View Full Version : CoreCodec/H.264 Codec "CoreAVC"
Dark Eiri
1st March 2009, 19:38
@ Dark Eiri
A Turion 2GHz is roughly equivalent to a 3.4GHz P4! so there is nothing wrong with a 3GHz P4 having problems ;)
That's good to know, I always saw it as a 'big' Celeron. :p
Thanks for the info!
carlo_0000
1st March 2009, 23:52
@ Dark Eiri
A Turion 2GHz is roughly equivalent to a 3.4GHz P4! so there is nothing wrong with a 3GHz P4 having problems ;)
@ Shinigami
DXVA doesn't always work, does it? ;)
@ carlo
An Athlon 64 isn't 2800+ anymore if overclocked to 2.5GHz, it is even faster than an Athlon 64 4000+!
Could you tell us your average load with the Seven Pounds trailer please? with a graph if possible. I can't imagine your CPU not maxing out.
dou you have a download link ?
beause i search google , found only video for quicktime, i don't like quicktime and probably do not use coreavc, and i was unable to download the .mov (only 96ko)
DJ Bobo
2nd March 2009, 00:15
You can find the link in the previous page, where I posted the graph. Just click on "Seven Pounds"
carlo_0000
2nd March 2009, 01:49
@ original speed 1.8G
not fast enough and some lag
pict: http://ftp1.dommel.be/picture_library/1800mhz.JPG
oc 2.5ghz
didn't hit 100% when playing
pict: http://ftp1.dommel.be/picture_library/2500mhz.JPG
and on my athlon x2 4200+ (olso oc @ 2.5ghz) cpu usage is 20-24%
BetaBoy
2nd March 2009, 03:37
DJ Bobo, yes we updated the requirements page.... but it is hard to balance the 'recommended' requirements to real life usage. For example the requirements page is based on non paff/mbaff content, ie; typical content for Apple trailers. We can refine it more for sure.... any recommendations? I am willing to have the community, ie; d9 rewrite what we should put there.
ChronoCross
2nd March 2009, 07:34
DJ Bobo, yes we updated the requirements page.... but it is hard to balance the 'recommend' requirements to real life usage. For example the requirements page is based on non paff/mbaff content, ie; typical content for Apple trailers. We can refine it more for sure.... any recommendations? I am willing to have the community, ie; d9 rewrite what we should put there.
If you let the d9 community write it, it will be a 12 page document chronicaling every feature available in the h264 spec and what's viewable in different resolutions on 30 different processors. Might be a bit of overkill for a more non-power user audience.
DJ Bobo
2nd March 2009, 10:12
@ carlo
Something's wrong here. You sure you're not using DXVA? 'cause the Seven Pounds Trailer would play in DXVA mode if you don't disable the correpondant option in MPC-HC.
There can't be so much difference between a 1.9GHz X2 (70-75%) and a 2.5GHz X2 (20-25%)
@ BetaBoy
I guess you can do it like most game editors would: minimum and recommended requirements. The minimum would correspond to what you have on the website right now, the recommended to what is common among the scene, may be something close to the DivX Plus HD profile at common bitrates (5 to 6 Mbit/s for 720p content (eq. to a DVD-5 rip) and 9 to 10 Mbit/s for 1080p content (eq. to a DVD-9 rip)).
Sure, you won't make everybody happy, there are always the 5% that won't agree ('cause they're using unusual settings), but at least the requirements would be much more realistic.
If you can't decide on what is "common", I guess you could always post a poll with several encoding options and then go with the majority and base the recommended requirements on that.
Cheers
_DW_
2nd March 2009, 15:32
DJ Bobo, yes we updated the requirements page.... but it is hard to balance the 'recommend' requirements to real life usage. For example the requirements page is based on non paff/mbaff content, ie; typical content for Apple trailers. We can refine it more for sure.... any recommendations? I am willing to have the community, ie; d9 rewrite what we should put there.
