CoreCodec/H.264 Codec "CoreAVC"

[ChronoCross]

So you claim all those people are liars? Anyway, I think you are failing to understand the point that foxyshadis is trying to make. Why contact support at all if you can fix something yourself?

And given proper instructions even retards can safely edit the registry.

What I'm claiming is just because those few had issues doesn't mean that's how it is with every issue.

I understand his point but obviously he didn't fix his issue without breaking the law.

[BlackSun]

After investigation, the reseller (ejunkie) was sending the wrong file to new customer. If you get the "A dynamic link library (DLL) initialization routine failed" error, please visit the following URL:

http://www.coreavc.com/retrieve

Type your Paypal email address that you used to buy CoreAVC, an email will be sent to that address that contain:

- Your serial number
- A download link which is not from e-junkie and that provide the proper file.

I added an entry in our knowledge base about it.

[Romario]

people in China doesn't even know what is serial number, or in Albania.

[Disabled]

After investigation, the reseller (ejunkie) was sending the wrong file to new customer.
Unfortunately it wasn't just new customers. I got a wrong version too that stopped working yesterday, but thanks to Viper Zx and Dr.Penix I came up with the idea that it might be a wrong version and got a corrected one from your link posted.
Would have been great if you would have mailed your customers, as the error message given was not very descriptive. And as molitar shows, ppl got confused.

[Shinigami-Sama]

What I'm claiming is just because those few had issues doesn't mean that's how it is with every issue.

I understand his point but obviously he didn't fix his issue without breaking the law.

I'm fairly certain foxy is a girl ( Remembers that 'how did you get your nick' thread from a while back )

anyways...


foxy also said 2 to 200 hours, so some people do get speedy service, while sadly, what looks like more than a handful do not speedy service


and back on topic
is there an ETA for the next release?

[KoD]

people in China doesn't even know what is serial number, or in Albania.

I think you meant to say: "People in China don't need serial numbers. They reverse engineer the software and make serial number generators for themselves." Self-help taken to the extreme. What's so expert in editing the registry when you can help yourself like that !

PS: for people that have no sense of humour, I'll say I'm quite aware CoreAVC is the best h264 software decoder out there and that reply above is not meant to be rude.

[BetaBoy]

Hi!

Thanks for re-introducing the deblocking options. With them it's indeed possible to shave a few percent off CPU usage, which is my main concern.

However, 1.6.5 doesn't seem to make good use of both my (logical) CPU threads. Previous versions fared better in that regard.

Here is a comparison with a late build of ffdshow (rev1803_20080120_xxl), used on a high bitrate 720p x264 file.

http://img85.imageshack.us/img85/5127/core165xu2.jpg

Regards
S.


I had this on my todo to answer....

The likely cause for such a CPU condition started with the 1.6 release of CoreAVC NG (Next Generation) filter.

CoreAVC has frame based SMP and iirc ffmpeg is splitting the tasks at a lower level. In our case it could be several factors... 'frames' can produce an uneven preformance or the OS thread handling... but when CoreAVC is loaded it should start to use the other core at full load right away.

We will test this above internally by making a version that doesn't have "hints" for which cpu decoding thread it should use and see if there is a diff.

[foxyshadis]

foxy also said 2 to 200 hours, so some people do get speedy service, while sadly, what looks like more than a handful do not speedy service

That was my point. I'm not bagging specifically on core here, but ANYTIME you email support at any company, you're taking a chance that it's going to take hours or days to get a human answer, if ever. I'm a professional in another field, I deal with this, and even I forget to answer every email and request that comes my way. (Apologies to anyone who PMs me unanswered, btw, nothing personal.) I always prefer self-support options, since I might fix it myself long before support gets back to me. Five minutes beats five hours, you know?

Core is probably infinitely grateful that they had the foresight not to post a phone number on their web site. :p

And yes, I even have the icon there, but it never seems to help~ oh well~!

