CoreCodec/H.264 Codec "CoreAVC"

[Astrophizz]

Looking forward to trying out a trial version to compare with DiAVC which is even faster than DivX on my system. If Dark's chart is any indicator then I'd have to agree with Cyber-Mav that it doesn't look good. Maybe I'll see CoreAVC's "larger advantage on older cpus" since I have a Core Duo (not Core2). Exciting stuff either way.

[ranpha]

Nice new website, the same as my website provider support system. Is that affiliate feature really work? 20% off purchase price is really generous. :p

[dead_screem]

so what's the deal? the website still has 1.9.5 and I havn't gotten an email. for the record the last update email I got was for 1.6. 1.6.5 and up I had to use the resend form every time, comcast seems to be blocking your bulk emails. So how do I get my 2.0 update 20% off email now?

Edit: I found the new core account system, (unfortunately it doesn't use our old core account email/passwords). It used my paypal email as the account email. (I had originally used a different email for the old core account then the one I payed with paypal) I tried logging in to the new system with my paypal email and my old core acount pasword and that didn't work. I tried the forgot password link and it said validation email sent but I got nothing. My paypal email was with aol. can't submit a support ticket because I would need to access my account to even do that... so now what?

[lnatan25]

A small question, will the h264 decoding bugs also find their way into the CorePlayer Movile platform?

[dvy]

CoreAVC is definite fastest H.264 decoder in planet earth.



Note "CoreAVC 2.0 Intel" is using ICC, so you can see the benefit of that.

Dark Shikari，why don't compile CoreAVC of windows version in icl,icl is wondows version intel C++ compiler.I compile x264 in icl is faster than x264 compile in gcc.

[hippoth]

What about a CoreAVC version for Mac OS X ... is there anything in the pipe?

[blubberbirne]

When buyers get the email with the download link???

EDIT: I can't login in customer area. My eMail don't work!

[zn]

any news on linux version? or android market?

[BetaBoy]

When buyers get the email with the download link???

EDIT: I can't login in customer area. My eMail don't work!

We are doing exactly as I stated emailing current customers with the account login first, then releasing it to them. So users will start to see those emails over the next few hours as they go continue to go out. Then we will open purchasing to the public.

[BetaBoy]

Nice new website, the same as my website provider support system. Is that affiliate feature really work? 20% off purchase price is really generous. :p
Correct, we have an affiliate program in place. However the rate will be 15% for 2.0.

Also on price. Upgrades are $4.95 and the full version is $9.95 (discounted from $12.95 till xmas).

[BetaBoy]

Looking forward to trying out a trial version to compare with DiAVC which is even faster than DivX on my system. If Dark's chart is any indicator then I'd have to agree with Cyber-Mav that it doesn't look good. Maybe I'll see CoreAVC's "larger advantage on older cpus" since I have a Core Duo (not Core2). Exciting stuff either way.
I did state this a while back in this thread that we will not offer a trial version of 2.0 for the first couple of weeks. The trial is scheduled to be released January 7th as CES begins as it is expected to be included on Media and some press junket's (NVIDIA, etc.)

[BetaBoy]

A small question, will the h264 decoding bugs also find their way into the CorePlayer Movile platform?Its the same thing really... so all the current fixes are already in CorePlayer 2.0. That also includes CUDA which we added a few months ago, but CorePlayer 2.0 will not be released in the next couple of weeks as we are still working on UI elements and trying to integrate our 'Corenect' UPNP/DLNA client/server into it for CES.

[BetaBoy]

so what's the deal? the website still has 1.9.5 and I havn't gotten an email. for the record the last update email I got was for 1.6. 1.6.5 and up I had to use the resend form every time, comcast seems to be blocking your bulk emails. So how do I get my 2.0 update 20% off email now?

Edit: I found the new core account system, (unfortunately it doesn't use our old core account email/passwords). It used my paypal email as the account email. (I had originally used a different email for the old core account then the one I payed with paypal) I tried logging in to the new system with my paypal email and my old core acount pasword and that didn't work. I tried the forgot password link and it said validation email sent but I got nothing. My paypal email was with aol. can't submit a support ticket because I would need to access my account to even do that... so now what?
Accounts get added into the system as the emails go out. So, no email=no account. Account emails are still going out.

