I Want Better Quality Than Xvid

[firies2006]

what is better than xivd? i have been backing up all my DVD's to convenient movie files on my computer over the past month but the quality just aint good enough in my opinion. i want better. the pixelation for some movies is terrible with xvid, even when setting the target quality in AutoGK to 100%, which "supposedly" should result in no quality loss. This proves to me that xvid aint good enough.

now tell me. is there better out there or am i being too fussy? i know there are limits to DVD quality and Blu-ray would be much better but i can't afford that right now.

All i simply want is to have a nice convenient video format on my hard drive with DVD quality video resolution and original audio.

your opinions please!!

(P.S. I know VOB files do this but they are too bulky and awkward in my opinion.)

[filmboss80]

I would venture to guess that something is amiss in your AutoGK encoding settings. There are issues of bitrate, motion estimation, resolution, etc. that may not be properly set.

With that said, you ultimately cannot have it both ways. (That is, perfect quality and small file size.) To maintain the quality of the source, do not encode the source to a more compressed format. Which is more difficult to live with: poor video quality or "bulky" file size? You may have to choose between the two.

Before completely giving up on Xvid, experiment with different settings until you get something that looks as close to the original as possible.

[ron spencer]

Just save as ISO File...100% quality and no "messy" vobs

[manono]

...even when setting the target quality in AutoGK to 100%, which "supposedly" should result in no quality loss.

Where'd you get that? Make it up, did you? Not true at all, as you found out. You want good quality using XviD? Encode manually to quant 3 and choose a better quantisation matrix, one like the Fox Home Entertainment. This also assumes you know how to feed an AviSynth script into VDub(Mod) and also know how to configure XviD properly, perhaps a big assumption.

[jagabo]

I agree something is amiss with your Xvid encoding. But h.264 does offer better quality or smaller file sizes. Much longer encoding times and more CPU usage on playback though.

[lordsmurf]

H.264 will look better. If for no other reason than it has anti-block tech built into the codec (both encode and decode).
Also ability to make slightly smaller file sizes.

Use MP4 container.

[jagabo]

Xvid has deblocking and deringing in its decoding options. It doesn't work as well as h.264 deblocking and deringing. And it doesn't help on a set-top Divx/DVD player.

[lordsmurf]

The XVID one has to be manually enabled, too, in the player, to my knowledge. The H264 one is just there.

[jagabo]

The XVID one has to be manually enabled, too, in the player, to my knowledge.

It's enabled in the Xvid decoder config dialog. Start -> All Programs -> Xvid -> Configure Decoder. Once enabled it stays that way until manually disabled.

[lordsmurf]

The XVID one has to be manually enabled, too, in the player, to my knowledge.

It's enabled in the Xvid decoder config dialog. Start -> All Programs -> Xvid -> Configure Decoder. Once enabled it stays that way until manually disabled.

Well, now I feel stupid.

Thanks.

[firies2006]

Thank you for all your replies. I will admit that I am not as knowledgeable as most of you on this subject. all I use for xvid encoding is autogk which does all the work. I just ensure that I can get the best possible resolution for each encode I do and ensure that once the encode is complete that the compressibility of the encoded video file is 80% or more. if it is not I increase the file size by 700 MB, so I start at 700 MB, then 1.36 GB AND then 2.05 GB.

However, even if the video file has its maximum resolution and a compressibility of 95%, flaws still show. and as I already said, even when setting the target quality to 100%, these flaws are still there. this is why I don't like Xvid. I mean the huge pixel blocks that are visible in dark scenes in some Xvid compressed movies are terrible. is that what you mean by H.264 having anti-block technology lordsmurf, like that H.264 eliminates the chance of this happening? there most be a better codec for compressing like.

manono I have no idea what you are going on about with using quant 3 etc. so yes, u would be making a “big assumption”. Any chance you could enlighten me as to how you do this? After all, this site is called VideoHELP. Do you offer help or just insult those looking for help?!

