CoreCodec/H.264 Codec "CoreAVC" [Archive] - Page 49

Rectal Prolapse

8th December 2006, 17:55

That sucks. :(

If I ask for a new activation for the 2nd OS install, will it disable the 1st activation on the 1st OS? I hope not. It is physically on the same PC. I can even install it on the same hard drive on the same partition (satisfying the one copy only requirement). So why can't I use both, one at a time?

I hope this gets fixed...

Inventive Software

8th December 2006, 18:31

Careful you two. If you would like to rant at CoreCodec's so-called "activation", there's a ranting thread (almost) exclusively for it here (http://forum.doom9.org/showthread.php?t=118775). :D

The only solution I can see Rectal Prolapse is to buy 2 licenses. Either that, or use ffdshow for the editing if you can get away with it.

Foreigner999

8th December 2006, 18:54

kamaleon

12th December 2006, 17:36

Hi folks, how's it going, my first post here ever :D
Does corecodec work with 4:2:2? sorry but the search feature doesn't seem to gimme any results, might have to do with the :

I'm trying to know if it'd work on 4:2:2 satellite broadcasts.

Manao

12th December 2006, 17:51

I'm not aware of any 4:2:2 H264 encoders, except the reference software. So how would you get 4:2:2 H264/AVC satellite broadcasts ?

kamaleon

12th December 2006, 17:58

Hi manao, thanks for the answer. I didn't know that. Good to know now.

The reference software you mean's coreavc?

KoD

12th December 2006, 18:04

The reference h264 software is this one: JM (http://iphome.hhi.de/suehring/tml/).

kamaleon

12th December 2006, 18:22

thanks KoD. i'll try to figure out what "reference" means :D

Shinigami-Sama

12th December 2006, 20:59

thanks KoD. i'll try to figure out what "reference" means :D

referance to the defined standard

kamaleon

12th December 2006, 22:36

you mean there's the standard, and then there's one particular codec that's the "reference" :D

Shinigami-Sama

13th December 2006, 01:05

you mean there's the standard, and then there's one particular codec that's the "reference" :D

yes
they make the standard and there needs to be referance encoder so that tests can be made to ceritify that they create compatible datastreams

kamaleon

13th December 2006, 09:36

ah, now I see much better, sorry for the OT ;)

cc979

13th December 2006, 11:52

if you need a better search try

whatever you are searching for site:forum.doom9.org

eg. h264 encoder 4:2:2 site:forum.doom9.org

kamaleon

13th December 2006, 19:00

if you need a better search try whatever you are searching for site:forum.doom9.org
eg. h264 encoder 4:2:2 site:forum.doom9.org

Right on, thanks mate ;) should have thought about that :D

toytown

19th December 2006, 00:08

Hi i finally got my Coreavc 1.2 to work today and decided to try it with some new test clips that i got some time ago, but never had the chance to playback due to my bad system.

Now i have a c2d e6600 and im trying to playback a AVC/H264 stream with an average bitrate of 15.5Mbps (1920x1080) and noticing bad framerate and general jerky performance, when i checked task manager i noticed that core is only using 1 core of the processor. Checking the COREAVC page its meant to be SMP for the pro version that i have, now i never got a chance to test this before, due to having a single core machine. But can anybody else confirm this, or give me any reason why this would happen. If i check in WMP, its definetely using COREAVC as the codec, and i totally removed the previous version.

EDIT - I rolled back to a WHQL version of Nvidia drivers and the jerkiness is all gone, still not sure if CORE is using both cores, but the video is playing fine now.