[schweinsz]

yikes looks like divx is still damn quick. would be good to see how diavc compares with the above. i guess coreavc is only good for those who want cuda support, cant see people paying good money for a decoder thats hardly any quicker and in half the cases above slower than a free alternative. ??
From the results given by Dark Shikari, I believe the DiAVC alpha is still faster than the DiVX and coreavc2.0. I believe it is especially true when the DiAVC beta is out.

[BetaBoy]

schweinsz... no offense, but pls post results and not speculations and keep that talk in your thread, not this one.

[schweinsz]

schweinsz... no offense but pls post results and not speculations.
There are some results about coreavc1.9.5, divx and DiAVC.
http://forum.doom9.org/showthread.php?p=1331966#post1331966
http://forum.doom9.org/showthread.php?p=1334702#post1334702
http://forum.doom9.org/showthread.php?p=1344770#post1344770
http://forum.doom9.org/showthread.php?p=1347162#post1347162
http://forum.doom9.org/showthread.php?p=1347187#post1347187

I have not a coreavc 2.0, but allrough Dark Shikari gives that coreavc2.0 is comparible to divx, I believe it is slower than DiAVC.

I will give more results after you release the coreavc 2.0 to public between coreavc2.0 and the upcoming DiAVC beta.

[STaRGaZeR]

Yes, all the x86 asm since 1.9.5 has been written by me, including some from x264 that I wrote as well ;)

A bit has gone in the other direction too; x264's cacheline-split chroma MC was originally written for CoreAVC.

So we can forget about you RE'ing CoreAVC and improving libavcodec/ffmpeg-mt with that, right?

[squid_80]

I will give more results after you release the coreavc 2.0 to public between coreavc2.0 and the upcoming DiAVC beta.

How about you wait until your filter can output to multiple colorspaces, perform image levels adjustment, handle anamorphic widths properly and deinterlace. Also don't forget to provide the source files used for the benchmarks so they can be verified.

[edison]

there is a sync glitch when playing TrueHD clip with the 09-12-19 version MatroskaSplitter .

[Cyber-Mav]

From the results given by Dark Shikari, I believe the DiAVC alpha is still faster than the DiVX and coreavc2.0. I believe it is especially true when the DiAVC beta is out.

this is why im hesitant on making the purchase for coreavc 2.0
i hate to say it but its just too little way too late.

diavc beta will be the decider for me.

shikari could you rerun those speed test by locking affinity to 1 or 2 cores and see how performance racks up?

[tal.aloni]

there is a sync glitch when playing TrueHD clip with the 09-12-19 version MatroskaSplitter .

I can confirm that,
1. TrueHD have issues with both m2ts and mkv files, including the sample I posted @ page 256,
I can post more samples if needed.
Haali, is this part of your "core parser library"? I'm willing to help. (sent you an email)

2. I have a Blu-Ray demo disc that have two audio tracks: 5.1 DTS-HD and 2-channel LPCM,
I enjoyed 5.1 with the previous version,
but with the new version, I can only select the 2-channel LPCM audio track.

3. DD+ now works great. (starting with the new release).

Thanks,
Tal

[JohnnyFu]

Correct, we have an affiliate program in place. However the rate will be 15% for 2.0.

Also on price. Upgrades are $4.95 and the full version is $9.95 (discounted from $12.95 till after xmas).

Betaboy, is this the xmas edition you announced in 2008 ? Sorry, couldn't resist :D

Btw, I'd like to upgrade my lincense but I can't seem to find an upgrade link, huh?

http://corecodec.com/products/coreavc2

[LoRd_MuldeR]

It was already told that CoreAVC 1.x customers will receive an email with a unique upgrade code...

[JohnnyFu]

Aight', didn't read the forum lately, I just stumbled upon the Coreavc2 link while looking for a solution on how to play MKV's in Win7 MCE.

[BetaBoy]

Betaboy, is this the xmas edition you announced in 2008 ? Sorry, couldn't resist :D

Btw, I'd like to upgrade my lincense but I can't seem to find an upgrade link, huh?

http://corecodec.com/products/coreavc2

That link is not live yet.... it will be in about an hour as we send the emails out to current customers and take down CoreAVC.com to redirect them to CoreCodec.com

[BetaBoy]

There are some results about coreavc1.9.5, divx and DiAVC.

I will give more results after you release the coreavc 2.0 to public between coreavc2.0 and the upcoming DiAVC beta.