The requirements look fine to me but you might want to add under CPU something to the effect of "or equivalent AMD processor." Someone who is cursing through your site might just assume you don't support AMD processors and move on.
May sound silly but it can happen and adding a few words like that will prevent that.
lucassp
2nd March 2009, 15:45
you could also add 8800 GTS 512 to the supported CUDA graphic cards.
BetaBoy
2nd March 2009, 16:30
_DW_, lucassp.... suggestions added to the page.... DJ Bobo... thanx for the input, we will consider it.
CC.... yeah, but that's why I enjoy D9 as much as I do.
lucassp
2nd March 2009, 16:49
_DW_, lucassp.... suggestions added to the page.... DJ Bobo... thanx for the input, we will consider it.
I did make the 512 bold for a reason :) The 320/640 have the old G80 core which does not have VP2, while the 512 version is based on G92 core and does have the VP2 decoder.
BetaBoy
2nd March 2009, 17:23
changed
DJ Bobo
2nd March 2009, 17:30
The requirements look fine to me but you might want to add under CPU something to the effect of "or equivalent AMD processor." Someone who is cursing through your site might just assume you don't support AMD processors and move on.
May sound silly but it can happen and adding a few words like that will prevent that.
What was that? Are you even following the discussion? This is not about "or equivalent", this is about requirements being unrealistic. It won't change anything if they would put "or equivalent AMD processor" because an Athlon 64 2800+ is equivalent to a P4@2.8GHz and both are absolutely not sufficient to play 1080p!
Unbelievable! :mad:
_DW_
2nd March 2009, 20:21
What was that? Are you even following the discussion? This is not about "or equivalent",
Apparently not. I made a suggestion and it looks like it was taken. There was an open call to make suggestions, I made one. If you don't agree please feel free to make another suggestion. Personal attacks do nothing but lower the value of your character and contribute nothing to the discussion at hand.
I also don't think a AMD 2800+ is "equivalent" to P4@2800 but I think I'm bias....:p
carlo_0000
2nd March 2009, 20:45
@ carlo
Something's wrong here. You sure you're not using DXVA? 'cause the Seven Pounds Trailer would play in DXVA mode if you don't disable the correpondant option in MPC-HC.
There can't be so much difference between a 1.9GHz X2 (70-75%) and a 2.5GHz X2 (20-25%)
@ BetaBoy
I guess you can do it like most game editors would: minimum and recommended requirements. The minimum would correspond to what you have on the website right now, the recommended to what is common among the scene, may be something close to the DivX Plus HD profile at common bitrates (5 to 6 Mbit/s for 720p content (eq. to a DVD-5 rip) and 9 to 10 Mbit/s for 1080p content (eq. to a DVD-9 rip)).
Sure, you won't make everybody happy, there are always the 5% that won't agree ('cause they're using unusual settings), but at least the requirements would be much more realistic.
If you can't decide on what is "common", I guess you could always post a poll with several encoding options and then go with the majority and base the recommended requirements on that.
Cheers
no and no it s full software, my video card is fx5600 it does not support dxva
and olso it s coreavc codec 1.8 (no dxva)
and you sayed before 64 2800+ oc @ 2.5g = 4000+ that s no possible, 4000+ is faster and had more cache , it s more = to a 3400+ (maybe litle slower)
and on my athlon x2 4200+ there is olso no dxva (geforce 7600gt)
i think it s something wrong with your computer
and when i disable 1 core i taskmanager, i can still play 1080p with 40% cpu usage (80% for the used core)
i olso did some test, on my media center seperon 2200+ soket A readon hd2400pro
i underclock the cpu @ 900mhz, and i was still able to play a bluray with dxva, codec of mpchc (cpu around 75%)
1080p (1080i i can't play it it s corrupt deinterlace probleme)
DJ Bobo
2nd March 2009, 22:01
@ BetaBoy
Thanks
@ DW
Whatever
@ carlo
That's impossible. No way an AVC Blu-Ray would run on a 900MHz CPU, even when backed up with a Radeon HD. You sure your blu-ray disc is not MPEG-2?