Romario

19th December 2006, 01:16

Hi guys, just one thought. CoreAVC 1.2 is cracked. Finally.

sillKotscha

19th December 2006, 02:03

Hi guys, just one thought. CoreAVC 1.2 is cracked. Finally.

odd news (http://forum.doom9.org/showpost.php?p=916887&postcount=118)... wrong place to post...

pankov

19th December 2006, 02:14

Betaboy, ... or anybody from the CoreAVC support team. I'm still waiting for the upgrade e-mail
it's almost a month now ... may be I've lost track of time but it's very long anyway ... and I haven't received any response to my PMs, ICQ and e-mail messages (both to CoreAVC support and to Betaboy personally).
May be it's time for me to start looking for an alternative method of getting 1.2
I didn't believe I'll be forced to do this, though
:(

Rectal Prolapse

19th December 2006, 17:24

pankov, I feel your pain. I was not happy and vented my frustration here and got flamed for it.

Anyways, moving on:

toytown, it is possible that you still have an old version of CoreAVC kicking around in your system. Did you install any codec packs? I would do a complete search of your C: drive for coreavc.ax files (or whatever it is called) and unregister and delete them. Then reinstall CoreAVC 1.2.

Good luck!

Hans Ohlo

20th December 2006, 11:03

one thing i like to be fixed is the following:
http://img299.imageshack.us/img299/8547/coremessng9.jpg
the installer misses to remove the old entry in this software list. or it should tell me to remove the old version first (which the new version simply overwrites).

toytown

21st December 2006, 22:46

toytown, it is possible that you still have an old version of CoreAVC kicking around in your system

I checked before i posted ,there is only the one dll for core installed and its the 1.2 version registered to myself. With the 97.xx (i think) drivers i would get about 55% total cpu usage and jerky playback, this led me to believe it was only using the 1 core (50% for the video, 5% for other useless windows tasks running).

When i downgraded back to the earlier WHQL nvidia drivers, my cpu usage whilst watching the movie dropped to around 40% usage. Whilst bringing up the statistics i also get about 1-5 dropped frames after running for several minutes. But the cpu never peaked above 60% total usage according to task manager.

This leaves me to believe that its only using 1 core and that taskmanager is only showing usage on both cores due to the task switching cores constantly.

Is it possible that there is a SUPER HIGH quality test mp4 i can use, that would simply not play on a single core machine, or even a method of seeing if the application is really using multiple cores?

JohnnyFu

21st December 2006, 23:36

Is it possible that there is a SUPER HIGH quality test mp4 i can use, that would simply not play on a single core machine

Try this one: beyonce.at.the.bbc.1080mbaff.sample.ts (http://mirror05.x264.nl/public/force.php?file=./beyonce.at.the.bbc.1080mbaff.sample.ts)
CPU usage is almost 100% the whole time on my 3500+@2.7Ghz (single core).

Do you have the AMD Dual-Core Optimizer (http://www.amd.com/us-en/Processors/TechnicalResources/0,,30_182_871_9706,00.html) installed ?

JohnnyFu

23rd December 2006, 01:51

I just wrote a support ticket. 5min later i had a reply.

Ticket ID : 19982
Ticket Title : CoreAVC Installation

Dear ,

The Ticket is replied and the details are :

Status : Open
Category : Big Problem
Severity : Major
Priority :

Description :
Hello, I'm going to format and reinstall my Windows XP system now, as far as i know, after that, i will not be able to install CoreAVC again. This is unacceptable for me. I really hope you can provide a solution.... greetz Johnny I bought your product to support your company... don't throw it away.... otherwise it was the last time i supported you. I will not and can not buy a new license every 2-3 months when i clean up my system....

Solution :

Regards, HelpDesk Team

is this an error or do they kidding me ?

jchunter_2

23rd December 2006, 02:03

CoreCodec is still taking orders for the CoreAVC codec and charging credit cards for the codec but not delivering anything. :(

ChronoCross

23rd December 2006, 02:29

I just wrote a support ticket. 5min later i had a reply.

is this an error or do they kidding me ?

you do know that as long as you use the same hardware and windows key you can reactivate automatically right?

JohnnyFu

23rd December 2006, 05:05

you do know that as long as you use the same hardware and windows key you can reactivate automatically right?

no, i didn't know, thank you :) so... everything should be fine then.