What settings in autogk might be amiss?
How do I configure Xvid properly?
How can I get the best quality possible using Xvid?

[jagabo]

Xvid has three basic compression modes:

1) Single pass, constant bitrate. In this mode you pick a bitrate and the encoder uses that bitrate for the entire video. Some movies/shots may get more bitrate than they really need, and some may not get enough. You know how big the file will turn out (file size = bitrate * running time) but you don't know the quality. This mode isn't really useful unless you have some device or application that only supports constant bitrate files.

2) Single pass, target quantizer. In this mode you pick the quality you want, indirectly via a quantizer that can range quantizer can range from 1 to 31. Small quantizers give high quality, large quantizers give low quality. Every frame gets whatever bitrate is required to maintain that quality. You know what the quality will be but you don't know how big the file will turn out. A quantizer of 3 is usually a good compromise between file size and quality. This mode is good if you don't care exactly how big the file turns out but you want a specific quality level.

3) Multiple pass, variable bitrate. In this mode you pick an average bitrate. During the first pass Xvid examines the compressibility each frame. During the second pass it uses that information to encode the video giving more bitrate to shots that need it, less bitrate to shots that don't need it. In the end the average bitrate comes out to the value you specified. You know how big the file will turn out (file size = average bitrate * running time) but you don't know what the quality will be. This mode is useful if you want a file of a specific size. For example 700 MB so you get the best quality that will fit on one CD.

I don't use AutoGK but I know it supports a constant quality mode of encoding where you select a quality from 0 to 100 percent. It automatically computes a quantizer value to pass to Xvid to use in the second mode above.

If you are seeing significant macroblocking with AutoGK set 1400 or 2100 MB something is wrong. Your display's black level may be set too high. Here's a shot with severe macroblock artifacts from (intentionally) too little bitrate (Xvid's playback deblocking turned off):



And with Xvid's playback deblocking turned on:

[manono]

If you are seeing significant macroblocking with AutoGK set 1400 or 2100 MB something is wrong. Your display's black level may be set too high.

Yeah, now it turns out he's talking about the so-called 'dark blocks' or 'black blocks', the blocking you get with flat colors like a dark background or a single colored sky or wall or something like that. Not the kind with too low of a bitrate, although it can be improved a little bit with more bits thrown at it. But yes, it sounds to me like a monitor or TV calibration is needed to minimize or get rid of the things if they're so annoying.

manono I have no idea what you are going on about with using quant 3 etc. so yes, u would be making a “big assumption”. Any chance you could enlighten me as to how you do this? After all, this site is called VideoHELP. Do you offer help or just insult those looking for help?!

What settings in autogk might be amiss?
How do I configure Xvid properly?
How can I get the best quality possible using Xvid?

You're using AutoGK. You can't adjust any settings. You can't configure XviD, except maybe by switching to 1-pass Target Quality encoding, which won't help (much) in this case. You get the best possible with XviD by encoding manually using AviSynth scripts. I offer plenty of help around here but as long as you're married to AutoGK there's nothing much you can do about your encodes. Except for a proper monitor calibration, as mentioned. The 'black blocks' are sort of a fact of life with MPEG-4 encoding. It has nothing to do with AutoGK. Do you have any 'real' complaints about AutoGK, other than that one?

[PuzZLeR]

What's better than Xvid? Depends what you want.

Xvid encodes faster than H.264. Xvid can be edited a bit easier than H.264. Xvid plays in more places than H.264 (currently). H.264 compresses better than Xvid.

MPEG-4 formats such as Xvid/DivX, and implementations of H.264, offer great compression. But the compression they offer really is decent video at much smaller bitrates because MPEG-2 turns into massive blocks at those levels.

But if you're looking for "supposedly no quality loss" (which is impossible but I know you mean visibly no quality loss) then you'd need higher bitrates, and MPEG-4 offers very little advantage over MPEG-2 at that level. You may as well go the ISO route for convenient files that offer that quality level you seek since the compression advantage isn't really linear throughout.

