6 ms·

> "Inadequate" for gaming, yes. Inadequate for computing? Hardly. How about for watching movies? High framerates would obviously seem very important for gaming

by selmnoo 13y ago

> "Inadequate" for gaming, yes. Inadequate for computing? Hardly. How about for watching movies? High framerates would obviously seem very important for gaming, but it's not clear how important they are to movies/shows. Your thoughts?

majormajor 13y ago
Displaying 24fps content on a 60fps screen repeats frames. If you repeat half the frames of the 25fps stream twice, and alternating half of the frames three times, then you get 60 frames per second, with a stream looking like: AABBBCCDDDEEFFF... This isn't perfect, since really each frame should be on the screen for exactly the same amount of time, which is why 120hz TVs came about -- 120 is evenly divisible by each of 60, 30, and 24. But it's also pretty much good enough, or else there'd be a lot more complaining about watching movies on computers or other 60hz devices. A 30hz screen makes that harder, since you're going to have to repeat every fourth frame once, so the unevenness could be more noticeable since those frames are now twice as long instead of 50% longer. But I haven't tried one to tell you just how noticeable it would be. Native 60fps content (sports is the most common one I know of) would require throwing away half the frames, but at least the remaining frames would all be shown for the same amount of time. NTSC "video" content (stuff shot at 30 interlaced frames per second, so 60 fields per second, stuff like old TV shows) would work alright. You wouldn't be able to take advantage of deinterlacing methods that try to reconstruct 60 full frames, but it wouldn't be a huge loss.
- joosters 13y ago
  Drop the screen refresh rate to 24hz to match the video then...
  - majormajor 13y ago
    That was an option on CRTs since you could control the speed of the beam, but I'm not aware of any LCDs that have an adjustable native refresh rate. Ah, the good ol' days, when blacks were black and "non-native resolutions" weren't a thing. :)
    - derekp7 13y ago
      That's one thing I never understood. With CRT's, you still had discrete pixels (either clusters of 3 RGB dots, or 3 RGB bars). But no matter the resolution (within the CRT's limits), nothing looked blotchy. Whereas on an LCD, you can always tell when it isn't running in a native resolution (or even divisor). So what is the difference? Other than a CRT would naturally do analog interpolation, couldn't an LCD or graphics adapter be programmed to simulate the same effect?
      - majormajor 13y ago
        The hardware side is where things start getting out of my area of knowledge, but my understanding is that yes, CRTs had a native upper bound on their resolution which had to do with the dot pitch of the display. I would guess that the way this worked in practice is that for lower resolutions they were effectively partially lighting up multiple phosphors per pixel on each pass? I imagine the difference between this an an LCD is that the amount of time a single pixel was lit by the beam on a CRT is far, far lower than the refresh rate of an LCD. I believe there were monitors that went both ways with this, not just letting you run lower than native but also higher than native (I had a 17" CRT that claimed to support 1600x1200, but not without a lot of blurriness, for instance). edit: man, it's really hard to find specific info about how exactly that worked online these days
      - bhauer 13y ago
        I am not a hardware guy, but from my understanding a CRT uses an electron gun a distance from the screen to cause a matrix of phosphors to emit light. Hence the depth of CRTs. The electron gun can repeatedly scans from top to bottom, left to right (or whatever direction). It can be adjusted to make any number of vertical "lines" as it sweeps over the phosphor matrix, but the matrix has a fixed dot-pitch. Because an electron beam is being shot at the matrix from a distance, there's a certain degree of bleed-over to neighboring phosphors. Firing electrons at the phosphors at coordinate 100,200 means a little bit of energy is also received by 99,199 and other neighboring "virtual pixels." In other words, while the phosphor matrix was just as rigid as the pixel matrix of an LCD, the power source that caused light to be emitted was not perfectly matched 1:1 with the matrix. There isn't an electron gun for each pixel, there is a single electron gun for the whole display. By comparison, each pixel in an LCD monitor (putting aside the three internal pixels for colors) is an individual mask over a uniformly white back-light. The panel adjusts the opacity of the mask to allow more or less of the