Looks like Divx 5.1 is Released!

[rozemab]

go to

www.divx.com

[Digga]

whhooha, fiiiinaly!!!
download: http://www.divx.com/divx/
update pro-version: http://www.divx.com/divx/pro/upgrade.php

[danwatt]

The changes look nice. The encoder definately seems a bit slower (about 50-70% of the speed I used to get in 505 on "slowest" setting, right now I am testing "standard", "slow" and "slowest" in multipass mode). The feedback window looks nice, but, beware, it seems to take a big performance hit (my rendering rate drops 2 FPS or so in VDub when I have the graph and picture turned on).

Other than that, great work DIVX team!

[zag2me]

anyone got any mirrors? Divx.com is down, propabaly due to demand.

[Ruppl]

I cant find the Options GMC, Quarter Pixel and Bidirectional Encoding.

[Lord_KiRon]

Originally posted by danwatt
The changes look nice. The encoder definately seems a bit slower (about 50-70% of the speed I used to get in 505 on "slowest" setting, right now I am testing "standard", "slow" and "slowest" in multipass mode). The feedback window looks nice, but, beware, it seems to take a big performance hit (my rendering rate drops 2 FPS or so in VDub when I have the graph and picture turned on).

Other than that, great work DIVX team!

Are you sure it's slower ? 50%-70% is LESS then 100% so it usually means faster :)

Anyway, they claimed it 10% faster then 5.0.5 on most CPUs , so how it can be ?

Btw: if you got a chance to d/l it please please please upload to :

http://www.gooddvdstuff.com/files/ ...

[Tester]

6 mirrors at chip.de

good luck


http://www.chip.de/downloads/c_downloads_getfile_8943765.html?t=1062599024&v=3600&

[DigitAl56K]

"The feedback window looks nice, but, beware, it seems to take a big performance hit"

If you are using the Overlay window to watch for feedback you should be able to set the video mode to "Fast Recompress" unless you are applying filters in VDub, which should compensate for the drop.

"I cant find the Options GMC, Quarter Pixel and Bidirectional Encoding."

Click "Select Profile Wizard" to access these options. Note you must check "Disable profiles" on the first page of the Wizard to enable GMC/QPel on the second page.

@danwatt: Standard mode in 5.1 is approximately equivelent in functionality to Slowest mode in 5.05. Using Slow/Slowest mode in 5.1 introduces to varying degrees the new RD algorithm, which is much slower but produces much superior quality. You can normally get away with using Standard mode for all but the last pass in multipass, and then switching to Slowest mode. Incidentally, the MV file is now functional once more, so that after the first pass you should see quite a dramatic pickup in performance.

Update: Jazz just brought the official server back up

[Candock]

the divx playa is already included.
i just click it away at installation :rolleyes:

[Acaila]

@danwatt: Standard mode in 5.1 is approximately equivelent in functionality to Slowest mode in 5.05. Using Slow/Slowest mode in 5.1 introduces to varying degrees the new RD algorithm, which is much slower but produces much superior quality. You can normally get away with using Standard mode for all but the last pass in multipass, and then switching to Slowest mode. Incidentally, the MV file is now functional once more, so that after the first pass you should see quite a dramatic pickup in performance.One word of caution for other people, if you use enable the use of the MV file during the last pass the codec will use whatever performance setting the previous pass used (because it copies the motion vectors from the file).

e.g.
1st through 3rd pass at standard mode without MV
4th pass at slowest mode without MV

Result: no problem, perfect video

1st through 3rd pass at standard mode with MV
4th pass at slowest mode with MV

Result: 4th pass will be encoded at standard mode, not slowest

In short, you can enable MV files during all passes only if you use the exact same performance setting for all passes. If you intend to use slowest for the last pass, turn MV file off during that pass (but you can still use it to speed up the other passes).

[Sumster]

Hi... did anybody notice a difference with the decoder of Divx 5.1 vs. Kaukura? I have an Athlon 1800+, and I could use full post-processing (ie. both at MAX) before with only about 30-50% CPU usage. With Divx 5.1, I'm at 100% with ultra-choppy playback... not sure if there is some other issue at play. Even with automatic post-processing at full, it uses 80-90% with choppiness.

-Sumster

[LordIntruder]

I made a very fast test and I notice no difference between Kaukura and 5.1 speed talking.

For those that are still with Kaukura and older betas, in 5.1 there is no longer "old pass 2 mode".

BTW I like thoses cows being caught in the tractor beam. ;)

Well for thoses who want more info about 5.1, you should definitely read older topics on beta version, especially the last one on Kaukura:

http://forum.doom9.org/showthread.php?s=&threadid=59149

Kaurura is close to 5.1 and you'll find some nice tips.

I was waiting for the 5.1 final to ask that question. In Kaurura Divx team has introduced a new psychovisual mode that can works on Chroma operation (http://www.divx.com/divx/versions.php). It is still here in 5.1 so I would like to know what are we suppose to get by turning this option on? I did test it with Kaukura but the file that is needed to perform that task wasn't written!?!

I would like some feedback about people that used this option.

Thanks :)

[SeeMoreDigital]

Well, I don't know about any of you guy's but I think it would have been nice if DivX (ie Gej) had posted a message on the forum, informing us of this new version.

I just feel a little used and abused is all.

Cheers

[DigitAl56K]

Gej hasn't even posted over at DivX.com to announce the new version yet ;) Actually I hear he's taking a few days off, so please don't be offended. With all the work DARC has been putting into the 5.1 release I'm sure he's sick of the sight of computers!

[Darrius "Junto" Thompson]

Originally posted by DigitAl56K
Gej hasn't even posted over at DivX.com to announce the new version yet ;) Actually I hear he's taking a few days off, so please don't be offended. With all the work DARC has been putting into the 5.1 release I'm sure he's sick of the sight of computers!