I'm really not gonna say it again seeing that you are insisting to taking your Off-Topic discussion to new levels. IE; Stop sh*tting on the release, before its even out, got it?

Take 'your' comparison out of this thread. Mods notified.

[weasel_]

Problem with cuda nad high number of ref frame is fixed ?

[BetaBoy]

In case you all did not notice..... Haali has posted the new splitter now that is included in CoreAVC 2.0: http://haali.su/mkv/

[edison]

In case you all did not notice..... Haali has posted the new splitter now that is included in CoreAVC 2.0: http://haali.su/mkv/

http://forum.doom9.org/showpost.php?p=1354688&postcount=5370 :)

[sneaker_ger]

One Question: Do you sell a full featured High Profile* Decoder this time? Or is it again possible that in 2 years x264 will use a feature previously not used and everyone has to buy a new licence?
(This of course excludes new features that are added to the specs after the purchase, as you don't need to promise on those.)

*normal High profile, not High10 or with 4:4:4 support

The 1.x was already sold as featuring "High profile (http://web.archive.org/web/20071222072516/www.coreavc.com/index.php?option=com_content&task=view&id=30&Itemid=1)".

[lexor]

@http://corecodec.com/products/coreavc2 I see this (emphasis mine):
Get Matroska MKV and MKA support from the ppl here at CoreCodec that helped create Matroska in the first place!

Was that intentional? Seems oddly out of place. Also I think it would be a good idea to expand MF into its full name (possibly with a link to details). I doubt many people who do not visit this forum would know what MF is and why bypassing it is important.

[BetaBoy]

Intentional yes, but only to my auto-correct.

[Disabled]

The 1.x was already sold as featuring "High profile (http://web.archive.org/web/20071222072516/www.coreavc.com/index.php?option=com_content&task=view&id=30&Itemid=1)".

Thats why I asked for full High Profile support, so they can't give me the finger once again. Now that DarkShikari teamed up with them, theyre gonna win every speed comparison and I have to buy the upgrade... ;-)

[BetaBoy]

The 1.x was already sold as featuring "High profile (http://web.archive.org/web/20071222072516/www.coreavc.com/index.php?option=com_content&task=view&id=30&Itemid=1)".
With features subject to change. We do plan on releasing 4:2:2 and 4:4:4 at some point... but as I stated, we have aligned our efforts with that of x264's capabilities.

[sneaker_ger]

With features subject to change. We do plan on releasing 4:2:2 and 4:4:4 at some point... but as I stated, we have aligned our efforts with that of x264's capabilities.

There's no mention of that on the website I linked. We all know that "smart" weighted p prediction was already supported by the standard standard at that time.

P.S.: Your website still charges 12.95 $ in case you haven't noticed yet. (Yes, I know your site isn't online yet)

Thats why I asked for full High Profile support, so they can't give me the finger once again. Now that DarkShikari teamed up with them, theyre gonna win every speed comparison and I have to buy the upgrade... ;-)

A decoder that isn't able to correctly decode all compliant High profile videos can't claim to feature High profile support IMHO. But of course I'd like to hear the answer to your question as well. Just wanted to make sure that when they claim "High profile" support it doesn't necessarily answer your question.

[Dark Shikari]

There's no mention of that on the website I linked. We all know that "smart" weighted p prediction was already supported by the standard standard at that time.

A decoder that isn't able to correctly decode all compliant High profile videos can't claim to support High profile IMHO. But of course I'd like to hear the answer to your question as well. Just wanted to make sure that when they claim "High profile" support it doesn't probably won't answer your question.libavcodec, DivX, Mainconcept, Elecard, and DiAVC all have known bugs or limitations in decoding some valid High Profile streams. It's not as if CoreAVC is the only one; hell, the only decoder I know that has no known issues at all is JM. And that's probably just because I don't use it enough to notice the bugs.

FYI, a list:

libavcodec: There is at least one weird conformance vector it fails on, relating to MMCO + long term refs.
DivX: incorrect deblocking with duplicate refs
Mainconcept: incorrect motion compensation with duplicate refs
Elecard: broken direct mode with duplicate refs, incorrect handling of delta quants > 25
DiAVC: quite a few (but it's alpha, so it never claimed to be perfect yet, of course)

[Disabled]

A decoder that isn't able to correctly decode all compliant High profile videos can't claim to feature High profile support IMHO.
I totally agree with you, but they obviously don't. But what can you do, try to sue them about it? Because of 9.99$?

libavcodec, DivX, Mainconcept, Elecard, and DiAVC all have known bugs or limitations in decoding some valid High Profile streams. It's not as if CoreAVC is the only one;
Youre probably right, but CoreAVC are the only ones who got money from me, so they are the only ones I can complain to. And they are the ones with the most obvious artifacts (ie unplayable).
I wouldn't care if the not supported features werent in use, but they are now, so now I do care. I just don't want to shell out money again to a company that promises things, they don't want to keep.