Sharktooth

23rd December 2006, 05:19

you do know that as long as you use the same hardware and windows key you can reactivate automatically right?
The exact reason why i wont buy it.
I change HW very often...

24th December 2006, 17:29

The exact reason why i wont buy it.
I change HW very often...
After my last upgrade I see that I don't need not coreavc, nor powerdvd hwa for now, ffdshow plays even 1080p 24 Mbps streams with 60%/40% cpu cores load.....

CruNcher

24th December 2006, 20:35

yes for low complexity streams (progressive) the difference is only 25% now
so you can allready playback 720p H.264 @ 30 fps with 800 mhz (MMX,SSE1) x86

KoD

24th December 2006, 20:56

I would like to know where all these numbers come from. Does anyone care to provide some sample files and timed test runs with ffdshow and coreavc ?

foxyshadis

25th December 2006, 00:46

I would like to know where all these numbers come from. Does anyone care to provide some sample files and timed test runs with ffdshow and coreavc ?

http://forum.doom9.org/showthread.php?p=914647#post914647

0.4 tests single-threaded, 1.0-pre tests multithreaded. Only with dual core is the difference huge, and then only if you have something that can actually saturate the system, like hd-dvd.

Ignore fps in timecodec and concentrate on dfps, which measures total time to complete. (fps measures thread-time and gets faked out by the clock wonkiness in speedstep and c&q systems.)

KoD

25th December 2006, 18:35

lexor

26th December 2006, 16:19

What I would like to see are timed tests from Px and Cruncher that proves the numbers in their posts. I would definitely like to see how ffdshow benefits from the second core for Px and timed tests from Cruncher on that 800 Mhz machine with a 720p file. Let's keep on topic.
how is any of that on CoreAVC topic?

Romario

26th December 2006, 20:00

BetaBoy gets what he deserves, really!

27th December 2006, 01:43

ChronoCross

27th December 2006, 04:46

http://users.i.com.ua/~finish/HardwareMonitoring.hml
This is RivaTuner text log for playing first 1:47 of this file - http://multimediacom.free.fr/HD-DVD/H264/24Mbps/

This is in this thread why? Why don't you open a new thread for this as it's completely separate from coreavc

Selur

27th December 2006, 08:37

little question:

Atm I'm running WinXP but I plan to check out Vista in a couple of weeks (MSDN AA), do I have to fear any problems installing my coreavc pro version on vista?
And If I switch back to WinXP (restore an image) will CoreAVC still run?

Cu Selur

JohnnyFu

27th December 2006, 11:29

oh... FYI... current installer is not compatible with Vista... We will release a Vista capable installer later today.

that's all i know about Vista
Restoring an image wich includes an already activated CoreAVC should work i think. But i don't think you will be able to install/activate CoreAVC on Vista if it was activated on an XP installation before.

KoD

27th December 2006, 13:44

I agree those last posts were off topic, but someone should have said so when the fire started, not when the house was already burned to ground.

What I don't like is this "let's keep on bashing about everything, it makes us happy" and "nobody complains when we see bashing, we like it that way" attitude. If you have something to complain about, that's the activation system, don't make unfounded estimates on coreavc vs ffdshow performance. Prove your point with test results, those are the only things that matter.

P.S.: Px, how should I interpret that file ? Use Haali's timecodec for easy to interpret results.

27th December 2006, 23:50

This is in this thread why? Why don't you open a new thread for this as it's completely separate from coreavc
Maybe because current and future coreavc users interests in inffo about alternative codecs speed? ;)

ChronoCross

28th December 2006, 00:57

I think it's just a reason for you guys to get back at core for adding activation. Although I can alter benchmarks to say whatever I want.

It's never really about benchmarks. If it was then you wouldn't be posting it in a competitors thread you'd be posting it in your own thread.

Foreigner999

28th December 2006, 06:50

Although it may be ethically questionable im sure people looking to buy Coreavc would welcome a variety of information (accurate of course) that would help them make the best informed decision.

