell the difference, but it was possible. I successfully matched 12 of the 15 trials, and as guessing would produce 50% success we can use the binomial theorem to calculate the probability that I was just lucky.
p = (15 nCr 12) * 0.5^12 * 0.5^3
= 0.014
A statistically significant (p>0.05) result, so it is overwhelmingly likely that 16.7ms change in latency is perceptible. It's subtle, but the "feel" is noticeably different. However, I think the oldness and softness of my keyboard made things much harder, and I'm certain I would not have been able to tell the difference after drinking even a small amount of alcohol. Room for further testing here!