As for your A64 being faster than a 4000+, it's based on the frequency, the first 4000+ was a 2.4GHz CPU.
Anyway, it's like you're telling me that a single A64 @ 2.5GHz is equivalent to my X2 @ 1.9GHz. This can mean one of the 3 things:
1) My laptop is busted => can't be, check the 6th page (http://forum.doom9.org/showthread.php?t=135923&page=6) of the x264 benchmark: my results are better than steelista's Pentium D @ 3GHz.
2) CoreAVC is not well optimized for multi-threading? I guess I'll leave this one to BetaBoy to answer (only 33% boost with a second core?!)
3) Your player is skipping frames to keep things running. That's for you to answer.
Snowknight26
2nd March 2009, 23:22
That's impossible. No way an AVC Blu-Ray would run on a 900MHz CPU, even when backed up with a Radeon HD.
It's possible. Don't denounce it till you've tried it.
That's impossible. No way an AVC Blu-Ray would run on a 900MHz CPU, even when backed up with a Radeon HD. You sure your blu-ray disc is not MPEG-2?
I think that's quite possible since he was apparently playing back a decrypted stream. MPEG-2 vs AVC shouldn't make a difference when using full hardware decoding.
As for software decoding of the 1080p trailer, 1 core of a 2 GHz Core 2 Duo (T7200) seems to be enough with libavcodec, which is in line with carlo's numbers except that he probably meant 40-48 % overall CPU usage with the X2 4200+ instead of 20-24 % (which wouldn't be realistic).
DJ Bobo
3rd March 2009, 01:18
I'm not gonna argue with you guys about things you can learn from online reviews. You go check and you'll find that what he's claiming is impossible w/ AVC blu-ray discs. MPEG-2 is another story. A 1.8GHz Sempron is what it takes for blu-ray AVC with a Radeon HD3200 (http://www.tomshardware.com/de/AMD-780G-HDTV-Blu-Ray,testberichte-239963-3.html), probably a bit less with an HD2400pro, but definitely not 900MHz.
Back on topic: I just tested the Seven Pounds trailer on a Core 2 Duo @ 2 GHz too (roughly equivalent to Athlon X2 5000+), average was 46%, thanks for confirming that nm. I noticed that the second core was way more loaded than the first one, which may confirm my theory that multi-threading isn't well implemented in CoreAVC.
Guess we'll just have to wait for BetaBoy to comment on that.
Dark Shikari
3rd March 2009, 01:22
I'm not gonna argue with you guys about things you can learn from online reviews. You go check and you'll find that what he's claiming is impossible w/ AVC blu-ray discs. MPEG-2 is another story. A 1.8GHz Sempron is what it takes for blu-ray AVC with a Radeon HD3200 (http://www.tomshardware.com/de/AMD-780G-HDTV-Blu-Ray,testberichte-239963-3.html), probably a bit less with an HD2400pro, but definitely not 900MHz.
Back on topic: I just tested the Seven Pounds trailer on a Core 2 Duo @ 2 GHz too (roughly equivalent to Athlon X2 5000+), average was 46%, thanks for confirming that nm. I noticed that the second core was way more loaded than the first one, which may confirm my theory that multi-threading isn't well implemented in CoreAVC.
Guess we'll just have to wait for BetaBoy to comment on that.The amount that your cores are loaded has absolutely nothing to do with CoreAVC and everything to do with your operating system's scheduler.