And if using H.264, be careful of that deblocking thing with certain implementations - at lower bitrates they tend to treat the situation with blurring or over-smoothing video.

[firies2006]

Again I thank you all for your input and advice. Since reading your comments I have started to encode some of my movies using H.264 and I am already seeing great improvements over the Xvid ones I have done previously. I mean the visible reduction of “dark blocks' or 'black blocks” (as our cheery friend manono called them) with H.264 is amazing. I think I have found my new, improved codec.

I have been using HandBrake to do these encodes and I am using the Matroska container. Because quality is of the uppermost importance to me, I have been using the constant quality rate option in HandBrake as I amn’t too fussy about file size (within reason) and I have read that using an average bitrate isn’t the best option for quality.

Now I have mentioned quality a lot and PuzZLeR, you hit the nail on the head when you said "I know you mean visibly no quality loss" (and may I thank you for understanding my friend). This is what I mean by quality as I know quality is going to be lost by compressing the source, I just want the visible quality loss to be as minimal as possible.
H.264 definitely seems to meet this requirement.

There is just one thing; I have no idea what constant quality percentage to use for my DVD’s. I said file size isn’t important but this is within reason. Like there is no point ending up with a Matroska file of around 4.8 GB when the main movie on the DVD is around 5.1 GB. I am looking for a file size ranging anywhere between 1 GB and 3.5GB while ideally maintaining the highest resolution. What percentage should I be thinking of using? I am doing an encode right now using a 60% constant quality percentage to check and see what sort of resolution and file size it will produce but it will take about 3 hours to complete. It is slower but definitely worth the wait.

Plus how exactly do you read the log of the encode after and is there any information in this that is important? For example, it was only after a couple of months using AutoGK (which is a wonderfully simple and easy to use program for Xvid conversions manono) that I figured out how important the compressibility percentage in an AutoGK log is to an encode. After discovering this, I would only accept a compressibility percentage of 80% or more and increased the file size accordingly to meet this requirement (as I have previously mentioned above). So is there anything of importance I need to check in the log or elsewhere after a H.264 encode is completed?

Thanks a million in advance for your knowledge.

P.S. When doing an encode using H.264 I am using the original AC3 audio track and do not wish to downgrade.

[PuzZLeR]

I have read that using an average bitrate isn’t the best option for quality.

This has been argued for years. Given two encodes, one done with constant quality, the other with average bitrate, all other settings ceteris paribus with both resulting in the same bitrate there is no distinguishable difference that anybody has proven as a better encode. Having said that, the reason average bitrate isn't the best option is because you need a second pass, which will not be necessary if you desire a certain quality level over an exact file size

There is just one thing; I have no idea what constant quality percentage to use for my DVD’s. I said file size isn’t important but this is within reason. Like there is no point ending up with a Matroska file of around 4.8 GB when the main movie on the DVD is around 5.1 GB.

This is why I don't use x264 any more. At the higher bitrates the compression advantage of it diminishes quite rapidly and offers no real advantage over Xvid, or even MPEG-2. Since you can't edit it, it won't play anywhere really, or takes forever to encode I'd rather stick with other formats for when I seek this quality level.

With HandBrake, you would need something like 65%-70% to equate something like the command line argument --crf 18 with x264 (more or less) which is considered transparency with MPEG-2 source. But many times you end up with a big, big file, and a quality level that even a good MPEG-2 encoder can provide with (in many cases) a lower bitrate, or smaller file size at an equivalent quality setting. So why bother?

Then again, it's a shame. I do believe two of x264's GUIs, HandBrake and MeGUI, are the best encoder interfaces around catering respectively to two different encoding philosophies.

Again I thank you all for your input and advice. Since reading your comments I have started to encode some of my movies using H.264 and I am already seeing great improvements over the Xvid ones I have done previously. I mean the visible reduction of “dark blocks' or 'black blocks” (as our cheery friend manono called them) with H.264 is amazing. I think I have found my new, improved codec.

