DVD2SVCD / CCE quality

[kqt]

I ripped a DVD using DVD2SVCD. CCE setting is set to 5 pass VBR. The output was viewable but very jerky for the fast or even slightly fast motion scenes. I tried DVDx also and the results were better in regards to this problem. saw many posts raving about the quality of CCE. Has anyone experience this problem? Thx.

[mrbass]

What's your playback device? Is that on your computer? I do 2 pass vbr and for motion movies..matrix, etc. do 3 pass and notice no artifacts. I know panasonic dvd players and sometimes certain apex models can quite handle it. Does the audio ever break up? Is your avg bitrate over 1800?

[sunmiztres]

It sounds like your field order could be backwards. On the video tab ( i think) the second choice down on the left hand side. if this was checked the first time you encoded uncheck it and reencode and then check to see if the problem still exists. cce gives great quality and with 5 passes you should be looking at dvd qulaity. i think that is what your problem is or the bitrate is to hight for your player. if you used the original settings in dvd2svcd then the bitrate should be fine.

[kqt]

I used a Pioneer 434 for play back. Playing back on the computer is also a bit jerky though better. In DVD2SCVD, I used all default settings. Is there a way I can tell if the bitrate is above 1800? Also, I'll change the NTSC Field operation from automatic to off and re-try the rip. Hope it comes out better.

[sunmiztres]

go to the bitrate tab on dvd2svcd and it will tell you your bitrate settings. they are variable so they can be anywhere from 300 min to 2400 max depending on your movie.

i have the same problem. but only on a few dvd's.
29.97 fps interlaced.
ive played w/ all the settings but in a pan or mation scene i get a bad flicker or a ghost image that seems to jump.
ive tried this w/ dvd2svcd and w/out, and get the same thing. so it has to be a setting in cce... but im at my wits end. usually i do 3 passes 25 quality. no progressive frames, no zigzag. if anyone has a clue PLEASE fill me in.
thanks

[Kdiddy]

I was just about to start a new post on the very same topic..I keep reading about this "dvd like quality" that CCE produces, but yet to have seen where it is better than TMPG, if anything they are same, which is not quite "dvd like"...I am hoping/wating to be proven wrong...I ran test after test..NTSC @ 23 fps, runs to fast & jerks, I cant get a VFAPI wrapped 29 fps file to open when framserved by Vdub..they both framserve fine from Premiere though...played with all the settings, still no "dvd like quality"....so lets say I do Matrix, which IM sure most everyone has..tell me what settings I need to use in both framserving & CCE, in order to produce this quality, and so that the movie will fit on 2 disks..can anyone do that please??

[nsdn]

hmmm...yep I can confirm CCE and dvd2SVCD can put the matrix on 2 disks.....I just did terminator2 which is more than 120 minutes on 2 90 minute disks....and there is absolutly no artifacts. I do not think that using 80 minute disks would be all that much differant.

I think however that me being PAL and you being NTSC it is very important that you run inverse telecini on your DVD rips....since for the conversion from file 24 fps to 30fps there is a lot of interpolated fields/ frames...plus your having to compress a lot more data @ 30 FPS...let your dvd player worry about the FPS conversion... ( I am pretty sure that most of your players will handle a 24FPS SVCD tho i have to admit i have not read much up on it since I do not have to worry about it).

For PAL of course the differance between 24 and 25 FPS is bugger all and hardly worth doing. I tryed it (the PAL interlaced check in dvd2svcd)...and it seemed to loose a bit of clarity...when i looked at the avisynth script i see why....it does a verticle reduce by 2 (which does the deinterlace) then resizes to the svcd resolution.....i tryed changing this method to double weave but it just seemed to screw things up....anyway it works fine without it so I stopped playing with it. There is a filter around for virtual dub that does inverse telecini for PAL...but when i frame serve with virtual dub to CCE it goes about 30% the speed of avisynth....so it would take a day and a half using 4 pass VBR....stuff it....it works great anyway I am not that much of a perfectionist.

Another thing I noticed is that if the film is in 2.35:1 u can get more on a disk..simply because there is less active picture ....if u convert 2.35 to 16:9 (which I did for the 13th floor with FANTASTIC results) your using more bitrate....although I also did Jurassic park and the Lost world over the weekend and they are @ 2 hours in 16:9, but u need at least 4 pass VBR to get a near DVD result...I did these at 3 pass...and i can see slight ghosting ..thats being very very picky most people would never see it.

Anyway sorry for being so verbose....but thats about it....

side note....although the results from CCE have far surpassed TMPGE for dvd ripping.....for my DV footage.....I have not been able to get a near dvd result....dunno why yet i am still playing with it...but even setting the average bit rates well above what I have seen it use for DVD rips i can see artifacts with a lot of movement....trying deinterlacing and all that crap..TMPGE still rules tho for functionality and ease of use...the fact that CCE does not handle audio really sucks..cya

[nfl2k2]

I put the average bitrate and max bitrate in dvd2svcd the bitrate calc told me too (even subtracted 20 off that), and still it came out to 3 cd's. the movie is 104 minutes long. I know it's some stupid little thing, but can someone tell me exactly how to make dvd2svcd make me 2 800mb files?