Yepp Gej is on work travel and vacation. You can blame me for not being first to announce this release although even if we tried to someone in the video world beats us to it quite often. And, I just got off vacation yesterday. We're all taking turns over here taking a little time off. All of our Brains are fried over here. Most of us haven't taken a break in a long long time. I'll ask those with brain power still left over here to troll the boards and we'll answer as we can. I'd consider DigitAl56K a part of the official DivX Team so we'll just leave it to him to answer everything :)

Darrius "Junto" Thompson
Director R&D DivX

[LordIntruder]

I'd consider DigitAl56K a part of the official DivX Team so we'll just leave it to him to answer everything

LOL!!

So DigitAl56K I would like you to answer why in Divx 5.1 we have.....

:D :D

[danwatt]

So, let me just make sure I understand this. Usually I do four passes (at least in 5.0.x, that seemed to be the sweet spot). I should use MV for passes 1-3 in Standard, then NO MV in pass 4 in slow? Doing standard in the first 3 and then slow in the 4th will be the same as if I did slow for all 4 passes?

Oh, one other silly question. Do I want to "update log file" during passes 2-3? I never really cared to look into what that did back in 5.0.x.

[kl33per]

Originally posted by danwatt
So, let me just make sure I understand this. Usually I do four passes (at least in 5.0.x, that seemed to be the sweet spot). I should use MV for passes 1-3 in Standard, then NO MV in pass 4 in slow? Doing standard in the first 3 and then slow in the 4th will be the same as if I did slow for all 4 passes?

Oh, one other silly question. Do I want to "update log file" during passes 2-3? I never really cared to look into what that did back in 5.0.x.
1. Using MV for 1-3 and then Slow/no MV for 4 is not the same as Slow/Slowest for all 4. However, it's pretty close and definately saves time (i.e. there would be a very slight quality increase by doing all four passes in Slow/Slowest mode).

2. You definately want to check the "update Log File' check box during nth passes, or the passes will be pointless (i.e. they'll do nothing).

[Candock]

Originally posted by SeeMoreDigital
Well, I don't know about any of you guy's but I think it would have been nice if DivX (ie Gej) had posted a message on the forum, informing us of this new version.

I just feel a little used and abused is all.

Cheers

Gej said that on german television, that the 5.1 will appear next day!
This was Tuesday!

[JensG.]

Originally posted by danwatt
[B]So, let me just make sure I understand this. Usually I do four passes (at least in 5.0.x, that seemed to be the sweet spot). I should use MV for passes 1-3 in Standard, then NO MV in pass 4 in slow? Doing standard in the first 3 and then slow in the 4th will be the same as if I did slow for all 4 passes?
Acaila said that using slower settings at last pass doesn't make sense if you used mv file in the passes before.


------
edit: I am wrong.

[Wuntvor]

Originally posted by JensG.
Acaila said that using slower settings at last pass doesn't make sense if you used mv file in the passes before.
Actually he said that if you use the fourth pass WITH mv file its going to use the same settigns as the earlier pass.

So using 1-3 passes with mv file and then a fourth with slowest without the mv file is ok.

Atleast thats how i read it :-)

regards
/Wuntvor

[Babeinternational]

@Wuntvor: Yeah, you're right.

@DigitAl56k: Teacher's pet...;-)

[JensG.]

Damn, I had to read again!
In short, you can enable MV files during all passes only if you use the exact same performance setting for all passes. If you intend to use slowest for the last pass, turn MV file off during that pass (but you can still use it to speed up the other passes).
Right, the last slower pass has to calculate mv data on its own, or it will reuse the one from faster passes before.

[SeeMoreDigital]

Hi guy's

I have not had chance to test DivX 5.1 yet!

I am interested to know whether or not this version has broken through the 'DivX versus WMV9 950kbps' barrier.

Have any of you tried creating, full frame 720x480/576 DivX and WMV9 encodes at say 850kbps and compared them against each other!

Cheers

[Twilight_Fandom]

so..divx 5.1 will give choppy playback on a slower pc? like a p3 800?
i dont really care,cause i got a fast pc,but i want my vids to play on slower ones too.

Overally,is it worth changing to 5.1?what do u guys think?

[DigitAl56K]

-Edit-
I didn't notice page 2, some of this is a bit redundant, oops ;)

P.S. @BabeInt - Teachers pet wanna be! lol ;)

-Original post-

Originally posted by Darrius "Junto" Thompson
I'd consider DigitAl56K a part of the official DivX Team so we'll just leave it to him to answer everything :)


Two things:

Thanks Junto!

and

...thanks Junto.... :S

lol ;)

Originally posted by danwatt
I should use MV for passes 1-3 in Standard, then NO MV in pass 4 in slow?

Correct, this will give you quite a performance boost. You can use the MV file so long as you keep all of your settings consistent.


Doing standard in the first 3 and then slow in the 4th will be the same as if I did slow for all 4 passes?


It won't be exactly the same, but to give you some indication when I tested doing 2-Pass with Standard+Slowest against Slowest+Slowest with Manihi the PSNR difference was typically less than 0.25db. Naturally further testing is warranted with the 5.1 release build, and it will be interesting to read everyones results on this :)

Originally posted by JensG.
Acaila said that using slower settings at last pass doesn't make sense if you used mv file in the passes before.


Yes, this is why you must disable the MV file once you change to slowest mode.

[DigitAl56K]

Originally posted by SeeMoreDigital
I am interested to know whether or not this version has broken through the 'DivX versus WMV9 950kbps' barrier.


Be aware that when you compare these two codecs DARC has put a truely exceptional effort into achieving these quality enhancements while staying within the MPEG4 spec. On the other hand, WM9 AFAIK breaks spec and won't play on any of the MPEG4 enabled/DivX Certified hardware players that are now beggining to flood the market.

Consider this when you come to encode your videos, as the quality difference really is very marginal.

Example:
DivX running under Home Theatre profile will tend to equal to within a very small margin (and sometimes exceed) the PSNR score of WM9 running under its full Complex decoder profile. The resulting DivX can be played on any of the current breed of MPEG4 hardware players. The resulting WM9 couldn't even be played on a WM9 hardware player (according to the MS WM9 codec documentation) because the decoder profile is too high (Complex profile exceeds both regular and high-def WM9 profiles and is only suitable for decoding on a PC).

