CoreCodec/H.264 Codec "CoreAVC" - Page 295

me7 · 26th June 2010, 17:36

Sure: http://www.mediafire.com/?sharekey=8...4ac78345cbe4ce
I was able to reproduce the error in that sample by seeking back and forth.

Guest · 26th June 2010, 17:43

Thanks. Seeking is fine with that sample in DGDecNV, so it appears that CoreAVC did not implement the "back by one GOP" move when decoding open GOP streams (or did not do it correctly).

They could also skip leading B frames instead of using the "back by one GOP" move if accurate random access is not a requirement.

CruNcher · 26th June 2010, 20:11

Quote:

Originally Posted by me7

I think there is a problem with open-gop, committed to x264 a few days ago. Sometimes after seeking, the picture looks like this:

Skipping a few seconds back and rewatching the same scene is fine though, seems to be related to seeking. Happens with and without CUDA.
Encoded with "--open-gop display".

Doesn't happen with DivX Demuxer and your sample, i guess it's more a Haali splitter issue.
Ok it does happen though it seems to be a very random problem, doesn't happen on every seek.
It definitely doesn't happen with DivX H.264 Decoder so indeed it seems to be as neuron2 said that Decoder issue

bob0r · 30th June 2010, 11:51

LOL CoreAVC has this error for so long... (or in this case maybe haali) The same happens with BBC-HD. it always recovers and video is there to watch

ChronoCross · 3rd July 2010, 04:24

I've pretty much given up on all CoreCodec projects.

CoreAVC was great when it came out, but with updates/bug fixes taking 7 or more months it becomes less and less competitive as open-sourced decoders performance are rapidly increasing with each passing week.

CorePlayer 2 has been worked on for over 2 years now with the initial forum announcement stating this in August 2009. Now at nearly a year since that post CorePlayer 2 is no closer to being available.

CoreASP - 2 or 3 years since it was announced. Treading on Duke Nukem Forever Territory.

CoreAAC - announced around a year ago, got it's forum placeholder yesterday which means it's 1-2 years away from release.

This may seem harsh but I've always been a stanch supporter of CoreCodec and it's been nothing but disappointment for awhile now and the last 20 pages of this thread is supporting evidence of that fact.

fenomeno83 · 5th July 2010, 22:24

Quote:

Originally Posted by ChronoCross

I've pretty much given up on all CoreCodec projects.

CoreAVC was great when it came out, but with updates/bug fixes taking 7 or more months it becomes less and less competitive as open-sourced decoders performance are rapidly increasing with each passing week.

CorePlayer 2 has been worked on for over 2 years now with the initial forum announcement stating this in August 2009. Now at nearly a year since that post CorePlayer 2 is no closer to being available.

CoreASP - 2 or 3 years since it was announced. Treading on Duke Nukem Forever Territory.

CoreAAC - announced around a year ago, got it's forum placeholder yesterday which means it's 1-2 years away from release.

This may seem harsh but I've always been a stanch supporter of CoreCodec and it's been nothing but disappointment for awhile now and the last 20 pages of this thread is supporting evidence of that fact.

can you say what decoder that in your opinion works better than coreavc for h264?because I didn't find them

ChronoCross · 5th July 2010, 23:45

Quote:

Originally Posted by fenomeno83

can you say what decoder that in your opinion works better than coreavc for h264?because I didn't find them

DivX, DiAVC, libavcodec are all contenders who either come close or surpass CoreAVC in common situations.

libavcodec has the most bugfixes for problematic streams. I haven't really tested multi-threaded as I have a single core machine.

Audionut · 6th July 2010, 09:52

Quote:

Originally Posted by ChronoCross

CoreAAC - announced around a year ago, got it's forum placeholder yesterday which means it's 1-2 years away from release.

Funny you say that.

Quote:

Hello,

***** , we want to thank you for your past purchase of CoreAVC 2.0.0 .
Since you are a valued customer we are now offering you the ability to be one of the first to purchase the long awaited release of our CoreAAC Audio Decoder. The CoreAAC Directshow Plug-in for Windows Media Player is the culmination of over 2 years of work and fully supports AAC on any 32/64bit Windows OS.

We would like to pass on a $2.00 off discount code below ($5.95 purchase price). This discount code can be given to friends, family or maybe you just need CoreAAC 2.0 for another computer.

