is AVC really that good?

[nexx]

Originally posted by Blue_MiSfit
Would it be possible to interlace these 60p sources thereby making older machines capible of decoding HD (admitedly at the cost of interlacing)?


I'm no expert, but it is entirely possible to take a progressive image and create interlaced output (at half the frame rate). I've used MEncoder for this exact thing, I'm sure it's possible with other tools. So you'd get 720i. I dont see how it being interlaced will help, but there is half the amount of video data to be decoded. You could just half the frame rate?

[SeeMoreDigital]

Originally posted by nexx
I'm no expert, but it is entirely possible to take a progressive image and create interlaced output (at half the frame rate). I've used MEncoder for this exact thing, I'm sure it's possible with other tools. So you'd get 720i. I dont see how it being interlaced will help, but there is half the amount of video data to be decoded. You could just half the frame rate? Agreed....

I find it a bit odd as to why 60fps is being used to broadcast HD content anyway... unless it's to provide a "half arsed" solution for owner/users with non progressive displays!

I had originally got it into my head that maybe each one of the 60No "fields" had been turned into a full frame. But, apparently this is not the case, each successive frame is slightly different.

That said, when the whole world goes completely digital, with everybody owning a progressive digitally connected display, all video could be displayed at 24fps :D


Cheers

[Sharktooth]

i'm uploading a new clip made with xvid1.1EDP (thanks cruncher).

[Sharktooth]

Originally posted by JohnV
Try using 1 or 0 b-frames. I'm interested if that makes a big difference for you with your 50" screen.
I made several encodes with different settings.
It seems 2+ b-frames is the major cause of blocking. With 1 b-frame you can safely keep the deblocking strength to a lower value and gain more details.
I suggest you to set 1 (or max 2 bframes) as the default recode setting.

[len0x]

Btw, you should all realize that this test is actually really flawed as if you do a comp test on the source you'll see that for XviD its ~32% for x264 it's ~26%(1 bframe), so basically no sane person should encode the above sample at the given bitrate and resolution (at least 1800kbps is needed for the width of 720).

[CruNcher]

len0x ehh i wouldn't say so this hits for 1 Cd bitrates ok 896 kbps is a little over 1 cd wich would be in the region of 7xx sure for the content thats used here i would not go that low (trailer) but that's the thing we need to test @ wich bitrate does it still looks visualy pleasant (and compares against others) also in those extreme samples you and Gknot can't adapt as fast as the codecs feature change and compression get's improved i personaly never use any compression test to set my settings if i have a original i know how much ASP/AVC is capable off (at the moment) and set the factor by that not doing any compression test on the sources it's like alot of things a expirience question.

[len0x]

Its true that comp test for AVC is not really defined yet, but it certainly is for ASP codecs. So this test is like saying "I know its gonna look really really bad with ASP codec, but I wonder how bad its gonna look with AVC?". AVC codecs never promised us to deliver the same quality as ASP with bitrates lower than 50% of the latter. So you know you're not gonna be satisfied with results of any encoding at that bitrate and resolution even before you start...

P.S. Is it just me or everyone else finds it difficult to read CruNcher's post above without any punctuation? :)

[SeeMoreDigital]

Originally posted by len0x
Its true that comp test for AVC is not really defined yet, but it certainly is for ASP codecs. Personally speaking, I think there's a need to generate more Mpeg4 "comp tests" using SP and ASP modes. And at DVD pixel frame sizes. :eek:


Cheers

[dragongodz]

P.S. Is it just me or everyone else finds it difficult to read CruNcher's post above without any punctuation? :)
read it really fast while holding your breath, its not that hard. :)
no punctuation ? he has a fullstop at the end so what are you nitpicking about ? :D

[Sharktooth]

Anyone watched the xvid1.1edp clip?
IMHO it looks really good in terms of details. Sure, it has some blocking and ringing artifacts but it definatly has more details than any other encode...

[JohnV]

Originally posted by Sharktooth
Anyone watched the xvid1.1edp clip?
IMHO it looks really good in terms of details. Sure, it has some blocking and ringing artifacts Ah, now it's only "some".. :D
Sorry but in EDP's case I'd call it horrible blocking.. Details are nice though.

[Sharktooth]

well, i cant expect that an ASP codec doesnt block at that res/bitrate.
AVC has in-loop filter that should prevent it to block even at the lowest bitrates...
Anyways have you tried different b-frames settings while encoding the trailer source with recode?