If nero had the best decoder, I sure would want to know it when im looking up info on the best decoder available. I use it, this method, when choosing hardware for my computers so that i can make the best informed decision based on past experience of others.

Besides, Betaboy has in the past welcomed comparisons of coreavc against the competition.

28th December 2006, 11:42

It's never really about benchmarks. If it was then you wouldn't be posting it in a competitors thread you'd be posting it in your own thread.
You can read in mind of other peolpe? :D
Or you don't like when someone shows that there is no necessary in your CoreAVC? ;)

KoD

28th December 2006, 14:13

Px, I'm still waiting for those timed test runs.

Use null renderer so that no time is spent in graphic card driver calls and we would actually measure codec performance and not graphic card performance; this will also take care of bugs in various driver versions for some of the video renderer modes. Post a link to the file you run the test on, run several times for both ffdshow and coreavc. Use yv12 output only from both so that no time is spent in color conversions. Keep standard deblocking in CoreAVC and in ffdshow do not enable any of the inloop deblocking disabling options on the Codecs page. Run in ffdshow without any other option enabled, that is no resizing, no additional postprocessing, no deinterlacing (ffdshow is silly enough to try performing deinterlacing on non interlaced material and ruin the results). If your material is interlaced, then do test runs without performing deinterlacing in either CoreAVC (no deinterlacing) or ffdshow (deinterlacing tab disabled) and then a test run with software deinterlacing in both CoreAVC (software deinterlacing) and ffdshow (enable the Deinterlace tab and select one deinterlacing method; I'd recommend 5-tap lowpass, although the quality is lower than that of CoreAVC software mode).

And please post here for everyone to see. The PM you sent me did not follow the above procedure, you did not run coreavc and ffdshow separately or at least you did not mention which one was which, and you also changed renderer from one comparison to the next which made it look like an apple to pears comparison.

28th December 2006, 14:45

Px, I'm still waiting for those timed test runs.

Use null renderer ......
Ok, when I have enough time to do all that you described, I'll do it....
And fyi - ffdshow now multithreaded in output samples quering, so with null redner only one core used
Run in ffdshow without any other option enabled, that is no resizing, no additional postprocessing.
Fyi - timecodec ignores all this stuff ;)

And please post here for everyone to see. The PM you sent me did not follow the above procedure, you did not run coreavc and ffdshow separately or at least you did not mention which one was which, and you also changed renderer from one comparison to the next which made it look like an apple to pears comparison.
I see, that you don't understand what I wrote to you. Ok, once again: in file, presented above, in first and third columns from end there are cores usage when playing test file. You can see how cpu usage distributed between cores.
Second part, is the fast test of same file, that shows that timecodec if useless for such tests, because
Overlay mixer, 1 core at 100%, second on 3-5%
User: 38s, kernel: 0s, total: 38s, real: 40s, fps: 30.3, dfps: 29.2
VMR9, both cores at 100%
User: 45s, kernel: 0s, total: 45s, real: 51s, fps: 25.8, dfps: 22.8
Same problem with timecodec and hardware accelerated codecs, it don't use hardware acceleration...

clsid

28th December 2006, 15:43

CPU usage doesn't mean anything.

To measure the pure decoding performance you should use the null renderer. The only important value is dfps.

When the dfps goes below realtime playback, then stuff like the queue in ffdshow can become a factor and a comparison using Overlay or VMR-9 can be interesting.

Timecodec will use an instance of the ffdshow filter with whatever settings are active in ffdshow. So you must disable all filters in ffdshow.

CoreAVC and ffdshow both don't have hardware accelerated decoding. So that is not an issue.

ChronoCross

28th December 2006, 16:21

You can read in mind of other peolpe? :D
Or you don't like when someone shows that there is no necessary in your CoreAVC? ;)

I read the minds of the people who have been posting in this thread as the last 50 posts have all been coreavc bashing. So yeah I think it's reasonable to be a little skeptical of your intentions.

As for not needing coreavc. ffdshow still doesn't come anywhere near it. Sorry to say.