Well yes indeed. If you're looking to compress something to a smaller size, with an acceptable visible quality loss, then x264 is your best bet. Nothing squeezes down video file size better at the lower bitrates - something x264 is optimized for. But again, this advantage isn't linear. x264 will not do much better, if at all, than other lossy codecs at the higher bitrates.

If you were to use something like 45%-50% with HandBrake/x264 you will get a nice small file. Yes you will notice some lossiness, but it will be a much nicer encode at that bitrate than with Xvid or (goodness forbid) MPEG-2 at that level. That's what x264 is best and really useful at IMO.

But if you really seek indistinguishable quality from the original, I would rather just stick with ISOs in your case. In fact, ISOs will be exactly like the original without need to re-encode.

[poisondeathray]

This has been argued for years. Given two encodes, one done with constant quality, the other with average bitrate, all other settings ceteris paribus with both resulting in the same bitrate there is no distinguishable difference that anybody has proven as a better encode. Having said that, the reason average bitrate isn't the best option is because you need a second pass, which will not be necessary if you desire a certain quality level over an exact file size

Average Bitrate (ABR) isn't the same as CRF, ABR is significantly worse, this has never been argued. If you meant 2-pass variable bitrate (VBR), then if you did an encode and ended up with the same size as one with CRF, then the quality would be very similar. 2-pass VBR is better for devices (e.g. blu-ray player, TVIX, WD etc...), because VBV is not effective in CRF mode.

This is why I don't use x264 any more. At the higher bitrates the compression advantage of it diminishes quite rapidly and offers no real advantage over Xvid, or even MPEG-2. Since you can't edit it, it won't play anywhere really, or takes forever to encode I'd rather stick with other formats for when I seek this quality level.

True, it's more difficult to edit , takes longer to encode

False, that "it won't play anywhere" - many devices support h.264. Ipod, PS3, xbox360, Tvix, WD, Popcorn Hour, PSP, even some optical players are coming out, etc....

True, compression advantage diminishes (non-linear), but it's never as bad as MPEG2 or XviD. You get absurdly better compression at very low bitrates (sometimes you need 2-3x the size if you used MPEG-2 or Xvid to reach similar quality), but you still get great advantage over XviD or MPEG-2 at near lossless levels. ie. on average a certain "quality level" is reached at an earlier bitrate always, and it's usually 20-30% lower when using h.264. There is ALWAYS an advantage at every bitrate, objectively and subjectively. If you've found different results, you must be using a poor h.264 implementation or bad settings.

If you did encodes, varying the bitrate and plot on the X-axis , and measured PSNR or SSIM overy the Y-axis, you will see that MPEG-2 or XviD NEVER cross the line of h.264. i.e. it is NEVER better (or even close) to h.264. In fact, if you use a good h.264 encoder (e.g. x264) it eventually becomes true lossless (@ qp=0), but MPEG2 or XviD does not reach this. I've done 100's of tests that confirm this on multiple genres of sources, both at DVD bitrates, and HD bitrates like from a blu-ray source. Moreover, this trend is not just a trend. It's a FACT. There is not one source that I have ever come across that does not demonstrate this relationship. Similar results have been repeated by many others, and even in University studies. No offense Puzzler, but you are the 1st person I've ever heard say differently. Here is an illustrative example of a DVD source (MPEG2 wasn't represented, because it's worse than even xvid...):





In this example, if 0.95-0.96 was visually lossless while watching (which it is close in reality). This corresponds to ~4000kbps for h.264. It would take almost 2x the bitrate if you used XviD (and even more if you used MPEG2) for equivalent quality!

With HandBrake, you would need something like 65%-70% to equate something like the command line argument --crf 18 with x264 (more or less) which is considered transparency with MPEG-2 source. But many times you end up with a big, big file, and a quality level that even a good MPEG-2 encoder can provide with (in many cases) a lower bitrate, or smaller file size at an equivalent quality setting. So why bother?