Discount code: *****
Purchase here: https://customers.corecodec.com/cart.php?a=add&pid=8

You can share and use the code above as many times as needed until August 30th, 2010.

Thank you.

Regards,

Dan Marlin
CEO/CoreCodec, Inc.
http://corecodec.com
http://twitter.com/corecodec

Lets hope that this is the start of many releases overdue.
I for one, am desperately waiting for coreplayer 2 for symbian.

BetaBoy · 6th July 2010, 10:11

Quote:

Originally Posted by ChronoCross

This may seem harsh but I've always been a stanch supporter of CoreCodec and it's been nothing but disappointment for awhile now and the last 20 pages of this thread is supporting evidence of that fact.

- We have been listening to the feedback here while we are working on DXVA support for CoreAVC 2.1, and have been reviewing the feedback/bug reports being posted.... we may not respond here to every D9 post but we do read what is being said and thank those for all the great feedback.

- While that is going on we just released CoreAAC 2.0 tonight, see: http://forum.corecodec.com/viewtopic...p=14872#p14872

- CoreASP 2.0 is under 'reconstruction' as we felt it was best to rework the code some more (since its like 3 codec code bases in one... and I am sure of you code junkies know what I mean about ASP)... this also includes ARM NEON (for our SDK Licensees).

- CorePlayer 2.0 is still a work in progress, but we are getting closer everyday now with just a few bullet points to go before we release it... including Android.

BetaBoy · 6th July 2010, 10:24

Quote:

Originally Posted by bob0r

LOL CoreAVC has this error for so long... (or in this case maybe haali) The same happens with BBC-HD. it always recovers and video is there to watch

Is this really an error? filters are not meant to seek backwards, as they only decode what they're given.

To be more precise... As far as "back by one GOP".... we can't "go back by one GOP" because the filter can't seek, it just receives data... ie; it shouldn't be showing frames until it is sure they're not incomplete.

Audionut · 6th July 2010, 10:46

Quote:

Originally Posted by BetaBoy

- While that is going on we just released CoreAAC 2.0 tonight.

Quote:

Originally Posted by BetaBoy 01 JUL 2010

We have officially launched CoreAAC 2.0 for you to purchase

http://forum.corecodec.com/viewtopic.php?f=30&t=3908

ChronoCross · 7th July 2010, 02:46

Quote:

Originally Posted by Audionut

Funny you say that.

Lets hope that this is the start of many releases overdue.
I for one, am desperately waiting for coreplayer 2 for symbian.

Indeed, pretty funny indeed. I've been waiting for CorePlayer 2 for Windows and Linux for quite some time now.

Quote:

Originally Posted by BetaBoy

- We have been listening to the feedback here while we are working on DXVA support for CoreAVC 2.1, and have been reviewing the feedback/bug reports being posted.... we may not respond here to every D9 post but we do read what is being said and thank those for all the great feedback.

Just do a quarterly release for product updates. You don't even have to respond on any forums. All any of us care about is results, but 6-12 month release cycles don't exactly inspire confidence.

Quote:

Originally Posted by BetaBoy

- While that is going on we just released CoreAAC 2.0 tonight, see: http://forum.corecodec.com/viewtopic...p=14872#p14872

Interesting timing.

Quote:

Originally Posted by BetaBoy

- CoreASP 2.0 is under 'reconstruction' as we felt it was best to rework the code some more (since its like 3 codec code bases in one... and I am sure of you code junkies know what I mean about ASP)... this also includes ARM NEON (for our SDK Licensees).

Duke Nukem Forever has been redeveloped 3 or 4 times. I got a good laugh when I read this one. Thanks.

Quote:

- CorePlayer 2.0 is still a work in progress, but we are getting closer everyday now with just a few bullet points to go before we release it... including Android.

Non-vague release date?

madshi · 7th July 2010, 09:23

@BetaBoy, two questions:

(1) Will CoreAVC get support for h264 MVC (3D) decoding?
(2) Will CoreAVC get support for h264 High 10 Profile (10bit) decoding?

Thanks!

BetaBoy · 7th July 2010, 12:35

CC..... Is D9 the reason you state we 'don't reply'? We do our best to reply although your not likely going to see a reply on every forum to every post.... But we do our best.

