View Full Version : Divx 6 Comparisons?
Bathrone
19th June 2005, 14:02
Now that its been out for a few days, I havent found any comparisons on the Net.
Some things I was considering was how the new Divx compared with:
1. XVID
2. x264
3. Nero Digital
4. The older 5.2.1 codec
Anyone got any results to share?
DigitAl56K
20th June 2005, 00:17
Sagittaire has run some tests and posted results here:
http://forum.doom9.org/showthread.php?p=675164#post675164
ChronoCross
20th June 2005, 01:16
I don't really trust metrics tests. I mean it's really the only way to give acutal evidence about a codecs superiority but alot of times even with better metrics tests the encode can still look worse than another.
708145
20th June 2005, 01:52
I don't really trust metrics tests. I mean it's really the only way to give acutal evidence about a codecs superiority but alot of times even with better metrics tests the encode can still look worse than another.
I second that. While metrics give a first hint at how good a codec is only your eyes will tell you the truth :)
Please trust your eyes :D
And now that I post in a DivX thread already I want to congratulate DivX Corp. for their latest codec release. It sure took some time. :rolleyes:
bis besser,
Tobias
DigitAl56K
20th June 2005, 02:01
And now that I post in a DivX thread already I want to congratulate DivX Corp. for their latest codec release. It sure took some time. :rolleyes:
Ah, but the codec has been ready for some time! In fact, we're already making big progress on the next one!
Jump over to DivX Labs (http://www.divxlabs.com) to get yourself a copy of Helium. Just remember it's experimental, and may exhibit some issues ;)
Bathrone
20th June 2005, 03:12
Thanks all
Some info http://forum.doom9.org/showthread.php?t=90784
Im not totally clear on all this metrics. So, basically of all MPEG IV ASP codecs Divx 6 wins, but with the AVC crowd x264 wins all.
Looking at the stats there doesnt seem to be much difference, even divx against x264...or is there? How much are the points worth - stuff all or alot?
DigitAl56K
20th June 2005, 03:39
Looking at the stats there doesnt seem to be much difference, even divx against x264...or is there? How much are the points worth - stuff all or alot?
There isn't much difference. Of course, it all depends on the test you run. Sometimes x264 will win out against DivX, sometimes the opposite will be true. But even when one exceeds the other, they're usually very close. When you apply H.264/AVC to future real-world use cases, such as HD video, it usually isn't as much of an improvement as the press machine would suggest.
Yup.. one big machine with a few screws loose ;)
One story we continually try to tell, is that when you have a very good video technology with an entire format around it, supported in millions of devices, it doesn't necessarily matter that there is a video codec out there with a tiny bit of extra compression, especially when it's decode complexity is higher.
It's a hard story to tell to technology purists because they don't necessarily think in terms of the big picture, but in terms of whatever is the absolute best in one particular aspect at the time. Nevertheless, as I have said in many other threads, it's not the individual pieces on their own that count - and sure it helps that the individual pieces on their own are very good, but it's more how they come together in a format and what that format enables that is truely valuable to most people.
Bathrone
20th June 2005, 03:49
Yes, thats a very sound point.
I had alot of hassle with x264 - installing, transcoding, decoding for playback. Its not supported on my XBMC. Its not supported on my DVD.
With Divx, even the new Divx I can play on XBMC and my DVD.
When the tools eventually come out like a divx enabled dvd shrink app Ive been talking about - the whole package becomes easy to use and easy to deploy.
IgorC
20th June 2005, 05:37
Future standarts Blue Ray, HDDVD were already established. Future standart everywhere supported codecs are H.264, WMV9, MPEG2. Itīs just a matter of time. Already during this year.
DigitAl56K
20th June 2005, 06:55
That's great, but how many of us will have them this year? So far it's all pie in the sky. What percentage of people worldwide will have a blu-ray enabled device this year? I don't have a blu-ray device. Nobody at DivX has a blu-ray device. What store can I go to and buy one? Is there any content available for these devices? If my content is in MPEG-2 (and you know it will be for some time because all the professional authoring tools are MPEG-2 and this is what the studios are comfortable with) then you still have to deal with massive file sizes. You still have half the majors lined up on blu-ray, and half lined up on HD DVD.
