Low fps on x264 with E6600

[xionoland]

hi, i have always having trouble with the x264 format and tried different ffdshow, coreavc, matroska, powerdvd codecs ,cccp on all differerent players and vlc and it never smooth, only smooth once time in linux tho, but not for long. so i buyed i new machine
(e6600, 8800gts and vista32). And tried again, but it is still not smooth, it lagg a little, like its just 15fps or something, very vissible on movment in white background. So it is many friends who dont see the problem. But i see it the hole movie. Can someone pleas help me before i toss the new machine out of the window:devil:

[foxyshadis]

This is a playback problem, not an encoding one, so moved to players forum.

What resolution is the video? What fps? What bitrate? Have you used custom playback options in mpc or zoomplayer to be sure that you're actually using the codecs you think you are? Does it even happen when you enable hardware acceleration, with cyberlink's (powerdvd) decoder?

[addit]

Just to clarify is this a problem that you specifically see and many other people don't? If not, perhaps it is inherent in the source - what source is it? In any case, try watching the video with the lights off or in lower lighting conditions (visual acuity to changes in framerate is said to be lower).

Adam

*EDIT* Darn, just a little bit to slow today heh, foxyshadis beat me too it!

[xionoland]

an example is "bobby 2006 dvd5 x264 720P sample". It was smooth once in linux. the bitrate is 48000. ive use MPC classic now, and tryed nvidia purevideo yesterday, but isnt sure if purevideo was on. i havent change in option because i dont know that player to vell, any tips that i should turn off or on?

[xionoland]

Source: Bobby 720p CtrlHD.ts
Codec: x264 2-pass
Audio: 384K AC3 5.1
Bitrate: 4968kbps
Resolution: 1280 x 720 (1.78:1)
Subtitles: None muxed
Framerate: 23.976 fps

how do i enable hardware acceleration? And in vlc its even worse. tanks for any tip, i would love to see x264 smooth, for the quality looks super for that filesize...

[Delerue]

Really strange. 720p with 24 FPS isn't heavy.

IMO VLC is a good player, but it's not that good for playing H.264 videos. So try to avoid it. I suggest MPC + CoreAVC to maximize the performance. But, considering that you have a E6600, FFDShow can do the job pretty well. You can also try SMPlayer (http://smplayer.sourceforge.net/windows/download_en.php); it uses the pretty good MPlayer engine. It has been said that PowerDVD has a powerful hardware acceleration option that works great (only) with Vista. Try it. ;)

You should check the memory utilization too. Disable all the useless process just to test (i.e. AV, firewall, and all the tray stuff).

Good luck.

[burfadel]

hi, i have always having trouble with the x264 format and tried different ffdshow, coreavc, matroska, powerdvd codecs ,cccp on all differerent players and vlc and it never smooth, only smooth once time in linux tho, but not for long. so i buyed i new machine
(e6600, 8800gts and vista32). And tried again, but it is still not smooth, it lagg a little, like its just 15fps or something, very vissible on movment in white background. So it is many friends who dont see the problem. But i see it the hole movie. Can someone pleas help me before i toss the new machine out of the window:devil:

What driver are you using for Vista? Some of the earlier 8800 drivers weren't that good...

You should try the Forceware 160.04 for Vista, you should be able to find it. The 160.04 works flawlessly on a friends computer, I'm using 160.03 for Win XP x64 (it is a different driver), and works flawlessly too.

Worth a try. I think it may have something to do with the display output method. There's Overlay mixer, VMR 7 & 9 (both windows and renderless), EVR (new one in Vista and on XP with Net framework 3) to name a few. These all rely on the display driver. Since you have the problem across different players and codecs, I'd say thats your best bet!

[bill_baroud]

I just go a new laptop (a samsung Q45) with Vista and noticed the same thing : If you use VMR9 for playback (in MPC), it's stuck at 15fps even with a plain 640x480 x264 + subtitle mkv file.
That's too bad, now i've to install VSfilter (playback is flawless with overlay mixer) instead of using internal mpc renderer.

Some crippled output feature from Vista kicking in ?
(well i still have to see if there is recent build of mpc/ffdshow since i'm bit lagging behind video news these days...)

[burfadel]

I still think its a driver issue, the Vista drivers are still essentially in development!

[xionoland]

thanks for the tip, ive give it a try, but i have some problems in xp too, but after installed the older forceware 160.04 it get a little better, it is ok too watch a movie, but in action sceens its a little laggy. i think it is driver too, because it doesnt play smooth in vlc, witch dont use external codecs.