[BetaBoy]

We actually have two numbers on the sites to call.... and in about 2 weeks an 800 number for support and licensing questions.... but yes we agree that ppl want self serve options... this is what we are striving for. Not to get too OT... but we are also going the same route with CorePlayer, but with it (and for more obvious reasons) because of features like IM, Sharing, Purchasing we will have a more 'connected' approach, especially as we move outwards to platforms like the iPhone...

http://img405.imageshack.us/img405/4033/snap023225pr3.th.jpg (http://img405.imageshack.us/my.php?image=snap023225pr3.jpg)
and
http://img169.imageshack.us/img169/8582/snap023334mw1.th.jpg (http://img169.imageshack.us/my.php?image=snap023334mw1.jpg)

While those pics on the iPhone don't show the CoreAccount integration... in the next few milestones it will.

[sbcale]

Hello, I am trying to convert an IronMan trailer that is in h.264 to WMV using zambelli's wmcmd.vbs. Here is my graph and here is the resulting log from the encode that doesn't work.

Please help. I have tried using the Sonic h.264 decoder but it introduced green and purple blocky parts to the wmv.

:thanks:

http://scale.smugmug.com/photos/252189354_8bUYS-L.jpg

Microsoft (R) Windows Media Encoder Command Line Script Utility
Copyright (C) Microsoft Corporation. All rights reserved.

Warning: Encoder failed to retrieve duration attribute from source plug-in.
.

Warning: Encoder failed to retrieve source duration again.
Encode process might not continue or progress report may be inaccurate.



======== Encoding Completed ========

Audio :
Codec: Windows Media Audio 10 Professional
Expected bit rate: 384000 bps
Average bit rate: 384000 bps
Expected sample rate: 5859
Average sample rate: 5859
Dropped byte count: 0 bytes
Dropped sample rate: 0
Total bytes: 7151616 bytes

Video :
Codec: Windows Media Video 9 Advanced Profile
Average bit rate: 49408 bps
Expected fps: 24
Dropped frame count: 1
Total coded frames: 238
Average sample rate: 24.001
Dropped bytes: 39 bytes
Total bytes: 78468 bytes

Overall:
Encoding time: 18 seconds
Average bit rate: 433408 bps
File size: 7481135 bytes
File duration: 148.735 seconds

[qyqgpower]

First, thank you for adding support for MV out of specs, letting me play old buggy files without artifacts. But today I come up with serious artifacts when playback some m2ts files decrypted from BD.

http://www.image-load.eu/out.php/t23969_snapshot20080209142447.png (http://www.image-load.eu/out.php/i23969_snapshot20080209142447.png)

http://www.image-load.eu/out.php/t23970_snapshot20080209142602.png (http://www.image-load.eu/out.php/i23970_snapshot20080209142602.png)

I have made several tests:
CoreAVC 1.6.5.0 -> artifacts
CoreAVC 1.6.0.0 -> artifacts
ffdshow tryouts 080115 -> fine
ffdshow tryouts 080206 -> fine
Elecard -> fine
Cyberlink -> fine

so apparently it should be a CoreAVC issue.

here is a small logo clip, you can download it and reproduce the problem
http://www.live-share.com/files/304750/00001.m2ts.html

[BetaBoy]

Thx for the report.... we found this issue the other day and have already fixed it... its in cue for the next release.

[ACrowley]

@Momber ,I can report the same Issues.
CoreAVC 1.6.5 decoding doesnt use both Cores as as good as older Versions or ffdshow or Cyberlink.
The Core Usage is async.

Oh, and its not a DualCore Optimizer etc Problem here

[CruNcher]

@BetaBoy
http://cruncher.mufflastig.com/x264/webtest/h264-progressive-test.htm ;)

[bob0r]

TCPMP:
[-] Continue buffering when press pause.
[+] Easy double click on video for full screen.
[+] Video completely widescreen in full screen mode.

FLASH:
[+] Continue buffering when press pause.
[-] Video completely widescreen in full screen mode. (more like 4:3)

[-] Easy double click on video for full screen

Found the button :)

[BetaBoy]

@BetaBoy
http://cruncher.mufflastig.com/x264/webtest/h264-progressive-test.htm ;)

Nice.... again not too OT. But we are releasing our RTP/RTSP components into CPM and CPP on the 21st. Once thats out we will finish the progressive download work and alpha blending afterwards so i'll have a stand alone demo version for you to post like that.