A better measurement of CoreAVC's threading would be how much each thread is loaded.
ajp_anton
3rd March 2009, 01:35
An E5200 at 1200MHz is just a little bit too slow to decode the WallE Blu-ray with CoreAVC (from hard drive, .m2ts, decrypted).
carlo_0000
3rd March 2009, 02:35
I'm not gonna argue with you guys about things you can learn from online reviews. You go check and you'll find that what he's claiming is impossible w/ AVC blu-ray discs. MPEG-2 is another story. A 1.8GHz Sempron is what it takes for blu-ray AVC with a Radeon HD3200 (http://www.tomshardware.com/de/AMD-780G-HDTV-Blu-Ray,testberichte-239963-3.html), probably a bit less with an HD2400pro, but definitely not 900MHz.
Back on topic: I just tested the Seven Pounds trailer on a Core 2 Duo @ 2 GHz too (roughly equivalent to Athlon X2 5000+), average was 46%, thanks for confirming that nm. I noticed that the second core was way more loaded than the first one, which may confirm my theory that multi-threading isn't well implemented in CoreAVC.
Guess we'll just have to wait for BetaBoy to comment on that.
ho no ? you don't know what you speeking about
are my print screen al fake maybe ????
here playing beyonce experience m2ts file , it s VC1 video bitrate 42 mbit/s
cpu 900mhz
gpu readon 2400pro
http://ftp1.dommel.be/picture_library/vc1%20beyonce.JPG
here some h264 from Within Temptation Black Symphony 20mbit/s
http://ftp1.dommel.be/picture_library/Within%20Temptation%20Black%20Symphony.JPG
only probleme, it does not deinterlace
ADude
3rd March 2009, 03:04
Are there any performance improvements to the software decoder in 1.9.0 or is the hardware acceleration (and other fixes mentioned in the change list) the only difference ?
- Licensed CoreAVC Pro user
BetaBoy
3rd March 2009, 03:12
ADude... read the thread (but it looks like you did). No perf mentions in the changelog.
BetaBoy
3rd March 2009, 03:14
The amount that your cores are loaded has absolutely nothing to do with CoreAVC and everything to do with your operating system's scheduler.
A better measurement of CoreAVC's threading would be how much each thread is loaded.
Thx DS... you beat me ;-)
meatwad
3rd March 2009, 11:14
Has anyone else experienced issues with older versions of coreavc using vmr9 or vmr7 after unregistering/uninstalling 1.9? Some streams will no longer work with older versions once 1.9 has been registered/installed. 1.9 is fine, but the seeking issues have become too distracting for me and I wanted to use 1.6 again (still the most compatible version with the player and htpc software I use). Anyway, I've tried unistalling all my codecs, but the issue is still there. 1.6 will work after 1.9 was installed, but it's hit or miss on whether it will display a tv stream if I use vmr9 or vmr7 (which of course means I lose deinterlacing if I use the generic overlay). Is there an easy fix or am I stuck with 1.9 for the time being?
DJ Bobo
3rd March 2009, 12:03
@ Dark Shikari / BetaBoy
How would I do that? the idea jumped to my mind because somewhere along the changelog of CoreAVC there is something about better multiple core balance, which doesn't seem to work on the Core2 laptop (with XP), but seems to work fine on my X2 laptop (with Vista).
@ carlo
DXVA decoding is off topic, let's keep this for another thread.
But now that I've seen your new screens, I saw that you disabled deblocking. Could you have disabled deblocking in CoreAVC too? That may explain why your A64@2.5GHz is able to compete with my X2.
@ anton
And how much is needed to decode it properly from start to end? Mind you, my X2 should be slightly slower than a pentium dual core e2140 (1.6GHz).
DXVA decoding is off topic, let's keep this for another thread.
But now that I've seen your new screens, I saw that you disabled deblocking.
Skip deblocking = "None" means that deblocking is not skipped.
Secondly, those are libavcodec settings and they don't have any effect on DXVA decoding. I don't think that the current ATI and NVIDIA hardware implementations even allow disabling in-loop deblocking by an user-controllable setting.