[burfadel]

23.976 is actually on the edge of the human visual system. During movement, you'll see frame by frame the motion is in steps, but your brain smooths it out for you. Some people are more susceptible to this than others. If you have any Pal media have a look at that, as its essentially just slightly sped up from 23.976. If you don't see any jerkiness then thats you're answer! There's not too much you can do to fix it, unless you can apply some post motion compensation (like some televisions do). I don't know of any off hand sorry.

Note: Imagine playing a game. If the frame rate drops to 23.976 fps, because its all motion you do notice it, and yet thats the same framerate as movies :)

[Delerue]

23.976 is actually on the edge of the human visual system.

Do you have any scientific study that confirm this?

[Manao]

23.976 is actually on the edge of the human visual systemNo. If it was, interlacing wouldn't exist. And, alas, interlacing does exist.

What is on the edge of the HVS is 24 fps with motion blur. Still, 48 fps with a lot less motion blur will be a bit smoother, and sharper ( during motion ).

Try to deinterlaced a sport video ( ie 25i -> 25p, don't bob it ). You'll see it's clearly not smooth, but jerky.

[Delerue]

No. If it was, interlacing wouldn't exist. And, alas, interlacing does exist.

Interlacing began because it was the best option to keep a high resolution without the need of a huge bandwidth. The other reason is because the AC has 60 Hz (America/Japan) or 50 Hz (Europe), and the engineers of this time had problems to implement a system out of the sync with the AC; that's why we have 60 Hz for NTSC, and 50 Hz for PAL.

What is on the edge of the HVS is 24 fps with motion blur. Still, 48 fps with a lot less motion blur will be a bit smoother, and sharper ( during motion ).

Hmmm... I ask you the same thing I asked to burfadel. :)

[Trahald]

Flicker fusion threshold.. google that a bit. the study i linked below finds it at about 22hz

http://www.asu.edu/news/faculty_students/holloway_050305.htm

[Manao]

Interlacing began because it was the best option to keep a high resolution without the need of a huge bandwidthIncorrect. 25p and 50i uses the same bandwidth and the same resolution. Still, 50i has been chosen. And because the AC was at 50Hz, because 25Hz is doable easily from 50Hz.
I ask you the same thing I asked to burfadelNo need for a paper. Take an interlaced sport video. Deinterlace it ( with sangnom, or with something like separatefields.selectodd, don't use a blend deinterlacer, since it naturally adds motion blurs ), and watch it. It will be sharp, but somehow jerky. Blend deinterlace it, it shouldn't be that much jerky anymore, but it'll be less sharp.

[Delerue]

Incorrect. 25p and 50i uses the same bandwidth and the same resolution. Still, 50i has been chosen.

Hmmmm... Maybe you're right. I always read that common NTSC TVs has a resolution of 640x480, but it's interlaced. So if you want to transmit a P signal (keeping the same bandwidth), you have to reduce the resolution. Here (http://en.wikipedia.org/wiki/Interlacing) there's some information (including an article in the end that says what you said about P vs I).

And because the AC was at 50Hz, because 25Hz is doable easily from 50Hz.

Yeah. I said that. ;)

Another reason for the implementation of this tecnology is that at 30 Hz (or 25 Hz in Europe) the image generates A LOT of flicker, so double the frame rate appeared to be the best option. But doing that we'll have to double the bandwidth or make necessary a frame buffer. Both things were not viable at that time.

No need for a paper. Take an interlaced sport video. Deinterlace it ( with sangnom, or with something like separatefields.selectodd, don't use a blend deinterlacer, since it naturally adds motion blurs ), and watch it. It will be sharp, but somehow jerky. Blend deinterlace it, it shouldn't be that much jerky anymore, but it'll be less sharp.

I asked for a study because you said this: 'What is on the edge of the HVS is 24 fps with motion blur'. Can you provide a study that confirm this afirmation?

[Delerue]

Flicker fusion threshold.. google that a bit. the study i linked below finds it at about 22hz

http://www.asu.edu/news/faculty_students/holloway_050305.htm

Consequently, rates over about 22 hertz are perceived as a steady or unflickering light. (from your link)

With all respect, but this 'study' is a bullshit. First because this afirmation is insane and can be easily refuted with a display monitor or inside a cinema room. Second, because subliminal messages are pure myth (http://www.csicop.org/si/9204/subliminal-persuasion.html) and never were proved to be true (more (http://www.snopes.com/business/hidden/popcorn.asp)). Third, the article doesn't explain anything about the study itself.

