XviD - it's time to move on!

[Vanos_b]

Why do I say that? Because I have the feeling that there is no noticeable progress in the development of the codec. Don't get me wrong, I really apreciate all the great work you guys are doing, but I think that there would be other things to improve than curve compression (and this is the only thing I've seen changing in the last builds). To be frankly I don't care much about the DSfilter because I think a good codec wouldn't need it anyway, and I couldn't care less about encoding speed because the only thing that matters for me is quality. First of all I think it should be corrected the problem of dark objects moving on a blue background that is very anoing anywhere it appears the sky. This is a very old problem and I really don't understand why it hadn't been solved by now. And then there are the B-frames, that I think XviD cannot "survive" withought. I know the work on this is in progress, but still think that is time to put the other things apart and focus more on this. I have to say that for me the quality of the codec is the lowest quality it provides (during the period of a movie, lat's say), and even thogh XviD behaves very well on a great part of the movie materials there are some scenes where it looks very ugly (compared with DivX5, that I think for now rules). And one more thing about blocks that appear on fade in/out scenes (like the Warner Bros logo in the begining of some movies) - very blocky no matter what settings. DivX5 is one pass further, but XviD could be two.
Regards.

[Teegedeck]

I'm sorry but I don't see anything in your post that would qualify it as meaningful. We all KNOW that curve compression isn't a magic wand that will improve encodes; the developers ARE working on B-frames (but if people like you start nagging they'll start to ask themselves 'why am I doing this, I don't get paid for it'). So, as this thread won't trigger any meaningful discussion, I'll only post this answer and hope you got me.

[Nightweaver]

I don't really have much of a problem with the way they're working now, but I'll admit, quality probably isn't as high on my priority list as it is on yours. I'm definitely not ripping professionally - I'm more doing it to see if I can, and to see what it looks like, and occasionally to make convenience copies of DVDs that I can just keep on my hard drive. That, and I like watching how fast development progresses ... it gives me a warm fuzzy feeling, or something like that :D

[-h]

I think you could have said that with a bit more tact. We are doing this for free, because we enjoy it. If you'd like to steer the codec in another direction, learn C and x86 assembly, read the MPEG4 natural video spec, and start contributing.

In regards to your point on no noticable progress:

B-frame encoding/decoding are working in a private build, but due to several remaining issues with interoperability (and a VDub mod), they haven't been committed to CVS yet. That's because as soon as B-frames show up, everyone will start yelling and screaming about it even though they don't give all that great a benefit, and Isibaar doesn't want to draw such attention until it's working 100%. That you think XviD can't "survive" without them (and how exactly does a GPL'ed project "die"?) shows that you've bought into the marketing.

A number of motion estimation modifications are in progress, which should drop bitrates yet again, and probably improve performance on the vague things you mentioned (dark objects on blue background, fades).

Interlaced frame support is mostly working and will be committed in day or so.

Quarter-pel interpolation is being tested, and should also be committed soon, along with related compression benefits.

I'm not 100% sure why your post annoyed me so much, but it did. Giving orders to open source developers, without understanding why said developers actually work on the project, does tend to draw responses like this.

-h

[Asmodian]

Vanos_b did you mean to come off demanding? I have nothing in the way of coding skills to offer the xvid community and would like to see the problems you mentioned fixed but have no idea how hard that would be. Are the people that are spending too much time improving this wonderful free codec's direct show filter and curve compression (as you see it, personally divx 5 isn't that great without post processing either – and the better the curve compression the better the codec, it was that kind of 2 pass processing with nandub that made divx 3.11 usable) even the same people who would be interested in "fixing" the problems? Are you complaining that people spend their free time improving what they want to improve rather then what you think they should improve? or are you just offering a helpful suggestion because you don't think they (who ever 'they' are ;) ) noticed those problems and though all they could do was the DS filter? Remember xvid is free, no one is paid to work on it… you didn’t pay for it… (and neither did I).

edit:
oops came in late... but your post even annoyed me (and I haven't worked on xvid, being a chemist and new to the scene)

[Selur]

@Vanos_b:
Sorry but when I read your post I would say it's not "time to move on!" but time to "think before posting"

-h and all the other folks working on the codec are doing a REALLY amazing work there,.. haven't seen a project (even comercial) where the programmers are so commited to the cause a looong time,...

Cu Selur

Ps.: Xvid rocks!!

[The People's Elbow]

@Vanos_b: Your post shows ignorance, impatience and...that you ****! Sorry, but you should think about what and how you talk about certain things, if you have no clue about them! For example your opinion about a DSF and a "good" codec which wouldn't need one - lol!
I don't wanna write more, else this post might get deleted.
Just one request: Keep your mouth shut next time, makes everyone feel better!