Remember that he is playing a previously decrypted stream. Playing Blu-ray directly from the disc with PowerDVD is another matter and that would require a somewhat faster CPU because of software AES decryption and other overhead.
leeperry
3rd March 2009, 12:29
How would I do that? the idea jumped to my mind because somewhere along the changelog of CoreAVC there is something about better multiple core balance, which doesn't seem to work on the Core2 laptop (with XP), but seems to work fine on my X2 laptop (with Vista).
add this to your XP boot.ini, and update to SP3 of course :
/NOEXECUTE=ALWAYSOFF /FASTDETECT /USEPMTIMER /NODEBUG /TIMERES=9766
it will kill your batteries faster on a laptop, but it will also increase the system snappiness.
this is also a good idea, plus running your media player in high priority and all the background stuff in low(process.exe in a batch is your friend) :
Windows Registry Editor Version 5.00
[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\PriorityControl]
"Win32PrioritySeparation"=dword:00000016
DJ Bobo
3rd March 2009, 12:38
Skip deblocking = "None" means that deblocking is not skipped
Oh, got things mixed up, thanks for clarifying this (not good though, would mean that I go back to the "not well optimized" theory :devil:)
@ leeperry
Thanks, but the Core2 laptop is not mine, and I don't think my friend would like me "messing up" with the registry and boot files. I thought, it was about some program that can measure that...
Gleb Egorych
3rd March 2009, 21:59
Betaboy, when the next version is planned to be released? As I understand all major 1.9.0.0 bugs are already fixed.
BetaBoy
4th March 2009, 01:11
Gleb Egorych... in general I no longer announce releases unless its 'on deck' and through QA. We are still working on interlaced support atm so the best I can say is that there is no release scheduled for this week.
IgorC
4th March 2009, 02:34
An E5200 at 1200MHz is just a little bit too slow to decode the WallE Blu-ray with CoreAVC (from hard drive, .m2ts, decrypted).
Did you downclock your E5200? Because its stock frequency is 2.5 Ghz. I get smooth playback of blu ray 30 Mbit/s on HDD with E2160.
BTW, nice avatar with Shrödinger's equations.
BetaBoy
4th March 2009, 07:48
Some clarification for those from DVinfo....
Input Levels
- TV (16-235): (ITU-R BT.601) always assume the stream uses TV levels.
- PC (0-255): (ITU-R BT.709) always assume the stream uses PC levels.
- Auto detect: use the full range flag in the stream to determine Luminance range.
Note: The input levels setting allows the overriding of the stream's colorspace setting and affects the conversion to RGB color space when it is done by the decoder.
TV = ITU-R BT.601
and
PC = ITU-R BT.709
This is now in our KB and will be added to the next update (within the help tab).
leeperry
4th March 2009, 10:09
TV = ITU-R BT.601
and
PC = ITU-R BT.709
This is now in our KB and will be added to the next update (within the help tab).
too bad this is false information.
SD = 601 / HD = 709...except FRAPS videos, there's no PC levels videos to begin w/
squid_80
4th March 2009, 10:43
too bad this is false information.
SD = 601 / HD = 709...except FRAPS videos, there's no PC levels videos to begin w/
What? You can't determine the colorspace used based just on the resolution.
There is a flag in the H264 stream that specifies if it is full range or not. This setting allows the decoder to ignore that flag.
leeperry
4th March 2009, 10:47
What? You can't determine the colorspace used based just on the resolution.
There is a flag in the H264 stream that specifies if it is full range or not. This setting allows the decoder to ignore that flag.
oh sure there's a flag....which is sometimes wrong from what I read?
Yes you can define the YCbPr decoding matrix depending on the resolution.