[scharfis_brain]

where do I get a small sized working avc-playback filter from?

the mpegable avc doesn't work (MPC refuses playing)
ffdshow crashes
and nero's codec is really huge in file size.

[Sharktooth]

ffdshow shouldnt crash. try downloading the latest celtic_druid build.

[Tommy Carrot]

Originally posted by Sharktooth
Anyone watched the xvid1.1edp clip?
IMHO it looks really good in terms of details. Sure, it has some blocking and ringing artifacts but it definatly has more details than any other encode...
This edp encode apparently favors the low-motion scenes very strongly. Didn't you encode it in 1-pass cbr mode?

[JohnV]

Originally posted by Sharktooth
well, i cant expect that an ASP codec doesnt block at that res/bitrate.
Well, if we are talking about video quality, imo we should talk about what we actually see, and not what we expect to see..
If the point of this thread is to compare different codecs with this trailer, why not discuss without any expectations what we actually see.

It gets soon very hard to interpret this discussion if EDP only has "some" blocking and ringing only because "we can't expect more from ASP"..

[SeeMoreDigital]

Hi Sharktooth,

Have you thought about generating these encodes again at 1:1 and raising the bitrate accordingly!

I'm wondering whether some of the blocking and ringing effects might reduce if the encodes are generated without resizing...


Cheers

[JohnV]

Here's a clip with a bit newer filters using 3-bframes and fixed deblocking -2 with extra quality and all tools enabled.
http://morbo.org/nero/sp2tr_b3_deblock-2.mp4

Don't know if it's better or not than the one Sharktooth provides, you decide.
I should get today (monday) again new filters with hopefully even better quality.

[kurt]

sharktooth, nice little codec comparison! especially the x264 AVC file (with b-frames) is IMHO impressive. The last one from JohnV too - wich filters do you mean, johnV? new version from ateme?

BTW, wouldn't it be better to encode without resizing?

[Sharktooth]

Originally posted by JohnV
Here's a clip with a bit newer filters using 3-bframes and fixed deblocking -2 with extra quality and all tools enabled.
http://morbo.org/nero/sp2tr_b3_deblock-2.mp4

Don't know if it's better or not than the one Sharktooth provides, you decide.
I should get today (monday) again new filters with hopefully even better quality.
Well it's much better then the actual filter with the same settings.
However there are still some blocks but comparing with the "actual" encoder and the deblock strength you used it is much much better!
The facial details are also improved and the high-motion scenese are less "motion-blurred" and definatly better looking.
There's still some work to do with the "flashing" blocks but i'm glad to see that 3 b-frames is a choosable option now.

Originally posted by SeeMoreDigital
Hi Sharktooth,

Have you thought about generating these encodes again at 1:1 and raising the bitrate accordingly!

I'm wondering whether some of the blocking and ringing effects might reduce if the encodes are generated without resizing...
You mean anamorphic encode? No. I choosen a mid-high res and a medium bitrate (the usual settings for a 1 CD backup with newer codecs) to see how much AVC is better than other codecs for making 1DVD->1CD backup.

[JohnV]

Originally posted by kurt
sharktooth, nice little codec comparison! especially the x264 AVC file (with b-frames) is IMHO impressive. The last one from JohnV too - wich filters do you mean, johnV? new version from ateme?
It was our newer filters than in the current Recode release. These will be available in the next update. And there's speed improvement too.

[plonk420]

Originally posted by calinb
720p with AVC is tough without a GPU and hardware acceleration. The Nero H/W acceleration doesn't seem to work on my ATI 9800 pro or Nvidia FX 5200 yet. Bottom line: I can't do it with a 2.1 GHz XP-3000+ Barton (512L2/200/400FSB) or 3.2 GHz Prescott (512L2/400/800 FSB). They can barely handle 24-30 fps without partitions / CABAC. (Tell me what constitutes the main profile, again :))

For comparison: 720p mpeg2 plays fine on my old 1.5GHz Willamette and an ATI or Nvidia DxVA decoder (Dvico, Moonlight, ATI). A 1.8GHz T-bred can do it with non-DxVA decoder.

well, mpeg-2 and wmv9 accelleration is supposedly working already with geforce 6600/6600GTs and WMP10/nvDVD. can it be harnessed by MPEG-4 AVC/ASP decoders perchance? my 2400+ can't even keep up with some of the ASP encodes i've done...