I feel sorry for anyone who will have to run WM9->DivX5.1 conversions in the future :)

[DigitAl56K]

Originally posted by Twilight_Fandom
so..divx 5.1 will give choppy playback on a slower pc? like a p3 800?
i dont really care,cause i got a fast pc,but i want my vids to play on slower ones too.

Overally,is it worth changing to 5.1?what do u guys think?

Actually, the decoder included in version 5.1 is up to ~30% faster than the 5.05 decoder (which stands it firmly in-line with FFDShow on performance). Additionally, it now includes a feature that automatically manages the postprocessing/quality levels in real-time based upon decoding performance, so that your system will always decode at the maximum possible quality without dropping frames.

Aside from this, 5.1 also features a new DirectShow AVI filter that can play back partially completed/damaged AVI files.

All in all, even if you only use DivX for playback 5.1 makes for a major upgrade over 5.05.

[SeeMoreDigital]

Hi DigitAl56K

Welcome 'back' to the forum! Be aware that when you compare these two codecs DARC has put a truely exceptional effort into achieving these quality enhancements while staying within the MPEG4 spec. On the other hand, WM9 AFAIK breaks spec and won't play on any of the MPEG4 enabled/DivX Certified hardware players that are now beggining to flood the market. Actually I was not talking about the M$ mpeg4 WM9 codec I was talking about the WMV9 codec.

So, I feel my question still stands. As a full frame 720x480/576 WMV9 .avi encode does look better than DivX 5.0.5 (and the beta's) at 950kbps or below - Try some comparisions at around 615kbps. Play them both back using the DivX player 2.1 (or new 2.5) and see for yourself!

Cheers

EDIT: I'll mail you some encodes on a CD-R if you like!

[DigitAl56K]

Is there an MPEG4 WMV9 codec? I have seen only Windows Media Video 9 and Windows Media Video 9 Professional.

P.S. 'thanks' ;)

[SeeMoreDigital]

Originally posted by DigitAl56K
Is there an MPEG4 WMV9 codec? I have seen only Windows Media Video 9 and Windows Media Video 9 Professional.

P.S. 'thanks' ;) Well I'm confused more than ever now! I'm talking about WMV9 (VCM). The application that creates, WMV9 .avi files. I thought you were talking about Microsoft's early Mpeg4 codec(s)!

As I said, I'll mail you some 'full movie' encodes on CD-R if you like. So you can compare!

Cheers

[temporance]

So, I feel my question still stands. As a full frame 720x480/576 WMV9 .avi encode does look better than DivX 5.0.5 (and the beta's) at 950kbps or below - Try some comparisions at around 615kbps. Play them both back using the DivX player 2.1 (or new 2.5) and see for yourself!Perhaps 720x480 isn't the best resolution for DivX. I could believe that WMV9 uses a process to reduce the amount of data in the image, instead of resorting to stronger quantization (and artifacts). That process could be a simple resize (Would you know if WMV9 had used a high quality internal resize to 512x480?) or could be more complicated preprocessing or transform domain techniques.

[SeeMoreDigital]

Originally posted by temporance
Perhaps 720x480 isn't the best resolution for DivX. I could believe that WMV9 uses a process to reduce the amount of data in the image, instead of resorting to stronger quantization (and artifacts). That process could be a simple resize (Would you know if WMV9 had used a high quality internal resize to 512x480?) or could be more complicated preprocessing or transform domain techniques. The full movie encode test is infact a PAL DVD rip of StarWars 2 (136min). The output image pixel size for both DivX 5.0.5 and WMV9 VCM is infact 720x576. The WMV9 VCM encode was generated using the slowest method possible and took around 30 hours to complete. The DivX 5.0.5 encode was generated using 'standard' setting with 'light psy' and 'bi-di' and took around 6 hours to complete.

Without doubt the WMV9 VCM encode looks better than the DivX 5.0.5 encode. Even generating encodes using the DivX beta's produced no better results!

I have also generated XviD, WMV9.wmv and RV9 encodes from the same source. All have audio at streams at 64kbps! However, here the RV9 encode looks better as a whole but... like alot of RV9 encodes, something appears to be missing and for some reason or another, the image does'nt seem to be as engaging!

Well, that's what I've been able to come up with!

Cheers

EDIT: I've even created DivX encodes of the same source (again at 674kbps) using 1024x576 pixels. These look very good on a PC monitor or even 42" plasma. But the Xcard can't play them as it does not support frame sizes above 720x576

[Sumster]

Just another question regarding post-processing....

On my Athlon 1800+ with 512mb of RAM running a high bitrate 640x480 DIVX5 clip:

No post-processing: 28-36% CPU
Min deblocking, Min deringing: 29-40% CPU
Full deblocking, min deringing: 48-58% CPU
Full deblocking, medium deringing: 100% CPU
Min deblocking, Medium deringing: 100% CPU
Min deblocking, Max deringing: 100% CPU


Therefore, if I set deringing to anything other than MIN, I quickly max out an Athlon 1800+ using DIVX 5.1. In all previous versions, I was able to use max deblocking, max deringing with no issues and around 50-60% CPU usage (including Kaukara). Was deringing disabled on previous versions? Has it been enhanced in DIVX 5.1?

Of course, playback is fine using partial and full modes. Although, if I configure the codec to FULL AUTO, the setting does not seem to stick.

Thanks...

Sumster

[LordIntruder]

SeeMoreDigital:

You made tests on Kaurura and you wasn't satisfied. You can assume it will be the same on 5.1. This one is just Kaurura bug free. Maybe there is a little speed improvement? Maybe there is a little video quality improvement? Yes but if you were not satisfied with Kaurura vs WMV9, you won't be either here too.

Also I'm like DigitAl56K, I didn't know there were different WM9 codecs. I always heard about WMV9 from Microsoft and you tell there is another one, a different one? I should be interested to know more. I'm a bit at a loss. :confused:

Anyway, here are some tests I made with 5.1:

I encoded a 10min excerpt movie (15000 frames), PAL 4:3, B&W, bitrate: 800K. I use a 1800+ with 1Gb RAM and VDM in its latest release (1.5.4.1).