CorePlayer on Linux is a challenge for distributions as many of you may know. Many of our OEM customers have been using it on Linux for years now... But this is a controlled distribution and much easier for us to manage. We have discussed a public release for a while.... Let's see what we can do about it in CP2.

Madshi.... No that is CoreMVC... And something you'll likely see later this year. On 10bit.... It's been discussed and once we get past DXVA, we will go there.

madshi · 7th July 2010, 15:09

Quote:

Originally Posted by BetaBoy

Madshi.... No that is CoreMVC... And something you'll likely see later this year. On 10bit.... It's been discussed and once we get past DXVA, we will go there.

Good to hear, thanks. For 10bit, which FOURCC are you planning to use? If I may suggest, the FOURCC values on this page seem like good choices:

http://msdn.microsoft.com/en-us/libr...x#fourcc_codes

There are also FOURCC values for 4:2:2 and 4:4:4 listed there, in case you plan to support native 4:2:2 and 4:4:4 h264 streams, too. In any case it would be great, if you could offer an option with which your decoder always outputs the *native* format without any conversion. So basically for 8bit 4:2:0 you would use YV12 or NV12, for 10bit 4:2:0 you would switch to FOURCC P010, and for 10bit 4:2:2 you would automatically use P210 etc. That would be the ideal solution.

ChronoCross · 7th July 2010, 17:19

Quote:

Originally Posted by BetaBoy

CC..... Is D9 the reason you state we 'don't reply'? We do our best to reply although your not likely going to see a reply on every forum to every post.... But we do our best.

As I stated in my last reply I could care less about replies on any forum.....I'm more concerned when it takes six months to a year for any bugfixes on a product. I simply used doom9's 20-40 pages of bug/product discussion since your last release as an example of why you should release bugfix versions more often.

BetaBoy · 7th July 2010, 23:28

Quote:

Originally Posted by ChronoCross

As I stated in my last reply I could care less about replies on any forum.....I'm more concerned when it takes six months to a year for any bugfixes on a product. I simply used doom9's 20-40 pages of bug/product discussion since your last release as an example of why you should release bugfix versions more often.

What bugs are outstanding in CoreAVC 2.0 in the last 20-40 pages that we have not addressed??? If anything was a show stopper we would have released a new version, but as I've stated we have taken this time to concentrate on the 5+ flavors of DXVA 1/2 done as it is no small undertaking, and is one of the major features we wanted to get into it now before we move that code over to working on adding higher AVC profile support and then MVC.

BetaBoy · 7th July 2010, 23:37

Quote:

Originally Posted by madshi

Good to hear, thanks. For 10bit, which FOURCC are you planning to use? If I may suggest, the FOURCC values on this page seem like good choices:

http://msdn.microsoft.com/en-us/libr...x#fourcc_codes

There are also FOURCC values for 4:2:2 and 4:4:4 listed there, in case you plan to support native 4:2:2 and 4:4:4 h264 streams, too. In any case it would be great, if you could offer an option with which your decoder always outputs the *native* format without any conversion. So basically for 8bit 4:2:0 you would use YV12 or NV12, for 10bit 4:2:0 you would switch to FOURCC P010, and for 10bit 4:2:2 you would automatically use P210 etc. That would be the ideal solution.

Noted... Thx for the input.

ChronoCross · 8th July 2010, 00:36

Quote:

Originally Posted by BetaBoy

What bugs are outstanding in CoreAVC 2.0 in the last 20-40 pages that we have not addressed??? If anything was a show stopper we would have released a new version, but as I've stated we have taken this time to concentrate on the 5+ flavors of DXVA 1/2 done as it is no small undertaking, and is one of the major features we wanted to get into it now before we move that code over to working on adding higher AVC profile support and then MVC.

I wonder if you would say this to one of your corporate clients if they were complaining about these "non-showstoppers".

Anyway I'm tired of trying to convince you that you shouldn't release only major revisions once a year and that smaller minor bug fix releases would be helpful to customer satisfaction.

I've decided as a result of your reply to never purchase or recommend your products to anyone ever again. I know you won't find losing a single persons business all that harmful but oh well.