CRF has been rebalanced several times during x264 development. It will again shortly, when autoVAQ gets implemented. So what people expected from CRF18 a year ago isn't the same as what it is now. A big change was when psyRD made it into the main trunk, so bitrates increased for the same CRF level. The simple answer is MPEG-2 or Xvid don't provide the same quality at that filesize, otherwise you could have used a much higher CRF level. (i.e. on that particular source, a "visually lossless" result would have been reached at a lower bitrate or higher CRF level using x264 than using MPEG-2 or XviD)

x264 will not do much better, if at all, than other lossy codecs at the higher bitrates.

Completely false. A certain quality level will always be reached earlier by using x264 than something like VC-1, MPEG-2 or XviD. You can repeat the tests, I have done 100's both subjectively just watching on the plasma, and objectively (SSIM, PSNR), using difference masks, frame by frame using avisynth (interleave)

But if you really seek indistinguishable quality from the original, I would rather just stick with ISOs in your case. In fact, ISOs will be exactly like the original without need to re-encode.

I agree!

Cheers

[lordsmurf]

Hey, poisondeathray, according to that chart above, encodes in 1-3 Mbps, where many of mine are, MainConcept is a lot better than x264, as I've been saying. But even then, both appear to be close, whereas MC is a lot easier to work with. I'm still not seeing high quality on x264, but I'm willing to believe enough hoop-jumping and it will suffice. I know that's true of some other audio and video format encoders.

[poisondeathray]

Hey, poisondeathray, according to that chart above, encodes in 1-3 Mbps, where many of mine are, MainConcept is a lot better than x264, as I've been saying. But even then, both appear to be close, whereas MC is a lot easier to work with. I'm still not seeing high quality on x264, but I'm willing to believe enough hoop-jumping and it will suffice. I know that's true of some other audio and video format encoders.

Well lordsmurf you would be correct, for that specific example at the lower bitrate range. I can show you many other examples where the opposite is true. And a delta of 0.001-0.003 PSNR usually doesn't constitute "lot better", and visually is usually not noticable. But for Xvid and MPEG2 vs. almost any h.264 implementation is a joke; and the delta of 0.01-0.02 PSNR in this case does constitute "a lot better" and there is no contest.

The explanation is this particular test is that testing was done using MC SDK 8.2 (not public). The publically available, but expensive MCR 1.6.1 is based on "gimped" version of the MC SDK 7.6. This current version of Mainconcept Reference doesn't have any of of the higher quality features like complexity masking / lumimasking (similar to x264's VAQ), but higher end versions of Mainconcept implementation (e.g. Sonic Cinevision), and the SDK, do have these extra features.

I can tell you from my testing , overall, about 80%-90% of the cases favor x264 over using MCR 1.6.1. It might be bad settings that you are using where you are seeing MCR providing "better" results than x264. I am certain on the majority of sources, that (especially) subjectively, and to a lesser extent objectively (by psnr,ssim) you will see the difference. The biggest difference is MCR's lack of AQ. When you examine with avisynth using interleave() the difference is like night/day - it's that big! Entire textures and background are missing when you use encoders without AQ at low bitrate ranges. If you want I can post some examples with multiple sources. After dozens of tests, I can't find a source where x264 is consistenly worse than MCR 1.6.1. (x264 might be worse in 5% of frames, equal in 10% of frames, and better in 85% of frames which is the usual distribution that I find for progressive content; MCR is clearly better at interlaced h.264, no contest there.) If you have such an example that demonstrates where "MainConcept is a lot better than x264", can you please share a sample of the source?