Important: the 10min are slow-motion, people talking, walking quietly indoors or outdoors. I didn't make any fast-motion test till now.

Test A:
-------

1st Pass: Standard, BFrames, GMC, PSY Slow, NO MV --> 47min
2nd Pass: Slowest, BFrames, GMC, PSY Slow, NO MV --> 176min
------
Total: 223min

Test B:
-------

1st Pass: Slowest, BFrames, GMC, PSY Slow, MV --> 174min
2nd Pass: Slowest, BFrames, GMC, PSY Slow, MV --> 121min
------
Total: 295min

Test C:
-------

1st Pass: Standard, BFrames, GMC, PSY Slow, MV --> 36min
2nd Pass: Standard, BFrames, GMC, PSY Slow, MV --> 30min
------
Total: 66min


For those who wouldn't still know it, 5.1 Standard equals or even beats 5.05 Slowest. However if you want to get the most prepare yourself to wait a long long time but the final result worth it.

Test C is the fastest and quite frankly the result is very good. 3 times shorter than test A. There are some noticeable differences between test C and tests A & B, a lowered quality on some areas but it's quite acceptable in regard to the encoding time. Also when you compare 5.1 Standard against 5.05 Slowest you will notice much less macroblocks. I think this 5.1 Standard is going to become beloved from rippers. ;)

Now test A & B are very hard to differentiate but you can still see a sharper image on test B with a small but noticeable quality increase on some specific scenes or/and areas. So yes when you watch both tests A & B at a correct distance it will be hard to tell which one is the best. So if you want to get a very very good video quality, try test A and if you want a even better one: test B.

But remember tests A & B are very very closed.

Now those results need to be confirmed in fast-motion scenes. One thing is sure, with 5.1 combined with slow and even more slowest mode, macroblocks are gone or almost where you got several with 5.05.

One thing remains to me to be tested: make test B WITHOUT MV use to see if this option that lowers the encoding time would mean a lowered quality too, perhaps almost unoticeable but some tests have to be done on this area. I hope to do this job (test B without MV) tonight.

Stay tune! :D

[SeeMoreDigital]

Hi LordIntruder

Also I'm like DigitAl56K, I didn't know there were different WM9 codecs. I always heard about WMV9 from Microsoft and you tell there is another one, a different one? I should be interested to know more. I'm a bit at a loss. No you're right, there is only one WMV9 VCM codec (two if you include the beta.

However, if you have Windows Media Encoder 9 installed on your PC you do have the option of saving video streams as one of Microsoft's Mpeg4 codecs (as well as WMV9.wmv streams).

This is what I thought DigitAl56K was talking about!

Cheers

[LordIntruder]

Hello SeeMoreDigital.

Thanks for the confirmation. :)

[SeeMoreDigital]

No problemo!

The old brain is wondering today!

I decided to take a day off work and create my own home web site (more like 'web shyte' at the moment) using Apache HTTP Server 2.0.47.

Just waiting now for SeeMoreDigital.net to get registered and I'll be able to stick a few 'zipped' test files on it for download.

Cheers

[midi]

WM9 AFAIK breaks spec and won't play on any of the MPEG4 enabled/DivX Certified hardware playersSpeaking for myself and everyone I know personally (college crowd, I guess), none of us care in the slightest about hardware players. A computer is pretty much a necessity in today's world, a hardware MPEG4 player is a rarity.

I would much prefer DivX Networks worked towards making the best computer codec possible rather than worry about this. Monitors are in general much higher quality than TV, but most decent video cards and laptops come with video out anyway if you really need to display on one.

[LordIntruder]

You should revise your position even if you speak for yourself ;)

It is a great pace to see that hardware players are going to be released more and more and that divx network works hard on this job, thus people could watch their divx quietly on their sofas which is much more pleasant than look at a monitor screen don't you think?

You say you can connect your pc to your tv, not everybody can do it, not everybody have a graphic card with tv output, not everybody have the pc close to the tv, etc...

And those that don't have any pc, if they have friends that have divx content that break mpeg4 specs, if those ones bring their CD, no one could be read on their hardware player.

Also I really wonder if staying under mpeg4 specification avoid to have a much better codec. If it was so simple to have a very strong codec, I'm sure others people would have already break mpeg4 specs for a long long time and we would have the choice to use another codec that would kill every ones on the market right now.

[SeeMoreDigital]

I totally agree with you LordIntruder!

Posted by midi
Speaking for myself and everyone I know personally (college crowd, I guess), none of us care in the slightest about hardware players. A computer is pretty much a necessity in today's world, a hardware MPEG4 player is a rarity. There will be nothing rare about Mpeg4 in a year or so's time. Many chip manufacturers are combining Mpeg4 onto their Mpeg1/2 chipsets. Some of these chipsets are already available on 'stand alone' players.

It's just a matter of time before the 'big boy' manufacturers follow suit, just as they had to do with Mp3 audio. I remember somebody at Pioneer saying "nobody want's to listen to Mp3 audio on one of our players when they can have DVD audio". Well they 'all' certainly changed their minds about the format pretty fast!

My guess is that the new chipsets will have quite a lot of capabilities. But not all of them will be available to the consumer at first!

Let's have more stand alone's. Maybe we should start a poll!

Cheers

[DigitAl56K]

MPEG4 capable/DivX Cerified hardware players are really taking off now in a big way, you can probably pick one up for ~$100. Who is going to buy a regular DVD player without DivX support in a years time? KiSS is even planning to release soon such a player with video recording support.

Here's something to think about, if VCD conversions are so popular, why wouldn't people want hardware players when they can get much better resolution, quality, and all on one CD with DivX? :)

[valipod]

Did anyone make a test using QPel at slowest? Accidentaly I did one encode and surprisingly it came a LOT faster than without QPel :eek:

This only happens using slowest, things are the other way around when using "standard", there QPel makes things slower.

Somebody in another forum said, QPel is automatically disabled when using "slowest", but I must disagree :devil:, first because of the speed difference, and second, because of the quality difference, QPel makes things noisier.