@XviD Developers: There is just one point in Vanos_b's post that I have to agree with in some way. This "dark objects in blue sky thingy" is really worth to get rid of, especially for me as a guy who likes to encode anime content.
Anyway, keep up the awesome work, I really appreciate it! :D

greetz from the Elbow!

[droolian01]

This post anoyed me too. There is nothing wrong in suggesting where the programers ephasis should be, but your tone was very dismissive of the hard work that these guys have put into a briliant project. I have decided to do all my tv captures with xvid, despite the fact i can't play them with the xvid dsfilter on my machine (soon to be sorted) because the codec gives excelent results now, has developed at a lightening pace, has more to do (great!)and more importantly it is an honourable excercise. These bril people do this stuff for nothing out of sheer pleasure, i cannot share this (can't program) but can get similar pleasure in 'creating' my own encodes. The moment any one of these people cease to get pleasure from working on xvid then they will stop, your comments could cause this, and i will have to find another hobby.
Keep up the good work and ignore idiotic comments.

[Vanos_b]

From your replies I feel like I have to apologies to the developers of XviD (mainly -h), but not to the others. I'm sorry if my post annoyed you and IT WASN'T MY INTENSION TO GIVE ORDERS. I just wanted to know why the development is going this way. Sorry for the title, it wasn't chosed that well. I know you do it for free in YOUR own free time. And as I said I really APRECIATE this (although you may of thinked otherwise).

B-frame encoding/decoding are working in a private build, but due to several remaining issues with interoperability (and a VDub mod), they haven't been committed to CVS yet. That's because as soon as B-frames show up, everyone will start yelling and screaming about it even though they don't give all that great a benefit, and Isibaar doesn't want to draw such attention until it's working 100%.

That you think XviD can't "survive" without them (and how exactly does a GPL'ed project "die"?) shows that you've bought into the marketing.

I didn't knew that and I'm happy to hear it. but I still can't get it why do you say they don't give such a great benefit. Please explain, or at least tell me what is giving a greater benefit.

No i didn't. But I think this is essential for a noticable compression improvement (I tested it with DivX5 - 23%), and of course a quality improvement. I'm not saying this codec would disappear but if it can't keep up with the other codecs (commercial or not) the large public would lose interest in it. XviD behaves very well even withought B-frames and I think that with them it could beat DivX5, and that's why I'm so impatient to see them implemented.

If you'd like to steer the codec in another direction, learn C and x86 assembly, read the MPEG4 natural video spec, and start contributing.
I don't want to "steer the codec in another direction". I like the direction it goes, but it just didn't seemed (to me) that it moved. I know C and some assembly but I think your doing this better than me so there is no point to me for bugging in. Hopefully I will contribute, but for now I'm trying to get aquainted with the code.

...without understanding why said developers actually work on the project, does tend to draw responses like this.
Seriousely: Why are you doing this?

Regards.

[The People's Elbow]

Vanos_b wrote:From your replies I feel like I have to apologies to the developers of XviD (mainly -h), but not to the others.
- You should apologize to the whole XviD Community, if you ask me! ;)

...and I can tell why your first post is so annoying:
1. You said that all the codec improvements (speed, DSF, ...) which were achieved in the last 2 month are nuts! - The opposite is the cse I guess!
2. You said that B-frames are the only way to make XviD a competitor to other codecs (DivX5) which implies to the reader "XviD actually sucks". - Likely offends guys from the XviD Forum ;)
3. The whole post sounds like a complaint about XviD's status and developement.

...That were my intentions while reading your initial post!

- greetz from the Elbow

[Peters]

I moved on..xvid encoding for a 'long' time, i like this codec, i like opensource, i like people working on this for our pleasure, i like the 'philosophy' of these people

It's a great work guys, i'm not impatient, Xvid already gives me wonderful results...

I'm really happy this codec exists...

[athos]

i will not comment on Vanos_b's posts, but i would like to than -h for lettings us now what mayor things are going on in xvid development! this is all very exiting!

[Koepi]

So only appologize to -h?

what about foxer? isibaar? suxen_drol? gruel? all those people i can't list? Finally, me?

You leech just something from the net, knowing it is _ALPHA_ (not even beta), but keep comlaining and insulting us (as the developers).

I don't like your attitude.

You should think it over some more times.
Spare me with replies.

You definatly ****.

Koepi

[R3g]

I just want to tell you my pont of view on this question :

- Why is the development going that way ? Because the developpers want it.