white backlight through. The mask over a backlight is what gives LCDs a challenge with viewing angle (because there's a non-zero gap between the light source and the mask and the mask itself has a non-zero depth). Finally, as I understand it, a plasma television is a bit like both. It's a matrix of phosphors like a CRT, but with individual power sources for each pixel like an LCD. Because light is emitted at the surface, plasmas tend to have very good viewing angles. OLED then is, in my opinion at least, a spiritual successor to plasma. OLED emits light at its surface like an LCD or plasma, but uses no mask layer over a backlight, like a CRT or plasma. (Take all of the above with a grain of salt because I'm just a layperson when it comes to hardware.)
- crassus 13y ago
  Is this also the source of the ultra-real "soap opera" effect that ruins movies on new televisions?
  - underwater 13y ago
    No. That's the TV interpolating frames. The effect described would introduce a stutter.
    - bcheung 13y ago
      Ugh, I hate that effect. The frame rate dynamically changes based on when the algorithm can find frames to interpolate. It's so disorienting.
  - usefulcat 13y ago
    That effect can often be turned off. I know it can on my 2008-era samsung tv.
- Kiro 13y ago
  So how does it work with games where the FPS is variable? Shouldn't the mismatch in FPS and HZ create a lot of problems?
bhauer 13y ago
Admittedly, we've not yet watched a film on the Seiki. We have fired up a few Youtube videos at full-screen and aside from looking comical (due to the large lossy compression artifacts present even in Youtube's highest quality videos), the video seems as smooth as Youtube ever does, at least to my eye. I suspect for video from a higher-quality source, my opinion would not be so forgiving. But put briefly, for consuming crummy web content, I have no complaints. Don't get me wrong. I haven't yet bought one for myself principally because I am waiting for Seiki (or someone) to offer me a similarly priced 4K monitor with HDMI 2 (60Hz). If I got the current model today and a v2 of the Seiki arrived next month with HDMI 2, I would be sad. The 30Hz limitation is the most significant disadvantage of this display. So I wait. But if you don't want to wait, the 30Hz limitation is tolerable. At <$500 now, it's very hard to restrain myself from just buying another and being done with the temptation. More of my crazy thoughts: http://tiamat.tsotech.com/seiki-4k http://tiamat.tsotech.com/seiki-4k
- Locke1689 13y ago
  I don't understand why you would want HDMI 2. DisplayPort is pretty much better in every single way.
  - bhauer 13y ago
    DisplayPort is indeed superior and I'd gladly take a cheap UHD monitor with both HDMI 2 and DisplayPort.
- kllrnohj 13y ago
  YouTube is 30fps, so it should match up with vsync nicely.
jfb 13y ago
I wouldn't want to watch a movie at 30hz, or at 4k.
- potatolicious 13y ago
  Aren't all movies at 30Hz or 24Hz? I don't think there are any mainstream releases at a higher FPS...
  - tcpekin 13y ago
    The Hobbit [1] is the only movie I can think of that is shot at a higher frame rate, it being shot at 48 fps. [1] http://www.fastcocreate.com/1682085/will-the-hobbit-start-a-48-fps-movie-revolution http://www.fastcocreate.com/1682085/will-the-hobbit-start-a-...
    - potatolicious 13y ago
      Very cool. I don't think it'll be distributable at 48fps though, at least not via blu-ray, which seems capped at 30fps by spec.
      - selmnoo 13y ago
        Current direction of cinema trends raises lots of interesting technological questions. I'm especially intrigued by those concerning media discs. The next FIFA World Cup is set to broadcast in 8K, so I think one can reasonably predict there'll be movies recorded in 8K by 2018. Now, imagine, something the length of the LoTR trilogy, in 60fps, 3D, and 8K. You're looking at a lot of data right there... that's asking for discs to contain half a petabyte worth of data by around that time. It'll be interesting to see how that plays out.
    - jfb 13y ago
      There are some noises about higher framerate movie production; The Hobbit was the first, and it was not well received. Motion at 30fps and higher looks very different than motion at 24fps, and part of the resistance to HFR is simply unfamiliarity; however, a lot of shots in The Hobbit were not improved by the 48fps presentation; I think that we won't get good results until film projection is dead, and directors and DPs can present at arbitrary frame rates -- there's no reason that, given a digital projection, a movie couldn't use 48 (or 60!) for certain sequences, while retaining 24 (or 30) for others. It would add another technique for directors to use, just as video has.