[BetaBoy]

First, thank you for adding support for MV out of specs, letting me play old buggy files without artifacts. But today I come up with serious artifacts when playback some m2ts files decrypted from BD.

http://www.image-load.eu/out.php/t23969_snapshot20080209142447.png (http://www.image-load.eu/out.php/i23969_snapshot20080209142447.png)

http://www.image-load.eu/out.php/t23970_snapshot20080209142602.png (http://www.image-load.eu/out.php/i23970_snapshot20080209142602.png)

I have made several tests:
CoreAVC 1.6.5.0 -> artifacts
CoreAVC 1.6.0.0 -> artifacts
ffdshow tryouts 080115 -> fine
ffdshow tryouts 080206 -> fine
Elecard -> fine
Cyberlink -> fine

so apparently it should be a CoreAVC issue.

here is a small logo clip, you can download it and reproduce the problem
http://www.live-share.com/files/304750/00001.m2ts.html

We had a closer look at the file.... and there is something else going on.... can you pls tell us what was used to encode it... and more specifically...what build number. Upon looking at it is seems not to be spec compliant but we want to confirm this.

[CruNcher]

BetaBoy i doub't he knows with what this Blu-Ray was mastered (encoded) with ;)

[BetaBoy]

ok.... we see what the problem is now.... this is another file that has out of range Motion Vectors (even beyond the tolerence that we added). They mastered it this way? Wierd... I wonder what they used to encode it?

[CruNcher]

unless it's a X264 or something else Consumer transcode of course :D
But he clearly states I come up with serious artifacts when playback some m2ts files decrypted from BD.
so it was obviously a pro Encoder that did this then hehe hmm maybe Mainconcepts (aka Sonic Cinevision) or Cinemacrafts CC-HDe or some Hardware Encoder, tough i think almost impossible to find out without knowing what the Studio that created this normaly uses, you could call them tough and ask ;) :D

[Dark Shikari]

ok.... we see what the problem is now.... this is another file that has out of range Motion Vectors (even beyond the tolerence that we added). They mastered it this way? Wierd... I wonder what they used to encode it?I did not see this coming. Oh the irony that it turns out the biggest problem with out of range motion vectors is on Blu-Ray discs! :eek:

[CruNcher]

I would really like to test this stream on my G92 core but i can't get it :D live-share is so crapy slow here that it brakes up the download.

[BetaBoy]

I did not see this coming. Oh the irony that it turns out the biggest problem with out of range motion vectors is on Blu-Ray discs! :eek:

Don't jump the gun yet.... we have no info if this is BR disc or if its mainstream or not.

[bob0r]

I would really like to test this stream on my G92 core but i can't get it :D live-share is so crapy slow here that it brakes up the download.

Yeah, hehehe i put it in flashget and let it run all night, anticipating this problem, here is the solution:

http://files.x264.nl/00001.m2ts :)

[BlackSun]

CoreAVC interoperability problems list:
Sonic HD Demuxer (EVO) (Hardware Deinterlacing Detection Fails frames get droped)
Sonic MPEG Demultiplexer (EVO) (No Connection Possible)
Mainconcept Mpeg Demultiplexer (EVO) (No Connection Possible)
Arcsoft MPEG Demux (EVO) (No Connection Possible)

Practicly all of those Decoders from these Companies are interoperable to each others Splitter/Demulitplexer, only CoreAVC has major problems here.

Mainconcept Mpeg Demultiplexer:
Mainconcept H264 Decoder = OK
Arcsoft Video Decoder = OK
Cyberlink H264 Decoder = OK
Sonic H264 Decoder = OK (old Mainconcept Decoder)
Sonic Cinemaster Video Decoder = OK
CoreAVC 1.6 = No Connection

Sonic HD Demuxer
Arcsoft Video Decoder = OK
Cyberlink H264 Decoder = OK
Sonic H264 Decoder = OK
Sonic Cinemaster Video Decoder = OK
CoreAVC 1.6 = Hardware Deinterlacing Problems (frames skipped)

Sonic MPEG Demultiplexer
Arcsoft Video Decoder = OK
Cyberlink H264 Decoder = OK
Sonic H264 Decoder = OK
Sonic Cinemaster Video Decoder = OK
CoreAVC 1.6 = No Connection

Arcsoft MPEG Demux:
Arcsoft Video Decoder = OK
Cyberlink H264 Decoder = OK
Sonic H264 Decoder = OK
Sonic Cinemaster Video Decoder = OK
CoreAVC 1.6 = No Connection

Cruncher, can you build a working graph in MPC, and send me text in Pin Info page for the demuxer output pin? (for each demuxers). Thanks.