Rec. ITU-R BT.601-5 => SD(PAL/SECAM/NTSC)
Rec. ITU-R BT.709-4 => HD
a while ago I asked Haali if he could add this in HR, w/ x<1024=601, x>1024=709...I also use the same rules in ffdshow, everything's pretty fine ;)
BTW, I understand Haali is working for your firm....any chance you could let him improve HR a bit? or is this really out of scope? :(
squid_80
4th March 2009, 11:28
But there's nothing stopping anyone from encoding in SD using 709 or HD using 601, is there?
leeperry
4th March 2009, 11:44
But there's nothing stopping anyone from encoding in SD using 709 or HD using 601, is there?
well, if you encode HD to SD w/o using ColorMatrix() to convert from 709 to 601....sure :o
and if you encode an SD upscale to HD w/o converting from 601 to 709, same problem will occur.....but luckily the chances of encountering these 2 cases is rather thin, and these are badly encoded material anyway...not specs compliant.
DVD=601, HDTV/BD=709 :cool:
squid_80
4th March 2009, 11:56
and if you encode an SD upscale to HD w/o converting from 601 to 709, same problem will occur.....but luckily the chances of encountering these 2 cases is rather thin, and these are badly encoded material anyway...not specs compliant.
Not specs compliant? According to what specs, yours? Otherwise show me where it says this in the AVC specifications.
leeperry
4th March 2009, 12:04
Not specs compliant? According to what specs, yours? Otherwise show me where it says this in the AVC specifications.
well, it's not AVC specific AFAIK.
there's no such thing as SD commercial AVC, these are mostly encodes from SD sources I think...and SD requires Rec. ITU-R BT.601-5 YCbPr decoding matrix coeffs
and HD is using Rec. ITU-R BT.709-4, whatever in VC1/MPEG2/AVC/DivX...you name it :
http://www.google.com/search?hl=en&safe=off&q=HD+ITU-R+BT.709-4&btnG=Search&lr=
this only matters if users enable the RGB output in CoreAVC, YV12/YUY2 are pass-through...and either ffdshow or the graphic card drivers will (try to) guess whether it's 601/709 depending on the resolution when converting to RGB.
Audionut
4th March 2009, 12:33
there's no such thing as SD commercial AVC,
Bold claim. Just because you haven't come across it, doesn't mean it doesn't exist.
and SD requires Rec. ITU-R BT.601-5 YCbPr decoding matrix coeffs
Choose your words more carefully. SD can be any color space the author chooses. HD too.
leeperry
4th March 2009, 12:48
Bold claim. Just because you haven't come across it, doesn't mean it doesn't exist.
Choose your words more carefully. SD can be any color space the author chooses. HD too.
well the NTSC/PAL/SECAM specs give BT.601, sorry for that :
http://www.google.com/search?hl=en&safe=off&q=NTSC%2FPAL%2FSECAM+601&lr=
decodes all variations of PAL, SECAM, and NTSC signals into standard ITU-601 compatible component colour values
just like HD is using Rec.709, this is mandatory :o
probably you're mixing YCbPr decoding matrix coeffs and gamut coordinates....HD can indeed use many different types of gamut(Rec 709/SMPTE RP 145/EBU Tech. 3213), and so does SD(SMPTE RP 145/EBU Tech. 3213)
and I always choose my words very carefully, just so you know :p
Audionut
4th March 2009, 13:03
well the NTSC/PAL/SECAM specs give BT.601, sorry for that :
http://www.google.com/search?hl=en&safe=off&q=NTSC%2FPAL%2FSECAM+601&lr=
A google page. Come on.
decodes all variations of PAL, SECAM, and NTSC signals into standard ITU-601 compatible component colour values
What a decoder from your google link is capable of has no bearing on the ability for SD to be encoded in any color space.
just like HD is using Rec.709, this is mandatory :o
No it's not.
I'm relying on my mobile phone atm until I get my adsl transferred.
I'll be more than happy to post samples of SD and HD material encoded in varying color spaces when I get adsl back.
squid_80
4th March 2009, 13:27
and I always choose my words very carefully, just so you know :p
Not carefully enough. You're talking about SDTV and HDTV as if they mean the same thing as SD and HD.
carlo_0000
4th March 2009, 18:16
@dj bobo
no debloking is enable in coreavc
and for mpchc it s enable too, i m not english but for me
skip debloking : none, it s mean that it do not skip deblokin so it s = to enable no ? because it no skip
no ?