To be honest , PSNR and SSIM testing isn't very good. It's an objective measure trying to measure a subjective trait. x264 usually has lower SSIM, PSNR values than it should , because of psy options (essentially "noise" is added to the picture, making it look more detailed). Recent builds have the psy-rd @ 1.0 (which IMO is too high), which is one of the reasons why people are seeing higher bitrates at the same CRF level. But , of course this is adjustable. I can show you many examples where going by PSNR alone is way off (i.e. the picture with the lower PSNR is clearly better). So as you already know and have preached in other threads, you have to examine by several methods across several sources/genres to be more complete

Cheers

[PuzZLeR]

Yawn... Ok, I've decided to wake up and re-join this thread – after I ran a few more tests (again) and particularly because I also laughed my head off with a certain recent post from x264’s home field saying that “x264 can reach transparency at around 2500Kbps” (from a DvD source).

Wait… let me think about this a moment….

HAHAHAHAHAHAHAHAHA!… YEAH RIGHT! Black screen maybe? Certainly for low motion, but not in general.

And what can you do with the Fan-Boys?

Ahh Poison, I knew you, a “Fan-Boy” as well (no offense), would jump on this thread soon enough to rescue your beloved x264. Again.

Listen my friend, it's really quite simple. I don't care. Let me put it bluntly. As long as I see a blur or a smear within the encodes I will not let it go. Yes, blah, blah, blah, latest versions, blah, blah, blah, AQ, blah, blah, blah artificial complexities and biased numbers, whatever. YAWN.

That blur and smear has NEVER left, and, moving along as a function of bitrate, it still remains long after MPEG-2 reaches, at least a respectable form of, “transparency” – regardless of the supposed “better features” of the x264 encode.

So yeah, you can show me all the pretty little graphs, with the pretty little colors and the data from the little university students (who notoriously “download” a lot), etc, whatever, the "better features" of the x264 encode take a backseat to (slight but apparent enough) blurry video in my eyes. And BTW I can make an argument on how such results are based on an outdated concept, or a subjective paradigm/hypothesis, and how they are constantly misinterpreted by many - including by you (logarithmic, not linear). But, I assure you as well, the professionals that I deal with don’t take them seriously anyway.

I still see this slight mist within the video, and so do others, even from the latest builds and even at CRF=18 (whatever that’s supposed to be). And the Fan-boys still defended x264 when this effect was so much worse not too long ago. So why should I expect any less cheerleading when indeed it has been corrected somewhat in recent times (albeit with higher bitrates)?

And you do have a point about x264 having nicer encodes, even at the higher bitrates visually… I’m not blind… BUT ... we still have that blur.

Now I understand that a softness is indeed commonplace among several H.264 encoders - even professional ones can be guilty of it - but with x264 it feels like a smear. I have never been able to shake it off without high, high bitrates – which, again, hurts its usefulness and certainly invites other options for high quality video in my opinion - which is my point.

Unfortunately it’s STILL there years later. It’s faint, it’s light, but there is still a slight but apparent smearing that still resides over the “nice encode”. My friend, I can’t take this smear. To me it’s like a stink over a nice encode and it kills everything.

And I could care less if it plays on some exotic player or some PS3 either...(iPod high quality video? You kidding me?)

I still say x264 is a great accomplishment for what it is, and the best in the world IMO, for mobiles and “distributed” Internet content, but, given its flaws, and numerous and fragmented compatibility issues (which includes blu-ray and interlaced encoding as well) it's simply not practical IMO for serious, or professional type of, video. Sorry my friend. I’m far from convinced otherwise.

But I will commend someone like you instead, contributing, in one way or another, to its development instead of being like me who will bitch over a free tool and just give up on it. In sincerity, I do actually appreciate this. Hy hat is off to you.

Oh, and by the way, I will be out of town for a few days so wishing another fellow countryman a “Happy Canada Day” holiday coming up. Not sure where you live though - Toronto? If so, lunch is on me.

Cheers,
Geordie.

[SatStorm]

I'm with PuzZLeR on this:
When I compare my x264 results with the realtime results of my new Hauppauge USB HD PVR (which still has bugs and driver issues), the x264 results looks "bad".

I don't know what kind of H264 encoding this card do, but it is better x264 and Mainconcept.