[Inventive Software]

I did not see this coming. Oh the irony that it turns out the biggest problem with out of range motion vectors is on Blu-Ray discs! :eek:

+1. This is astonishing! :devil:

[ChronoCross]

I did not see this coming. Oh the irony that it turns out the biggest problem with out of range motion vectors is on Blu-Ray discs! :eek:

Once again the reason that standards need to be used.

[BetaBoy]

Point is we still know nothing about the file, if it was a Blu-Ray disc or not and what was used to encode it (we determined it was not x264)... yet assumptions are being made... -1.

Note that today we...
- Fixed: AVCHD M2TS blockiness
- Fixed: DS deinterlacing
- Fixed: INI for missing crop1088 and the value

We looked into the CPU core unbalanced issue and found that it is not caused by CoreAVC NG as we first thought but looks to be libcoreavc and an artifact of the windows scheduler... we'll continue to work on it.

We are also looking into a report by `md on a potential high FPS issue as it relates to CPU usage.

[BlackSun]

- Fixed: INI for missing crop1088 and the value


No that has already been fixed in the 1.6.5, I just forgot to commit the fix after doing the release :)

[Inventive Software]

No that has already been fixed in the 1.6.5, I just forgot to commit the fix after doing the release :)

Doh! :D

[BlackSun]

Doh! :D

Oh don't worry the fix was in the 1.6.5 :D

[Momber]

I had this on my todo to answer....

The likely cause for such a CPU condition started with the 1.6 release of CoreAVC NG (Next Generation) filter...

@BetaBoy: Thanks for the reply. I will follow further development with much interest.

S.

[odditory]

When is there going to be a CoreAVC ENCODER?

[ADude]

We looked into the CPU core unbalanced issue and found that it is not caused by CoreAVC NG as we first thought but looks to be libcoreavc and an artifact of the windows scheduler... we'll continue to work on it.
Please keep us posted on this one.

This could make a big difference in performance on dual core PCs...

[Shinigami-Sama]

When is there going to be a CoreAVC ENCODER?

when coreavc 2.0 comes out
so...
when pigs fly?

[BetaBoy]

/me notes to remind Shinigami-Sama to take pictures of our flying pigs then.

[BetaBoy]

Please keep us posted on this one.

This could make a big difference in performance on dual core PCs...

I'm not too sure of this... we have been testing the CPU load and even though its uneven, the performance (other then in high FPS case) seems to be about the same.

But give us a few days to continue to work on it... in the meantime Haali is looking into the directhshow pin issue.

[ADude]

I'm not too sure of this... we have been testing the CPU load and even though its uneven, the performance (other then in high FPS case) seems to be about the same.

But if the load on one core hits 100% at any point, then it is a problem.

Sure, if the total load is low enough that a single core is close to sufficient, then it won't matter.

The graphs posted earlier seemed to indicate that one of the cores was hardly being used at all...

[Momber]

The graphs posted earlier seemed to indicate that one of the cores was hardly being used at all...
True, but this may depend on the video stream, too. The graphs I posted earlier were captured during a fast-moving scene in a high bitrate 720p x264 re-encode.

Last night however, I tried Core 1.6.5 on a 1080i H.264 TV cap ("Contact" H.264, remuxed to mkv) and for the first time ever with my "antiquated" CPU I was able to watch such a cap smooth and stutter free without resorting to DXVA (Cyberlink).
Used Core, deblocking "skip always", deinterlacing "off", Overlay Mixer, YUY2. Zoom Player 5.00.

I'm fairly ecstatic about that, actually.

http://img227.imageshack.us/img227/4928/coreh2641080imf1.th.jpg (http://img227.imageshack.us/my.php?image=coreh2641080imf1.jpg)

As you can see, the load-balancing is better there. Always keeping in mind that my CPU only has "virtual" multithreading, as it's a P4 Prescott 3.0 GHz.

Greets
S.