BetaBoy · 8th July 2010, 03:10

and answering with a statement rather then answering what bugs we have not addressed and or replied too helps who? As far as releases... We think we have been good in the quality of the last few 1.9.x and 2.0 releases and making them as bug free as possible. I attribute that to 1) Our devs doing QA, 2) The amazing job the beta test group does 3) Sites like here at D9 and, 4) The feedback we get from our OEM customers... and to that... you are right, if there were any major bugs in 2.0 as it stands right now we would be aware of it, and we are not.

26th June 2010, 17:43	#5882 \| Link
Guest Guest Join Date: Jan 2002 Posts: 21,901	Thanks. Seeking is fine with that sample in DGDecNV, so it appears that CoreAVC did not implement the "back by one GOP" move when decoding open GOP streams (or did not do it correctly). They could also skip leading B frames instead of using the "back by one GOP" move if accurate random access is not a requirement. Last edited by Guest; 26th June 2010 at 17:52.

7th July 2010, 12:35	#5894 \| Link
BetaBoy CoreCodec Founder Join Date: Oct 2001 Location: San Francisco Posts: 1,421	CC..... Is D9 the reason you state we 'don't reply'? We do our best to reply although your not likely going to see a reply on every forum to every post.... But we do our best. CorePlayer on Linux is a challenge for distributions as many of you may know. Many of our OEM customers have been using it on Linux for years now... But this is a controlled distribution and much easier for us to manage. We have discussed a public release for a while.... Let's see what we can do about it in CP2. Madshi.... No that is CoreMVC... And something you'll likely see later this year. On 10bit.... It's been discussed and once we get past DXVA, we will go there. __________________ Dan "BetaBoy" Marlin Ubiquitous Multimedia Technologies and Developer Tools http://corecodec.com

8th July 2010, 03:10	#5900 \| Link
BetaBoy CoreCodec Founder Join Date: Oct 2001 Location: San Francisco Posts: 1,421	and answering with a statement rather then answering what bugs we have not addressed and or replied too helps who? As far as releases... We think we have been good in the quality of the last few 1.9.x and 2.0 releases and making them as bug free as possible. I attribute that to 1) Our devs doing QA, 2) The amazing job the beta test group does 3) Sites like here at D9 and, 4) The feedback we get from our OEM customers... and to that... you are right, if there were any major bugs in 2.0 as it stands right now we would be aware of it, and we are not. __________________ Dan "BetaBoy" Marlin Ubiquitous Multimedia Technologies and Developer Tools http://corecodec.com Last edited by BetaBoy; 8th July 2010 at 07:15.

26th June 2010, 17:36	#5881 \| Link
me7 Registered User Join Date: Mar 2007 Posts: 217	Sure: http://www.mediafire.com/?sharekey=8...4ac78345cbe4ce I was able to reproduce the error in that sample by seeking back and forth.

30th June 2010, 11:51	#5884 \| Link
bob0r Pain and suffering Join Date: Jul 2002 Posts: 1,337	LOL CoreAVC has this error for so long... (or in this case maybe haali) The same happens with BBC-HD. it always recovers and video is there to watch

3rd July 2010, 04:24	#5885 \| Link
ChronoCross Does it really matter? Join Date: Jun 2004 Location: Chicago, IL Posts: 1,542	I've pretty much given up on all CoreCodec projects. CoreAVC was great when it came out, but with updates/bug fixes taking 7 or more months it becomes less and less competitive as open-sourced decoders performance are rapidly increasing with each passing week. CorePlayer 2 has been worked on for over 2 years now with the initial forum announcement stating this in August 2009. Now at nearly a year since that post CorePlayer 2 is no closer to being available. CoreASP - 2 or 3 years since it was announced. Treading on Duke Nukem Forever Territory. CoreAAC - announced around a year ago, got it's forum placeholder yesterday which means it's 1-2 years away from release. This may seem harsh but I've always been a stanch supporter of CoreCodec and it's been nothing but disappointment for awhile now and the last 20 pages of this thread is supporting evidence of that fact.

7th July 2010, 09:23	#5893 \| Link
madshi Registered Developer Join Date: Sep 2006 Posts: 9,140	@BetaBoy, two questions: (1) Will CoreAVC get support for h264 MVC (3D) decoding? (2) Will CoreAVC get support for h264 High 10 Profile (10bit) decoding? Thanks!