by selmnoo 13y ago

majormajor 13y ago

Displaying 24fps content on a 60fps screen repeats frames. If you repeat half the frames of the 25fps stream twice, and alternating half of the frames three times, then you get 60 frames per second, with a stream looking like: AABBBCCDDDEEFFF... This isn't perfect, since really each frame should be on the screen for exactly the same amount of time, which is why 120hz TVs came about -- 120 is evenly divisible by each of 60, 30, and 24. But it's also pretty much good enough, or else there'd be a lot more complaining about watching movies on computers or other 60hz devices. A 30hz screen makes that harder, since you're going to have to repeat every fourth frame once, so the unevenness could be more noticeable since those frames are now twice as long instead of 50% longer. But I haven't tried one to tell you just how noticeable it would be. Native 60fps content (sports is the most common one I know of) would require throwing away half the frames, but at least the remaining frames would all be shown for the same amount of time. NTSC "video" content (stuff shot at 30 interlaced frames per second, so 60 fields per second, stuff like old TV shows) would work alright. You wouldn't be able to take advantage of deinterlacing methods that try to reconstruct 60 full frames, but it wouldn't be a huge loss.

joosters 13y ago
Drop the screen refresh rate to 24hz to match the video then...
- majormajor 13y ago
  That was an option on CRTs since you could control the speed of the beam, but I'm not aware of any LCDs that have an adjustable native refresh rate. Ah, the good ol' days, when blacks were black and "non-native resolutions" weren't a thing. :)
  - derekp7 13y ago
    That's one thing I never understood. With CRT's, you still had discrete pixels (either clusters of 3 RGB dots, or 3 RGB bars). But no matter the resolution (within the CRT's limits), nothing looked blotchy. Whereas on an LCD, you can always tell when it isn't running in a native resolution (or even divisor). So what is the difference? Other than a CRT would naturally do analog interpolation, couldn't an LCD or graphics adapter be programmed to simulate the same effect?
    - majormajor 13y ago
      The hardware side is where things start getting out of my area of knowledge, but my understanding is that yes, CRTs had a native upper bound on their resolution which had to do with the dot pitch of the display. I would guess that the way this worked in practice is that for lower resolutions they were effectively partially lighting up multiple phosphors per pixel on each pass? I imagine the difference between this an an LCD is that the amount of time a single pixel was lit by the beam on a CRT is far, far lower than the refresh rate of an LCD. I believe there were monitors that went both ways with this, not just letting you run lower than native but also higher than native (I had a 17" CRT that claimed to support 1600x1200, but not without a lot of blurriness, for instance). edit: man, it's really hard to find specific info about how exactly that worked online these days
    - bhauer 13y ago
      I am not a hardware guy, but from my understanding a CRT uses an electron gun a distance from the screen to cause a matrix of phosphors to emit light. Hence the depth of CRTs. The electron gun can repeatedly scans from top to bottom, left to right (or whatever direction). It can be adjusted to make any number of vertical "lines" as it sweeps over the phosphor matrix, but the matrix has a fixed dot-pitch. Because an electron beam is being shot at the matrix from a distance, there's a certain degree of bleed-over to neighboring phosphors. Firing electrons at the phosphors at coordinate 100,200 means a little bit of energy is also received by 99,199 and other neighboring "virtual pixels." In other words, while the phosphor matrix was just as rigid as the pixel matrix of an LCD, the power source that caused light to be emitted was not perfectly matched 1:1 with the matrix. There isn't an electron gun for each pixel, there is a single electron gun for the whole display. By comparison, each pixel in an LCD monitor (putting aside the three internal pixels for colors) is an individual mask over a uniformly white back-light. The panel adjusts the opacity of the mask to allow more or less of the white backlight through. The mask over a backlight is what gives LCDs a challenge with viewing angle (because there's a non-zero gap between the light source and the mask and the mask itself has a non-zero depth). Finally, as I understand it, a plasma television is a bit like both. It's a matrix of phosphors like a CRT, but with individual power sources for each pixel like an LCD. Because light is emitted at the surface, plasmas tend to have very good viewing angles. OLED then is, in my opinion at least, a spiritual successor to plasma. OLED emits light at its surface like an LCD or plasma, but uses no mask layer over a backlight, like a CRT or plasma. (Take all of the above with a grain of salt because I'm just a layperson when it comes to hardware.)
crassus 13y ago
Is this also the source of the ultra-real "soap opera" effect that ruins movies on new televisions?
- underwater 13y ago
  No. That's the TV interpolating frames. The effect described would introduce a stutter.
  - bcheung 13y ago
    Ugh, I hate that effect. The frame rate dynamically changes based on when the algorithm can find frames to interpolate. It's so disorienting.
- usefulcat 13y ago
  That effect can often be turned off. I know it can on my 2008-era samsung tv.
Kiro 13y ago
So how does it work with games where the FPS is variable? Shouldn't the mismatch in FPS and HZ create a lot of problems?