Here is the test data:

1000 frames
1st pass, slowest, No MV, Fast Psy, GMC, BIDI and QPel: 180 seconds
2nd pass, slowest, No MV, Fast Psy, GMC, BIDI and QPel: 184 seconds

1st pass, slowest, No MV, Fast Psy, GMC, BIDI NO QPel: 495 seconds!!!
2nd pass, slowest, No MV, Fast Psy, GMC, BIDI NO QPel: 520 seconds!!!

Can anyone explain this? Gej?

[Acaila]

My guess it's a bug, and that the performance setting "slowest" gets turned into "standard" when you use QPel at "slowest" mode.

Care to run another encode at "standard" without QPel and see if it's equally fast as your "slowest" with QPel?

[dare]

Regarding HW players:

they are definately on the way to spread all over the market. And since Phillips is launching such a player with 3.11 support it is soon to be considered a requirement to be able to sell a DVD player at all.

If taking this in consideration QPel is alos not to be used since no HW player supports this.

My 2c

/dare

[DvdKhl]

I didn't test the new DivX jet but it has still the same bug:
if you open the codec settings and hit ok... after that everything like post processing is set back to defaults

Or is it just me?

[valipod]

Originally posted by Acaila
Care to run another encode at "standard" without QPel and see if it's equally fast as your "slowest" with QPel?

Well I did that and:

1st pass, standard, No MV, Fast Psy, GMC, BIDI and QPel: 170 seconds
2nd pass, standard, No MV, Fast Psy, GMC, BIDI and QPel: 186 seconds
(so pretty close to slowest AND QPel)

1st pass, standard, No MV, Fast Psy, GMC, BIDI, NO QPel: 90 seconds
2nd pass, standard, No MV, Fast Psy, GMC, BIDI, NO QPel: 100 seconds

Well... intrigued of the results I've redone the "slowest with QPel test" now and (maybe my CPU is cooler now ;)) the results came identical to the "standard with QPel" test. As a final blow, I compared the two output AVIs binary: identical. So the conclusion is pretty obvious:

Chosing QPel with BIDI and GMC in the slowest mode makes the codec to go back to "standard", while keeping the other options.

Thanx for sharing a little bit of experience with the junior member ;)

[DigitAl56K]

It seems like there may be a bug that prevents you selecting Full Auto PP (we're looking into it).

Also, QPel (as far as I know) is currently unsupported in combination with Slow/Slowest modes. As Acaila suggested, encoding will proceed as if you had selected Standard mode should you attempt to combine these two options.

[SeeMoreDigital]

Originally posted by DigitAl56K
... Also, QPel (as far as I know) is currently unsupported in combination with Slow/Slowest modes. As Acaila suggested, encoding will proceed as if you had selected Standard mode should you attempt to combine these two options. This very issue reminds me about various members comments that were posted a while ago, which is. If a user selects an option that the DivX can't do. Why does'nt a polite 'warning message' appear!

After all it's not a very nice feeling to discover that after hours of encoding, your selected option has not worked as planned!

Cheers

[commando]

Can anyone suggest a good way to do tests? I want to run tests using a few different settings on DivX 5.1. Right now what i'm doing is using a 300MB VOB file, and doing a full encode of it using GKnot. It's taking 20 minutes for the first pass encode on standard more, and I guess it'll take quite a lot longer for the 2nd pass on slowest mode.

My questions are:
- Could I encode it to a standard AVI, then just text DivX compression settings? If so, what program do I use for a basic avi encode from vob? And once I have that how do I do DivX compression.

- Alternately, as I understand the 1st pass basically just gathers statistics. After the first time I do a full encode, could I specify a file that contains those statistics, so I can just do a single pass?

I hope that makes sense... all I want is a good way to compare speed/visual quality on DivX 5.1!!!! Thanks for any advice.

[LordIntruder]

Now I wonder how this possible bug has passed the validation. Whether they don't know about it, odd as some discussions went along about it and if they know about this bug it would be great to indicate it on the "what's new" page or, as you propose, to popup a warning message that under some circumstances you can get a strange result.

By the way I made a test D (see higher for Test A, B & C). I also forgot to mention that my modulation bitrate was set to zero. Usually I prefer to set it to 0.1

Test D:
-------

1st Pass: Slowest, BFrames, GMC, PSY Slow, no MV --> 174min
2nd Pass: Slowest, BFrames, GMC, PSY Slow, no MV --> 174min
------
Total: 348min


Well it's a bit better than test B which is similar but use MV file. A bit better when you watch very carefully.

I look at test A, B & D again today and they are really really close. I even notice today that on some areas test A is a bit better than test B!! However I can say we have a sharper image on test B but it is very hard to notice, you have to watch carefully and very close to the screen and you never watch a movie like that.

In conclusion I continue to use test A adjustements for my future encodes, no need to increase the encoding time by selecting the ones used in test B or D, you only see small differences at 20cm from the screen. That tests are for slow-motion movies. For fast-motion it may change and using slowest for all passes may increase quality. Tests have to be done.

[SeeMoreDigital]

Hi LordIntruder
In conclusion I continue to use test A adjustements for my future encodes, no need to increase the encoding time by selecting the ones used in test B or D, you only see small differences at 20cm from the screen. That tests are for slow-motion movies. For fast-motion it may change and using slowest for all passes may increase quality. Tests have to be done. I have come to much the same conclusion!

With all available DivX settings, I would have thought that using 'slow' and 'slowest' modes, would have revealed far greater and apparent differences to an encode. As it happens, the differences are rather minor and almost totally lost when viewed on a hardware player!

On the other hand, if you were to use WMV9 VCM (at around 760kbps) and encode the same clip using the 'Slowest' and 'Second Slowest' modes. You see 'stark' differences between the encodes, especially with respect to the background movement areas, which become very much more stable and engaging. Unlike DivX....... Shame!

Cheers

[DigitAl56K]

I tend to disagree with that, particularly at lower bitrates Slow and Slowest modes can offer quite strong differences, although when viewing on a hardware player/tv setup its pretty difficult to spot artifacting in any case.