I really think any improvment brought to the code IS important, even on curve compression and DSFilter. I also really think that anyone who doesn't get involved in the developpment (like me) should stay away from judging the choices of the developpers. Making suggestions or reporting experience is one thing, saying "THIS is the thing to do" or "THIS is totally useless" is just unacceptable. With posts like yours, the developpers will one day decide to continue the developpment for their own, and we will all miss that great codec.

You are free to say what you think, but please be more carefull with the way you say it.

[philippas]

@The People's Elbow, i agree with you.

@Vanos_b
I think Xvid gives much better quality than divx5. The only thing that divx5 does "better" is that it blurs the image and remove more colour, that's why it appears that it gives higher compression ratio.

Also about b-frames for xvid, virtualdub or any other application currently available is not a suitable for non linear encoding.
By -h
Since B-frame encoding requires frames to be fed to the encoder "out of order", there is talk of creating a custom encoding program to generate XviD AVI's containing B-frames. These AVI's will be playable with all the normal tools (Media Player, PowerDivX, etc.), it's just the encoding that's tricky

Maybe you are wondering why divx5 can use b-frames! Because it can use only 1 b-frame between two keyframes it doesn't give much improvement in compression. Also when divx5 uses b-frames it compresses them at quant=4 (correct if I’m wrong), no matter if you limit the quant values, which is "cheating" and it doesn't give good results but blurs the image even more.

[Apo]

This is indeed very annoying.
Everyone who is working on Xvid is really doing a great work and not only h but also Isibaar, Koepi, Nic and everyone else who is spending their spare time to improve this codec and is so kind to share their improvments with us. They could also hold them back and have fun with it alone. I really hope that the developers just don't pay attension to this thread.

To the developers: YOU ALL ARE DOING A WONDERFUL JOB. DON'T GET
OFFENDED BY SUCH STUPID STATEMENTS LIKE Vanos_b IS MAKEING, JUST IGNORE IT.

Apo

[wing1]

@vanos_b

You have to take into consideration that xvid is still young and it has come a long way to run neck to neck with divx5. Look at how long divx4 and divx5 took for developing time ( are these guys being paid to do that? :D ).

I see xvid is being improved tremendously on a daily basis. Where else can you complain about a certain aspect of the codec and you get response with a few hours? That alone is worth my time to continue testing out the codec. I know that you want b-frame and so do a lot of other folks who are testing out this codec. It will all come together soon enough.

Like you, I don't care about curve compression algorythm or 2 pass encoding algorythm, but there are others who do use that for encoding purposes. Nevertheless, the benefits do trinkle down to other aspects of the codec while it is being tweaked by the developers; It may not be apparent with what you want to use it for, but it does to a smaller degree. I do see that with every new releases, but then I would be imagining all of these improvements :D

I do see that xvid will be the best codec out there for mpeg4 in a near future from the first day I start to use this codec. It will take time but it will be worth waiting for it.

As always, work with what is suitable for you.

Just remember one thing : Good things come to those who wait :D

[Vanos_b]

@Koepi:
From your replies I feel like I have to apologies to the developers of XviD (mainly -h),

I thought you were a serious mod, and I'm sorry you proved me otherwise. I never intended to offend anyone. And I won't respond to your insults (or anyone else for that matter). I appologiesed once for not putting it (my ideas) the right way. I am just sorry for what this thread turned into, as my only wish was to express my sugestions.

:(

[-h]

I didn't knew that and I'm happy to hear it. but I still can't get it why do you say they don't give such a great benefit. Please explain, or at least tell me what is giving a greater benefit.

Because they don't. The only reason they decrease the file size in DivX5 is because the quantizer is doubled for B-frames. This creates a rather annoying effect (for me anyway), and lowering the B-frame quantizer through the registry eliminates most of the space-saving.

I'm not saying this codec would disappear but if it can't keep up with the other codecs (commercial or not) the large public would lose interest in it.

And I'm saying that isn't something that motivates us.

Seriousely: Why are you doing this?

Because data compression fascinates me, and this is a great project to play with.

-h

[Koepi]

Time to close this thread IMHO. If some of the other mods want to do it, I agree, I won't as for sake of free speech.

@vanos_b:

You're not improving things when adding one insult to another. My point of view doesn't change, it's fixed now as you proved me right.

@All:

Thanks for the positive feedback. I guess it's time to update the documentation of XviD, hopefully I can do that in this new week.

Best regards,
Koepi

[skyline]

From reading between the lines, I feel that vanos_b was actually trying to provoke some discussion on where the development of xvid is at.
I think he/she suceeded.

Thanks Xvid developers and long live Xvid!

[Teegedeck]

I agree with you, Koepi. I would have preferred if the thread would have been ignored, but now we have a discussion that only produces bad feelings. Let's just forget about it, start over again and have a nice rest of the day, what do you say? :)

Thread closed