DJ Bobo
4th March 2009, 18:45
@ carlo
OK, so you're saying that deblocking is enabled in CoreAVC right? (Deblocking: Standard)
That would mean that under similiar circumstances, you're achieving with an A64@2.5GHz what I'm achieving with an X2@1.9GHz.
To remove any doubts, I'd like you -if possible!- to give me the average CPU load using the Throttlewatch program (http://www.softpedia.com/get/Tweak/CPU-Tweak/ThrottleWatch.shtml) (this program doesn't need any installation, just decompress somewhere)
You just start Throttlewatch, start the Seven Pounds trailer, click back on the Throttlewatch window, wait until the trailer reaches 20 seconds, hit F5 to start logging, wait until the trailer reaches 2 minutes and hit F5 again. You should get a journal.txt file in your Throttlewatch directory with an average load specified at the end.
If possible also a screenshot of MPC playing the trailer with statistics enabled to check for dropped frames (preferably a screenshot after the player finished playing)
Merci :)
leeperry
5th March 2009, 00:23
I'll be more than happy to post samples of SD and HD material encoded in varying color spaces when I get adsl back.
I'm talking about everyday consumer content Mr NitPicker :D
DVD/VHS rips/DVB/SDTV/HDTV/HDDVD/BD/VCD/SVCD you know...sure you can encode a movie in FCC YCbCr colorspace using some obscure gamut Germany was using during WW2...but then don't count on any consumer PC software to decode it properly me thinks, not w/o some heavy avisynth scripting in ffdshow.
Not carefully enough. You're talking about SDTV and HDTV as if they mean the same thing as SD and HD.
oh well, decode 1080p in 601 if you like, and keep 709 for full range videos only...I'm cool w/ that :cool:
BTW the major lag I had at the beginning of movies(when CUDA was enabled) was apparently due to the "delay compensation" option in Ozone4 that was fighting against it, disabling it fixed the problem altogether.
when I enable it, it says 68ms...in 48Hz :confused:
http://www.image-load.eu/out.php/i148333_delay.jpg
carlo_0000
5th March 2009, 17:19
@ carlo
OK, so you're saying that deblocking is enabled in CoreAVC right? (Deblocking: Standard)
That would mean that under similiar circumstances, you're achieving with an A64@2.5GHz what I'm achieving with an X2@1.9GHz.
To remove any doubts, I'd like you -if possible!- to give me the average CPU load using the Throttlewatch program (http://www.softpedia.com/get/Tweak/CPU-Tweak/ThrottleWatch.shtml) (this program doesn't need any installation, just decompress somewhere)
You just start Throttlewatch, start the Seven Pounds trailer, click back on the Throttlewatch window, wait until the trailer reaches 20 seconds, hit F5 to start logging, wait until the trailer reaches 2 minutes and hit F5 again. You should get a journal.txt file in your Throttlewatch directory with an average load specified at the end.
If possible also a screenshot of MPC playing the trailer with statistics enabled to check for dropped frames (preferably a screenshot after the player finished playing)
Merci :)
ok done, it s 38%
some higher whayt i thinked with taskmanager
but it say clock frequency 4570mhz (little bug because it s not possible, it s 2.5g /core)
http://ftp1.dommel.be/picture_library/journal.txt
http://ftp1.dommel.be/picture_library/x2coreavc.JPG
DJ Bobo
5th March 2009, 17:39
@ carlo
Sorry I wasn't clear enough, I asked for your numbers with the Athlon 64 @ 2.5GHz, not with the X2. That 1080p works on the X2 is pretty much established.
By the way, your links don't work.
vBulletin® v3.8.11, Copyright ©2000-2025, vBulletin Solutions Inc.