Note that Slow PV is less intense than Fast PV.

[SeeMoreDigital]

Originally posted by DigitAl56K
I tend to disagree with that, particularly at lower bitrates Slow and Slowest modes can offer quite strong differences, although when viewing on a hardware player/tv setup its pretty difficult to spot artifacting in any case.

Note that Slow PV is less intense than Fast PV.
Well, here's one for the DivX web site 'suggestion's box! Why not have some short 'low 750kbps' test encodes on the DivX web site (using the same source in PAL and NTSC), using various 2pass VBR encoding settings.

I am sure this would be a big help for some of 'us' end users. As we would have something to gauge our encodes by!

Cheers

[Zep]

Test D:
-------

1st Pass: Slowest, BFrames, GMC, PSY Slow, no MV --> 174min
2nd Pass: Slowest, BFrames, GMC, PSY Slow, no MV --> 174min
------
Total: 348min




how come you did not turn on MV?

If 1st Pass you use slowest (as you did) then you can use the MV file
for 2nd Pass and save tons of time. That is why the time is 174 minutes
for each pass. Your first pass should be set to Standard and leave the
rest alone. Please post the times then


Anway the problem comes when using MV when you do

1st Pass: Standard, MV
2nd Pass: Slowest, MV

which then uses the low grade MV data made on the 1st Pass set at standard.


at least this is how it works here in my tests :)



It would be great if you also did the tests with MV on both passes too and posted
it here. That second pass should be much faster and still same quality since
1st pass will be done in slowest. Which brings up a question...


which is better for speed and/or quality


1st Pass: Standard, no MV
2nd Pass: Slowest, no MV

or

1st Pass: Slowest, MV
2nd Pass: Slowest, MV


I guess I need to do some tests huh :D


thx


Zephead

[LordIntruder]

Note that Slow PV is less intense than Fast PV.

What do you mean by that DigitAl56K? I didn't make deep test on PV but on those I did I was unable to see any difference between fast and slow PV. As Divx state that slow PV allow to avoid macroblock or such, I prefer to use this one.

If you have made some tests in that area I should be very interested to read your conclusion if any. :)

Thanks.

@Zep:

Please read the entire thread. Also In my last post I stated:

"By the way I made a test D (see higher for Test A, B & C).

That means that in an earlier post I made some others tests and what you request have already be done. :sly:

[Zep]

@Zep:

Please read the entire thread. Also In my last post I stated:

"

That means that in an earlier post I made some others tests and what you request have already be done. :sly:


You missed my point which was that test D it moot because no one should
or will ever use those test D settings and comparing them to ABC is IMHO
a waste of time since ABC will be used and D will not.



Also you never tested (call it test E and another not used setting)

1st Pass: Slowest, MV
2nd Pass: Standard, MV

which you should do if you truly feel test D is not a waste of time
and you enjoy testing interestingly off track settings :)

[Assault]

@ Zep

I think you didn't read the first page of this thread where Acaila posted that when you use MV in both passes the performance settings of the first pass get used during the second pass. So your test E would give you the same results as test B. ;)
Furthermore I don't think test D is useless because it delivers the best possible quality. People with fast machines will use those settings.

Assault

[Owen]

Bitrate control for one pass real time capture is brocken.(STILL)
It seems to be stuck at 4000kb. Encoding 720x576 PAL in Home Theater Profile (no Qpel)gives about 5Gigabytes per hour. Regardless of bitrate setting.
Quality is very good, but at that bitrate it had better be.
I reported the same problem with the beta release.
5 Gig per hour is just ridiculous. I am after 2Gigabytes per hour and DVD qualiy.
Will this ever get fixed?

Regards,

Owen

[SeeMoreDigital]

Hi Owen,

Although I don't capture to DivX myself. I'm absolutly staggered that this fault has not been corrected!

What with all the other problems that are being written about DivX 5.1.0 I don't think it will be very long before DivX 5.1.1 comes out!

Let hope it's sooner than later.

Cheers

[Ravendark]

Originally posted by Sumster
Hi... did anybody notice a difference with the decoder of Divx 5.1 vs. Kaukura? I have an Athlon 1800+, and I could use full post-processing (ie. both at MAX) before with only about 30-50% CPU usage. With Divx 5.1, I'm at 100% with ultra-choppy playback... not sure if there is some other issue at play. Even with automatic post-processing at full, it uses 80-90% with choppiness.

-Sumster

Something is going wrong I think.
I am working on a T-bird 1.2 and at full post-processing I am using 80% of cpu...

[temporance]

Originally posted by Owen
Bitrate control for one pass real time capture is brocken.(STILL)
It seems to be stuck at 4000kb. Encoding 720x576 PAL in Home Theater Profile (no Qpel)gives about 5Gigabytes per hour. Regardless of bitrate setting.I've never seen this bug in single pass bitrate mode. Filesize is pretty accurate, especially with longer (1+ hours) clips.

What application are you using for the capture? Do you also see this bug when using the same DivX settings in other apps?

[SeeMoreDigital]

temporance makes a good point!

It could well be a problem at your end.

Try doing some caps with an application such as MediaPlayer Classic

Sorry I can't be anymore help. I personally use an ADS 2.0 for capturing direct to Mpeg2 at around 6000-9000 kbps. If there's anything worth keeping I rencode to a 2pass DivX file later!

Cheers Owen

[Subwars]

@zep

you should try reading the whole post before you start making comments and end up making a fool of yourself :stupid:


@LordIntruder

thanks for your work on them tests it's helped me and i'm sure alot of other people as well


i've done a few tests myself different ways with a 1.8 celron clocked to 2.1 and with 5.0.5 i get around 25 fps... but with divx 5.1 on standard and pv & b-frames turned off i can get around 10 fps... but on slowest it gets down to 1-2 fps so even if your just to use that slowest on say the 4th pass the last one... that 1 last pass is gonna take like an entire day just by itself... so is it really worth all that extra time.. how about it just being on standard you say that it's still better on standard then 5.0.5 maybe that could be the way to go without each rip taking days and still get an improvment over 5.0.5... not that i recon it's bad anyway the quality

[zag2me]

Originally posted by midi
Speaking for myself and everyone I know personally (college crowd, I guess), none of us care in the slightest about hardware players. A computer is pretty much a necessity in today's world, a hardware MPEG4 player is a rarity.