With H264, at 4.000kb/s it looks like mpeg 2 at 9.000kb/s. But mpeg2 at 9.000 is not enough some times, so H264 at 4.000kb/s.

It is silly, but I start thinking mjpeg again myself...

H264 and X264 is great. But probably it is not enough for some people, including me.

[quide]

Hi,

There is just one thing I still don't get it:
Why there isn't a 4th option that does a "Multiple pass, target quantizer"?

I just don't care about the target size and want variable bitrate (to get the smallest file possible) for a certain quality.

Cheers!

Xvid has three basic compression modes:

1) Single pass, constant bitrate. In this mode you pick a bitrate and the encoder uses that bitrate for the entire video. Some movies/shots may get more bitrate than they really need, and some may not get enough. You know how big the file will turn out (file size = bitrate * running time) but you don't know the quality. This mode isn't really useful unless you have some device or application that only supports constant bitrate files.

2) Single pass, target quantizer. In this mode you pick the quality you want, indirectly via a quantizer that can range quantizer can range from 1 to 31. Small quantizers give high quality, large quantizers give low quality. Every frame gets whatever bitrate is required to maintain that quality. You know what the quality will be but you don't know how big the file will turn out. A quantizer of 3 is usually a good compromise between file size and quality. This mode is good if you don't care exactly how big the file turns out but you want a specific quality level.

3) Multiple pass, variable bitrate. In this mode you pick an average bitrate. During the first pass Xvid examines the compressibility each frame. During the second pass it uses that information to encode the video giving more bitrate to shots that need it, less bitrate to shots that don't need it. In the end the average bitrate comes out to the value you specified. You know how big the file will turn out (file size = average bitrate * running time) but you don't know what the quality will be. This mode is useful if you want a file of a specific size. For example 700 MB so you get the best quality that will fit on one CD.

I don't use AutoGK but I know it supports a constant quality mode of encoding where you select a quality from 0 to 100 percent. It automatically computes a quantizer value to pass to Xvid to use in the second mode above.

If you are seeing significant macroblocking with AutoGK set 1400 or 2100 MB something is wrong. Your display's black level may be set too high. Here's a shot with severe macroblock artifacts from (intentionally) too little bitrate (Xvid's playback deblocking turned off):



And with Xvid's playback deblocking turned on:

[jagabo]

Why there isn't a 4th option that does a "Multiple pass, target quantizer"?

Target quantizer mode is like saying "remove this much detail from every frame". There's nothing to be gained by a second pass. You would just be repeating the exact same thing. You already have the smallest file size (depending on other settings you used) for the requested quantizer.

VBR requires two passes because the encoder needs to know about the entire movie before deciding on the best strategy for compressing it. During the first pass the encoder is just examining the video and saving a log of how difficult each shot is. During the second pass it does the actual encoding using the information from the first pass. That way it can allocate more bitrate to shots the need it, less to shots that don't, and still meet the requested average bitrate.

[deadrats]

what is better than xivd? i have been backing up all my DVD's to convenient movie files on my computer over the past month but the quality just aint good enough in my opinion. i want better. the pixelation for some movies is terrible with xvid, even when setting the target quality in AutoGK to 100%, which "supposedly" should result in no quality loss. This proves to me that xvid aint good enough.

now tell me. is there better out there or am i being too fussy? i know there are limits to DVD quality and Blu-ray would be much better but i can't afford that right now.

All i simply want is to have a nice convenient video format on my hard drive with DVD quality video resolution and original audio.

your opinions please!!

(P.S. I know VOB files do this but they are too bulky and awkward in my opinion.)

try using xvid's "hi def" profile (i know your target resolution us sub-hi def), my personal experience is that it produces the best quality, in some cases on par with h264.

also don't use to little bit rate, the quality will suffer.

[branch]

False, that "it won't play anywhere"

That's not what he said. The "really" matters.

[jagabo]

It's called hyperbole.