[Sasovics]

Guys,

I must say the compatibility problem with AviSynth still persist even with the new version of CoreAVC 1.6.5

Still getting the annoying:

"CAVIStreamSynth: System exception - Access Violation at 0x0, reading from 0x0"


My graph is as follows:

Haali Splitter(EVO file) -> CoreAVC Video Decoder

Then my AVS script is as follows:

Code:

DirectShowSource("feature.grf",video=true,audio=false,fps=23.976,framecount=xxxx)
ConvertToYV12 ()

NOTE: I have no problem with above setup if I use the CoreAVC v1.5.0.1

Can you please guys try to focus little more on this issue too ? ... as I am not the only one complaining about this
, and unfortunately so far did not see any answer to this issue from CoreCodec guys :(

[BetaBoy]

Sasovics... I know its alot to keep up.... but this has been discussed several times over the past few pages and is planned to be fixed with the next release.

[Sasovics]

Sasovics... I know its alot to keep up.... but this has been discussed several times over the past few pages are will be fixed with the next release.

Thanks a lot BetaBoy. At least I know you are aware of this issue and are planning to fix it in the next release ;) Fair enough!

I can still use the 1.5.0.1 version in the meantime!

[CruNcher]

Point is we still know nothing about the file, if it was a Blu-Ray disc or not and what was used to encode it (we determined it was not x264)... yet assumptions are being made... -1.

Note that today we...
- Fixed: AVCHD M2TS blockiness
- Fixed: DS deinterlacing
- Fixed: INI for missing crop1088 and the value

We looked into the CPU core unbalanced issue and found that it is not caused by CoreAVC NG as we first thought but looks to be libcoreavc and an artifact of the windows scheduler... we'll continue to work on it.

We are also looking into a report by `md on a potential high FPS issue as it relates to CPU usage.


Betaboy does this means you fixed the problem of that sample, because i now tested it on my G92 core in Hardware Decoding and couldn't see any issues but the issue is clearly visible with CoreAVC especialy before that frame that's shown here you can see something even more strange Slice Decoding problems it seems also.

[molitar]

Question.. Is their any real purpose in CoreAVC when I now have an ATI 3870 video card? Since I now can run it in DXVA mode for x264 and use NO CPU USEAGE at all! It seems to me but I could be wrong their is no need for CoreAVC anymore until it supports Hardware GPU. I mean what advantages if any does CoreAVC have?

[Dark Eiri]

Well, I guess the 3870 cards have the same problem every other DXVA2 card has till now. It doesn't decode "out-specs" x264/H264, which happens to be 90% of the encodes out there.
So, yeah.

[oddball]

For it to decode it needs to be flagged with profile 4.1 and BluRay and HDDVD are. x264 is but only on a few encodes. Most people do not use it. Shame really. It makes no difference to size or quality.

[Threedcoder]

BetaBoy,

I see you work for CoreAVC. Well, I just spent my $7.95 to download the decoder only to discover it doesn't work with AviSynth. :) I see Sasovics has the same problem, only he has the older 1.5.0.1 version to fall back to.

However, I don't since I just purchased. Is there anyway I can download an older version (specifically 1.5.0.1) from the CoreAVC site so that I can continue working until a fix for this AviSynth issue is fixed?

Thanks.

[richardlynn]

Betaboy,

Could you please explain the options in the version1.6.5,like deinterlacing and deblocking?I'm puzzled about the new options and can't find any instructions about them.

[BetaBoy]

It doesn't decode "out-specs" x264/H264, which happens to be 90% of the encodes out there.
So, yeah.


1) Where did you get the 90% info from? From our demographic studies you are way off... that is unless your talking about open source and consumer, but even then your % is not even close. In any case pls posts facts not speculations.
2) You are wrong again.... If you have not followed this thread or the last release (which at times is alot too follow) we support out of range MV. There was one post from one persons saying he had an issue with non-standard BV with MV, but never a) sent info on the sample b) never stated a title c) never stated what was used to encode it with. So without more info...

[BetaBoy]

Betaboy,

Could you please explain the options in the version1.6.5,like deinterlacing and deblocking?I'm puzzled about the new options and can't find any instructions about them.

I will update the first post in this thread with the 1.6.5 changes...