I would much prefer DivX Networks worked towards making the best computer codec possible rather than worry about this. Monitors are in general much higher quality than TV, but most decent video cards and laptops come with video out anyway if you really need to display on one.

You are joking right?

In the near future every new dvd player will be able to play mp4 and people will begin to "look back at the days" when people actually watched video content on those teenie weenie computer monitor screens, or actually had to turn on their computer to watch a video. This new codec is a step in the right direction for hardware player playback.

Just to add to this, I have done a number of encodes for my K|I|S|S dp-500 in divx5.1 and the quality is outstanding.

[Sirber]

My computer screen is far more detailled than my TV. Also, Who'll buy a hardware player with 2 codec included while a computer can play them all?

[bond]

Originally posted by DigitAl56K
Also, QPel (as far as I know) is currently unsupported in combination with Slow/Slowest modes. As Acaila suggested, encoding will proceed as if you had selected Standard mode should you attempt to combine these two options.seems like dxn tries everything to get people away of advanced codec features like qpel (which is really a great feature!)
also where is my beloved original 2-pass in divx5.1???

[zag2me]

Originally posted by Sirber
My computer screen is far more detailled than my TV. Also, Who'll buy a hardware player with 2 codec included while a computer can play them all?

-People who want silence
-People who dont own a computer
-People who want to watch a movie in their living rooms
-People who want a playback device that is small and looks good
-People who cant afford a computer
-People who want a nice remote
-People who dont want to spend days configuring a device
-People who dont want to buy an operating system

LOL:D got a bit carried away there, but I see the market for hardware players exploading in the near future. Kiss have sold 1/4 million units already. I like the way this new version of divx stays in the MP4 standards, they obviously realise that hardware players are the way forward for this kind of technology.

[SeeMoreDigital]

Nice one zag2me.

Have I mentioned before that Sigma Xcard users can now playback and view WMV9 .wmv files on their TV's with JovePlayer..... I did oh well!

But hang onto your hats Mr JP recently held a poll asking users if they wanted RV9 and QuickTime capability too!

I know it's still a PC device but my computers in another room and I am able to talk to it via snazzy remote. I don't even have to swich on a PC monitor to access the files!

Cheers

[Sirber]

Originally posted by zag2me
-People who want silence: water cooling
-People who dont own a computer: does that exist?
-People who want to watch a movie in their living rooms: Tv-Out
-People who want a playback device that is small and looks good: HTPC
-People who cant afford a computer: Duron aren't expensives
-People who want a nice remote: MPC handle remote control
-People who dont want to spend days configuring a device: Windows is plug and play
-People who dont want to buy an operating system: Linux is always fun :D

just kidding :D

[Lord_KiRon]

And people who want to use divx pro should visit warez link removed by moderator

[commando]

Originally posted by Lord_KiRon
And people who want to use divx pro should visit (warez link removed by moderator)

Whoever wrote the scrolling page should be shot.

[SeeMoreDigital]

I hear quite a lot of people are waiting for the Philips DVD 737 'stand alone' to arrive. And why not when it has a chipset that promises this: -

"Philips has announced that its DVD+R/RW reference design now can be integrated with the pnx1300 and pnx1500 Nexperia™ media processors. The pnx1300 and pnx1500 have support for Ethernet, DivX, MPEG4, H.263 H.264, RealVideo, Windows Media 9, MP3 Pro, Dolby Pro Logic 2, Dolby Digital and JPEG2000 making the platform very flexible for the manufacturers."

Correct me if I'm wrong, but I think that just about covers everything!

Cheers

[Sirber]

Cool!!!!!

I want that!!! Hum... does it supports Matroska?

[Lord_KiRon]

Originally posted by commando
Whoever wrote the scrolling page should be shot.

(Warez discussion removed by moderator)

[deXtoRious]

(Warez discussion removed by moderator)

[jggimi]

6) No warez, cracks, serials or illegally obtained copyrighted content! Links to content of a questionable nature, asking for, offering, or asking for help/helping to process such content in any way or form is not tolerated. Thread closed.

[jggimi]

Thread reopened at the request of SeeMoreDigital and Sirber.

[SeeMoreDigital]

Blimey! Thanks jggimi. Thanks Sirber.

And I thought the most power I ever had was pushing the 'wait' button on a street crossing!

We better start posting something useful here now.

Thanks again

[Sirber]

Thanks jggimi :D

[amango]

Originally posted by SeeMoreDigital
I hear quite a lot of people are waiting for the Philips DVD 737 'stand alone' to arrive. And why not when it has a chipset that promises this: -

"Philips has announced that its DVD+R/RW reference design now can be integrated with the pnx1300 and pnx1500 Nexperia™ media processors. The pnx1300 and pnx1500 have support for Ethernet, DivX, MPEG4, H.263 H.264, RealVideo, Windows Media 9, MP3 Pro, Dolby Pro Logic 2, Dolby Digital and JPEG2000 making the platform very flexible for the manufacturers."

Correct me if I'm wrong, but I think that just about covers everything!

Cheers

You are wrong. The Phillips DVD 737 use an chipset from ESS, which is build also in very cheap DIVX players you can get for 99 Euro in Germany (Yamada DVX 6000/6100 for example). The new pnx-chipset is a new chipset from Phillips, but there are no products yet available.

[temporance]

Originally posted by SeeMoreDigital
"Philips has announced that its DVD+R/RW reference design now can be integrated with the pnx1300 and pnx1500 Nexperia™ media processors. The pnx1300 and pnx1500 have support for Ethernet, DivX, MPEG4, H.263 H.264, RealVideo, Windows Media 9, MP3 Pro, Dolby Pro Logic 2, Dolby Digital and JPEG2000 making the platform very flexible for the manufacturers."I'm not excited about this yet. Just sounds like a very powerful (read "expensive") media processor. And manufacturers will have to pay all the licensing fees for each of these standards. (very expensive). So, I can't really see this being successful. Just look at Philips' track record with the Trimedia ;).