bhauer 13y ago

Admittedly, we've not yet watched a film on the Seiki. We have fired up a few Youtube videos at full-screen and aside from looking comical (due to the large lossy compression artifacts present even in Youtube's highest quality videos), the video seems as smooth as Youtube ever does, at least to my eye. I suspect for video from a higher-quality source, my opinion would not be so forgiving. But put briefly, for consuming crummy web content, I have no complaints. Don't get me wrong. I haven't yet bought one for myself principally because I am waiting for Seiki (or someone) to offer me a similarly priced 4K monitor with HDMI 2 (60Hz). If I got the current model today and a v2 of the Seiki arrived next month with HDMI 2, I would be sad. The 30Hz limitation is the most significant disadvantage of this display. So I wait. But if you don't want to wait, the 30Hz limitation is tolerable. At <$500 now, it's very hard to restrain myself from just buying another and being done with the temptation. More of my crazy thoughts: http://tiamat.tsotech.com/seiki-4k http://tiamat.tsotech.com/seiki-4k

Locke1689 13y ago
I don't understand why you would want HDMI 2. DisplayPort is pretty much better in every single way.
- bhauer 13y ago
  DisplayPort is indeed superior and I'd gladly take a cheap UHD monitor with both HDMI 2 and DisplayPort.
kllrnohj 13y ago
YouTube is 30fps, so it should match up with vsync nicely.

jfb 13y ago

I wouldn't want to watch a movie at 30hz, or at 4k.

potatolicious 13y ago
Aren't all movies at 30Hz or 24Hz? I don't think there are any mainstream releases at a higher FPS...
- tcpekin 13y ago
  The Hobbit [1] is the only movie I can think of that is shot at a higher frame rate, it being shot at 48 fps. [1] http://www.fastcocreate.com/1682085/will-the-hobbit-start-a-48-fps-movie-revolution http://www.fastcocreate.com/1682085/will-the-hobbit-start-a-...
  - potatolicious 13y ago
    Very cool. I don't think it'll be distributable at 48fps though, at least not via blu-ray, which seems capped at 30fps by spec.
    - selmnoo 13y ago
      Current direction of cinema trends raises lots of interesting technological questions. I'm especially intrigued by those concerning media discs. The next FIFA World Cup is set to broadcast in 8K, so I think one can reasonably predict there'll be movies recorded in 8K by 2018. Now, imagine, something the length of the LoTR trilogy, in 60fps, 3D, and 8K. You're looking at a lot of data right there... that's asking for discs to contain half a petabyte worth of data by around that time. It'll be interesting to see how that plays out.
  - jfb 13y ago
    There are some noises about higher framerate movie production; The Hobbit was the first, and it was not well received. Motion at 30fps and higher looks very different than motion at 24fps, and part of the resistance to HFR is simply unfamiliarity; however, a lot of shots in The Hobbit were not improved by the 48fps presentation; I think that we won't get good results until film projection is dead, and directors and DPs can present at arbitrary frame rates -- there's no reason that, given a digital projection, a movie couldn't use 48 (or 60!) for certain sequences, while retaining 24 (or 30) for others. It would add another technique for directors to use, just as video has.