[sneaker_ger]

Don't get me wrong. I'm no angry customer complaining about not getting fixed weight p. But I don't like the marketing which tries to make it look like it's x264 fault that CoreAVC is buggy. On the recent CoreAVC page there was a URL listing its limitations (good thing!) - but it wasn't even linked (bad thing!). If I sell a decoder and claim that it features High profile support I have to be prepared to be taken up on that. (But to be realistically: CoreAVC is no 2000$ product - it only costs as much as few packs of cigarettes so customers should not be too demanding)
In this case I just wanted to make clear that giving a "marketing friendly" answer, like "CoreAVC 2.0 supports High profile" would not be sufficient, as they already had claimed that in the past.

[Cyber-Mav]

we have aligned our efforts with that of x264's capabilities.

so playing videos that are not encoded with x264 will result in a negative impact on speed?

this is very insteresting indeed.

[sneaker_ger]

so playing videos that are not encoded with x264 will result in a negative impact on speed?

this is very insteresting indeed.

We all know that impact on decoding speed depends on the settings and not the software used for encoding.

[Dark Shikari]

We all know that impact on decoding speed depends on the settings and not the software used for encoding.It's quite possible for a decoder to take speed shortcuts based on things it knows about the encoder. I don't know any that do, but it could be done.

[sneaker_ger]

It's quite possible for a decoder to take speed shortcuts based on things it knows about the encoder. I don't know any that do, but it could be done.

Thanks for the clarification. I didn't think in that direction at all. I love how you make at least 10 people a day look stupid. :(

[BetaBoy]

sneaker_ger... your points had been discussed 100+ pages back and we have taken the steps in notify what is and is not. In fact we have done this since the first day it was sold as well as changed when we were notified on issues.

[BetaBoy]

We are sending out emails now for users who purchased it within the past 60 days for the free update. Then come the update emails with the $4.95 upgrade discount for current customers.

[BetaBoy]

For the $9.95 launch use the coupon code: 20LAUNCH or just click this link: https://customers.corecodec.com/cart.php?a=add&pid=1&promocode=20launch

[the_corona]

That's great, if only we had a trial to see if it was worth it......or at least a few benchmarks by people with different proc architectures. I'm guessing nehalem owners arent exactly impressed so far (not that they should care much in the first place). What about Atoms/Athlon 64/Core Duo/Core 2 Duos?

I will not make a blind purchase.

[BetaBoy]

Since its finally live... I would like to thank the staff here for their hard work on the release, Haali for his simple solution for bypassing MediaFoundation in MKV's, Dark Shikari for his input, Clsid for his help with the installer... and even though you don't see it because its in our CoreAVC SDK... Thx to Mans for his kick ass work on ARM NEON Cortex A8 which makes CorePlayer rock on the iPhone 3GS.

[Disabled]

https://customers.corecodec.com/cart.php?a=add&pid=1&promocode=20launch
That link still states "- SMP (supports 4 CPU Cores)". Wasn't 2.0 to support more cores like DS made me think (http://forum.doom9.org/showthread.php?p=1354537#post1354537)?

[BetaBoy]

That's great, if only we had a trial to see if it was worth it......or at least a few benchmarks by people with different proc architectures. I'm guessing nehalem owners arent exactly impressed so far (not that they should care much in the first place). What about Atoms/Athlon 64/Core Duo/Core 2 Duos?

I will not make a blind purchase.

Based on feedback you will see across the board improvements in overall speed with all processors. Noting however that the Motion Vector and WeightP changes did slowdown performance about 1%... this is why we switched to ICC to get that 2% gain in overall speed.

[BetaBoy]

That link still states "- SMP (supports 4 CPU Cores)". Wasn't 2.0 to support more cores like DS made me think (http://forum.doom9.org/showthread.php?p=1354537#post1354537)?
We are still in the middle of changes.... that one is next.