[Manao]

And because the AC was at 50Hz, because 25Hz is doable easily from 50Hz.Ok, that didn't make any sense. I wanted to say : "And because the AC was at 50Hz, [i]since 25Hz is easily doable from 50Hz".

'What is on the edge of the HVS is 24 fps with motion blur'. Can you provide a study that confirm this afirmation?No. However :
I don't see any jerky motion when watching movies ( progressive, 24/25fps ), either at the cinema or on my TV ( shown at "50Hz", that is a field - not a frame - every 20ms ) or my computer ( 25fps with a LCD monitor refresh rate of 60Hz ).
I believe in the guys who chose 24fps in the first place. They wouldn't have uselessly overengineered the stuff, seeing how costly the fps was at that time. Imho, they chose the lowest watchable framerate, so by definition the one "on the edge".
As for the "motion blur" part, well, without motion blur, at 25 fps, I do see jerkyness, and I'm really sure my eyes aren't better than the average guy's.

[Delerue]

Ok, that didn't make any sense. I wanted to say : "And because the AC was at 50Hz, [i]since 25Hz is easily doable from 50Hz".

Hmmm... So you're saying that none of the reasons I told is correct? Well, what's the reason interlaced was invented?

No. However :

Ok. I asked because I'm looking for that for so much time, and I never found anything. :(


I don't see any jerky motion when watching movies ( progressive, 24/25fps ), either at the cinema or on my TV ( shown at "50Hz", that is a field - not a frame - every 20ms ) or my computer ( 25fps with a LCD monitor refresh rate of 60Hz ).
I believe in the guys who chose 24fps in the first place. They wouldn't have uselessly overengineered the stuff, seeing how costly the fps was at that time. Imho, they chose the lowest watchable framerate, so by definition the one "on the edge".
As for the "motion blur" part, well, without motion blur, at 25 fps, I do see jerkyness, and I'm really sure my eyes aren't better than the average guy's.

And if I told you that I can prove that 24 FPS (even with motion blur) isn't enough? Or better: I can notice the diference between 24 and 48 FPS easily.

[Trahald]

Thats the thing.. you have 2 points.. the point where a sequence of pictures ceases to look like a individual pictures, and the point where it looks like perfect motion. when i look at film it does look fluid. but a 60+ fps game seems to look more fluid. blurred 24fps is (near) the point where most people stop seeing a slide show. is it perfect, no. just because your brain has difficulty keeping up, to an extent it can detect the less than perfect motion.

i think most people can tell the difference between 24-48. but if you put a 24fps movie in front of the average person... then a 48 fps movie in front of them.. then a 10 fps movie in front of them.. (without saying anything) they will only comment about the 10 fps movie.. thats the difference.

[Delerue]

when i look at film it does look fluid. but a 60+ fps game seems to look more fluid.

To me isn't fluid anyway. :(

But I think that the point here is the difference between common people and rigorous people. My mom can't tell the difference between 10 FPS and 48 FPS (yeah... really odd, but true). But some friends of mine and me, that play video-game and computer games since ever (hehehe), can't tell the difference between any framerate below 60 FPS; one of these friends can't notice more than 60 FPS!

What I don't understand is the idea (myth) that the human eye has a specific limit, when we, in fact, never found a scientific article about it. Maybe we must try to avoid the idea that our limits are the human limits. What you think?

[Manao]

Well, what's the reason interlaced was invented?People want two things when it comes to video :
spatial resolution
temporal resolution ( framerate )
However, at the time interlacing was invented, the bandwidth available wasn't enough to have what was considered enough in spatial resolution ( 720x576 / 720x480 ) and temporal resolution ( 50 fps / 60 fps ). Interlacing allows a tradeoff : when there's no motion, ie when the eyes are most interested by the spatial detail, interlacing offers the full resolution. When there's motion, the eyes tends to favor motion, and that's what interlacing will favor too.

Basically, interlacing is a way of "compressing" the video. It was the "best" way available at that time. And even now, when all new displays are progressive, interlacing is kept, because even though it's a mathematical nightmare, and makes every encoding standard more complicated ( PAff, MBaff in AVC ), it still requires less bandwidth for almost the same visual quality than progressive at full framerate.


Or better: I can notice the diference between 24 and 48 FPS easilyDid I ever say the contrary ? The 48 fps one will be sharper ( less motion blur ), and seems more real/lively.

In the same way, on your TV, you can always tell the difference between a movie and sports/news/shows/90s series ( they were still shot interlaced at that time ).

However, I think 24 fps is the limit under which you can stay all day watching TV without a headache. And it wouldn't do if people were having headaches after only 1 hour of TV a day, would it ?