The list of problems goes on. Don't expect any of this to become mainstream this year.
Anyway, this thread is on codec comparisons and we're going a little off-topic.
Has anyone else run some tests? :)
HalfHuman
20th June 2005, 08:31
Still haven't seen a codec comparison with the following
1. XVID
2. x264
3. Nero Digital
4. The older 5.2.1 codec
Anyone got any results to share?
i mean a codec comparison made by an independent party with images and possibly movie clips. i have seen one on THG, but they seem so in love with DivX. and this is one reason i don't trust their results. in their oppinion XviD is "good" quality compared to mighty DivX.
and yes you are quite a little bit off topic.
bond
20th June 2005, 09:24
Im not totally clear on all this metrics. So, basically of all MPEG IV ASP codecs Divx 6 wins, but with the AVC crowd x264 wins all
note that this comparison is of course only valid for the clip used in that comparison (harry potter 2 trailer), and imho a fast motion movie trailer isnt representative for a normal movie...
Doom9
20th June 2005, 09:27
i have seen one on THGDon't trust a site that doesn't respond to input on their video encoding techniques by somebeody who runs a video compression site... enough said.
The last codec comparison took 1.5 weeks of holidays.. how many of you are willing to spend that every time a new codec comes out? No? No again? And you there in the corner no as well? I thought so. Just so that you realize a properly done comparison is a huge, unpaid effort.. that's why you see so few of them. And then there's the chance that by the time your article is out, there's a new codec revision that fixes a particilar issue you've had, etc.
For now, the best recommendation has already been given: Please trust your eyes Compare those codecs you're interested in, and use whichever floats your boat.
Inventive Software
20th June 2005, 09:48
Don't trust a site that doesn't respond to input on their video encoding techniques by somebeody who runs a video compression site... enough said.
The last codec comparison took 1.5 weeks of holidays.. how many of you are willing to spend that every time a new codec comes out? No? No again? And you there in the corner no as well? I thought so. Just so that you realize a properly done comparison is a huge, unpaid effort.. that's why you see so few of them. And then there's the chance that by the time your article is out, there's a new codec revision that fixes a particilar issue you've had, etc.
For now, the best recommendation has already been given: Compare those codecs you're interested in, and use whichever floats your boat.
I second the time it takes to do one of these comparisons. The time planning the thing should be factored in as well (mine took over 4 months), and allowances made for updated codecs days, maybe hours, before you do the tests. I've been doing these tests for the last 4 weeks, and it's still only 1/4 finished. Currently I'm in 6th form, so I can set the codec on it's way, go to school, come back, see how long it's got left roughly and wait for the time. Because the codec takes so long to process (average encoding speed is between 6 and 7 FPS), I know I can be gone for about 4 hours, come back and know the codec will still be going.
My philosophy is to properly test the codec beforehand to iron out any problems. Rigourously test the options you plan to use, if they don't work, go to the next codec, test those options. If you have to find a workaround, do so, but notify the devs. That's why they rely on the input. A practical example is me telling On2 about the 2-pass encoding.
Codec tests take a while, especially if there's lots of codecs to test. I'll admit, this one's taking a long while to do, but the next one may take even longer (more codecs to test) or less (less encoding time) or about the same! I'm getting a new machine to do the legwork, so less encoding times for me are on the cards!
708145
20th June 2005, 11:15
Jump over to DivX Labs (http://www.divxlabs.com) to get yourself a copy of Helium. Just remember it's experimental, and may exhibit some issues ;)
Well, but since the codec doesn't come with a CLI I'm not as interested as I'd otherwise would be. But no worries, I'll test it sometime but definitely not anytime soon. Way too busy on other things.
Any chance to provide Linux64 and Windows64 CLI encoders for use in scripts? You should have tools like that for development anyways :p
bis besser,
Tobias
vBulletin® v3.8.11, Copyright ©2000-2026, vBulletin Solutions Inc.