[SeeMoreDigital]

Yes,

It's all got pretty confusing now with regard to the Philips DVD 737. Even some of the Philips reps don't know what's happening with this product.

There are even rumours that the reason why the product is so long coming is because some people within Philips would prefer to use there own chip and not the ESS. So who knows, we may see the same badged player with different chipset in it!

Either way, the Nexperia chip will end up in standalone DVD players soon.

So it may be worth waiting to see what does happens!

Don't you just love cutting edge technology!

Cheers

[Sirber]

if this append, I'Ll be in heven :D. But, what kind of container does it supports?

[SeeMoreDigital]

Originally posted by Sirber
if this append, I'Ll be in heven :D. But, what kind of container does it supports? I think it's a little too early to say at the moment.

However with the correct firmware upgrades there's no reason why it should'nt support just about any container, as the chip looks like it will be able to support just about any codec!

Cheers

[Sirber]

Maybe it's a CPU with software codecs, like RCA Lyra HDD :)

[DaveQB]

the new "slowest' with DivX 5.1 sounds like XviD on max quality settings (VHQ-4, moton search 6 - ultra high etc)

1-2 FPS encoding sounds bout right.

i have to back midi and siber in a way.

a top quality codec regardless of spec's would be great!!
i watch all movies on my PC and on occasions run it through the main living rooms big TV for family viewing.

Never owned a stand alone DVD player and have no intentions to buy one.

must be just me :D

so in conclusion guys... DivX5.1 is the best DivX5 ever released ??

[SeeMoreDigital]

Originally posted by DaveQB
... Never owned a stand alone DVD player and have no intentions to buy one... must be just me :D Well last week I moved my standalone DVD player to another room for the kids to watch. I now use a DVD rom drive (which is connected to my 'Media PC' via USB 2.0), to watch all my movies.

Works really great. Better than expected!

Cheers

[Owen]

Ok people, I have tested Divx5.1 1 pass in Vdub and it obeys the rate control setting as expected.
But it does not obey rate control setting during real time capture using FlyDS.
I get 4Mbit encodes no matter what the setting.

What is going on here?

Regards,

Owen

[rivers]

decided to upgrade to 5.1 from 5.02, and after 2h or reading, understanding & testing there's still a thing in my head that need answer:

why 5.1 with perf setting to *standard* is still slower than 5.02 in slowest? and i mean a 2-pass with MV and every option turned off.

1st pass 5.1 mv: AVG 20fps
2nd pass 5.1 mv: AVG 25fps

with 5.02 was something near 20-25 for the first pass, and 35-40 for the second...

system: p4 3.00GHz, ddr400

tnx in advance

[DaveQB]

my understanding its encoding alot better; the final quality is alot better.

Encoding more the quality of XviD. I went from 25-40FRP with DivX 5.0.2 to 1-2 FPS with XviD, but quality alot better.

What i mean is that since they are now able to attain a better quality (and thus longer encode times) the settings needed to be scaled or adjusted to accomodate.

A bit like how all average size drinks are XL and not MEDIUM like they ahould.

[hhsg]

I have the same problem with you, i'm using VVCR to capture
version 5.0.4 has the same problem
we should report this bug to DivX support, and wait


Originally posted by Owen
Ok people, I have tested Divx5.1 1 pass in Vdub and it obeys the rate control setting as expected.
But it does not obey rate control setting during real time capture using FlyDS.
I get 4Mbit encodes no matter what the setting.

What is going on here?

Regards,

Owen

[Laz]

This may haved already been answered, but what quality setting do I use if I want divx 5.0.5 style results..
I used to be able to encode at 35+ fps but now tiwh 5.1 I can only encode at a max of about 15 on the 'standard' setting...

Anyone able to help me out?

Thanks in advance.

[SeeMoreDigital]

Originally posted by Laz
This may haved already been answered, but what quality setting do I use if I want divx 5.0.5 style results..
I used to be able to encode at 35+ fps but now tiwh 5.1 I can only encode at a max of about 15 on the 'standard' setting...

Anyone able to help me out?

Thanks in advance. Hi Laz.

First try using 'standard' for both passes. And make sure you disable the 'feedback window' and uncheck 'GMC'.

Depending on the bitrate you use, you could also uncheck 'Bi-Di' and the disable 'PSY' settings too. If you then enable and create 'MV log files' you will find your first pass to be as fast as DivX 5.0.5 and your second pass will encode 'very' quickly indeed!

Well worth a try, if you want speed!

Cheers

[Laz]

Thanks for the quick reply :)

I'm not really after speed, I'm just curious to know why its encoding 1/3 of the speed it usually does....

But Yeh, both passes are on 'standard', but I can't remember if those other settings were enabled or not.. (I'm doing the encode ATM)

The feedback window is open, but I've disabled the preview thing, and that seems to have helped a tiny bit...

Is there anywhere I could find optimum settings for Bi directional encdoing and the psycho vis enhancements for particular bitrates?

Usually around the 600-1200kbps mark..


thanks again :D

[SeeMoreDigital]

Originally posted by Laz
...Is there anywhere I could find optimum settings for Bi directional encdoing and the psycho vis enhancements for particular bitrates?

Usually around the 600-1200kbps mark.. That's always an interesting question but not really all that easy to answer! This is because everybody sees an image differently!

In my experience I don't bother with either Bi-Di, PSY or GMC when encoding above 1500kbps!

That said, if you are generating encodes for viewing on a TV monitor (instead of PC monitor) you could get with a lesser bitrate. But it all depends on the frame size you encode to!

My personal interest is in generating 720x576/480 anamorphic encodes that can fit onto a single CD-R for viewing on a TV. And as I've got a burner that will allow me to burn onto 800MB CD-Rs. This is the file size I aim for.

Cheers

[Laz]

I see...

Well anyway I tried an encode using the fastest settings and Its still bog slow @ 15fps..