[Delerue]

People want two things when it comes to video :[list]
spatial resolution
temporal resolution ( framerate )

Ok. Spatial resolution is good, but this article here (http://www.vxm.com/Progvsinter.html) says that the 'interlacing doubles the resolution' is a myth. What you think?

As for FPS, I can't understand why they choose 60 FPS (Fields Per Second) if cinema has 24 FPS (Frames Per Second) and is considered good to common people.

It was the "best" way available at that time. And even now, when all new displays are progressive, interlacing is kept, because even though it's a mathematical nightmare, and makes every encoding standard more complicated ( PAff, MBaff in AVC ), it still requires less bandwidth for almost the same visual quality than progressive at full framerate.

Not anymore. With digital signal, is better to compress progressive material. So 720p has a better result than 1080i, but both use the same bandwidth. The article I linked before told about this.

However, I think 24 fps is the limit under which you can stay all day watching TV without a headache. And it wouldn't do if people were having headaches after only 1 hour of TV a day, would it ?

I think you're confunding two different things: FPS and refresh rate. Low FPS doesn't make you feel headaches, but low refresh rate do.

[Manao]

Low FPS doesn't make you feel headaches, but low refresh rate do.If I take a 25fps video and remove one frame out every two, I obtain something I can't stand to watch for more than 5 minutes. I think it'll be the same for you. It's definitely not confortable to watch such low framerate videos.

As for refresh rates, indeed, anything under 80 Hz for CRT displays is likely to give me headaches, but that's another story.

Now, for the article : However, because of the much higher statistical correlation and lower level of aliasing present in the P signal, both can be MPEG-compressed to the same digital data rate with about the same subjective quality. Results like these have been reported by Bell Labs, NHK, RAI, and Project Race of the EU. Thus, there is no data-rate penalty for using P rather than I, and there are many advantages.From my testing, I would say that 60p requires at least 15% more bandwidth than 30i, at the same quality.

As for the 1080i / 720p spatial resolution, it's a complicated matter.

As I said, the spatial resolution matters mostly when the picture is relatively still, and in such a case, the authors agrees to say that 1080i wins.

When there's motion, indeed, the vertical resolution is reduced to something in between 540 & 1080. But the horizontal resolution is still 1920, so the spatial resolution is between 1 to 2 millions of pixel, so at least as much as 720p.

Finally, 1080i30 handles motion as well as 720p60.

A detail left out of the study, but still very important : films are still shot at 24 fps, so for film content, 720p60 is really inferior to 1080i30 ( which actually becomes 1080p24 in that case ).

So, in the end :
1080i30 requires roughly 25 to 40% more bandwidth than 720p60 at
the same quantizer
1080i30 has a high spatial resolution
1080i30 is interlaced - which is bad since most HD displays are progressive and so must deinterlace it, which isn't easy
1080i30 is clearly better for film content, provided you can pay the additionnal bandwidth over 720p60

I agree with the conclusion of the author : interlacing should disappear. I would really prefer 1080p60 over 1080i30. But 1080p60 was requiring too much bandwidth ( ~60/70 % more than 720p )

[burfadel]

I take it you didn't mean a smooth then jerky motion, but a consistent slight lie stop start motion?...(very slight)

[MarcioAB]

Why interlacing began ?

From one side, the cannon machinery inside the TV that throw electrons into the phosphor wall was limited at that time to 15750 scan-lines per second.

On the other side, because the 60Hz AC energy pulse, everything else inside TV became in sync with that and so 60 frames per second became necessary.

So 15750/60 = 262.5 scan-lines per frame, pulsing at 60 fps was everything that could be achievable at that time.

But phosphor takes longer to diminish the emitted light. So it became necessary to adjust the cannon to a bigger distance in the vertical and fire each frame with 1 scan-line of difference.

That will produce 525 lines interlaced being 262.5 refreshed at each 60Hz pulse and the other 262.5 lines still be there because phosphor did not diminished yet.

So, despite in reality we have 262.5 lines refreshed at 60Hz, we can say we have 525 lines refreshed at 30Hz.

And everything adjusted very ingenious providing a luxury full 525 lines at 60 fps in more static scenes and the blurry interlaced image at 30 fps in movement scenes. Similar math for 50Hz AC energy pulse.

We are used to watch this at short distance.

In the theater, watch film in a long distance at 24 fps (exhibited at 2x24 fps) is a totally different experience.

Now, watch film (24 fps) in the PC in a short distance ... I notice sometimes, very clearly, that is missing frames, but I accept it in the name of the art.