[kqt]

My original problem was probably due to the bad quality of the (bootleg) DVD I was trying to rip. Crap in... crap out!I retried with both the Matrix and Patriot and got pretty good results with 3-pass VBR. Like nfl2k2, I'm also interested in how to set DVD2SVCD to get the Matrix onto 2 800Mb CDs. Thanks in advance.

[Kdiddy]

Thanks NSDN for the informative post but I think you missed my point totally, I KNOW MATRIX can fit on 2 disk, I have done it and other movies several times...I know about the IVTC & all,...what I want to know is EXACTLY what settings in your framserver & CCE that you use to obtain this holy grail of "dvd like quality"...because I always get more halos/ghosting whatever ya call it with CCE 4 pass than I do TMPG 2 pass....so Im looking for the experience CCE guy to tell me why??.

[Truman]

I use dvd2avi, then AVIsynth/MPEG2DEC to resize, add borders and frameserve to cce. For encoder settings I recommend the cce manual, everything is explained quite well.

I guess, the "ghosting" depends on the resize filter. If you stay with Vdub, try "bicubic resize" instead of "precise bicubic resize" or add a blur filter (yes, have some more tests )

btw, you don't need to convert Vdub frameserved files with VFAPI. CCE accept *vdr.avi files. Or did I misunderstand this?

[Truman]

I forgot something to mention. "DVD like quality" you can only have, when you copy the VOBs. I am pretty sure, you don't expect exacly the same quality like DVD. How could a min500/ avg2000 / max2600-MPEG-2 look like a min1000/ avg5000/ max9800-MPEG-2?
Well, I think my (X/S)VCD have very, very good quality. It is very clear, very sharp, short said: perfect made (X/S)VCD. But it isn't "DVD like..". This is impossible. Yes, many people talk about "dvd like quality" results with cce. I tended to do that too after the first succesful conversions, because I have been very impressed. But honestly, talking about dvd quality if you mean svcd quality is bulls***.

[Kdiddy]

finally a good honest response..thanks Truman, yeah I didnt expect "dvd like" at those bitrates either, but I keep seeing all these post about it... yeah ya did misunderstand that one part...when I convert my 29.97 fps "d2v" file with VFAPI, in order to have a fake AVI, I then open Vdub and try to framserve it to CCE, but wont accept it, CCE also wont accept that fake AVI when directly opened as well...

I will try your method though...but in your honest opinion, those this method considering all the steps, does the quality output improve that much over TMPG to prove worth...dont factor in speed, because it is not an issue with me, ONLY quality..thanks

[Truman]

I did not compare TMPG's MPEG-2 to CCE's MPEG-2 yet. If you feel like do that, please report your results here.

CCE wont accept a *vfapi.avi, but VirtualDub accept the it? That is strange..
You need cce sp 2.50 to encode, v.2.62 doesn't accept Vdub's and AVIsynth's files.
Maybe you should reinstall the VFAPI codec.
You can also try to feed any avi file into CCE and edit the .ecl (cce-project). Have a look at my reply here.

[nfl2k2]

could someone still tell me why i have a 3 800mb files? thanks

[Truman]

I don't use DVD2SVCD, so I can not give you the settings. But I calculated with FitCD. To fit 104 min on 2 CDs, set 1868 kbps average video bitrate and 224 kbps audio. (or 1963 + 12
If you want to add menus or something like this, substract few kbps.
Do you really got a 3800 MB file? That would mean, you didn't end with 3 but 5 CDs...

[Kdiddy]

Well I reached my conclusion, and IMO with my eyes, on my DVD player to my TV, (a lot "my"s in there), TMPG produces a better quality product than CCE, I love CCE's speed, but it just wasnt enough...it wasnt a runaway by any means, but I could notice enough of a difference to convince me that TMPG was the route I should go.....I should note, on my PC, they are virtually identical, but I watch all my movies on my standalone...so some factors could come in to play with my standalone...

[Truman]

Oh, comparison finished alreaddy? Anyway, I am interested in the results. I assume, the source was DVD. How did you frameserve to CCE? Do you use cce sp 2.50? If you don't mind, please post the .ecl file here or email me.
I am going to compare both too, but this will take 2 or 3 days.

[Kdiddy]

yeah I was only using 2 min clips, so it wouldnt take all night, then finished this morning...I frameserved CCE 2.5 from Vdub & frameserved CCE 2.6 out of Premeire 6.0...those 2 pretty much produce same results..sorry didnt save th .ecl file...but Ill email you my exact settings & a TMPG template when I get the chance..prolly later on today

[nfl2k2]

I'll just stick with tmpgenc. The quality is almost the same, plus I know how to use it good and it's faser. thanks anyways.

[Kdiddy]

well the 3+ VBR passes on CCE took the same amount of time as TMPG's 2 pass, so the speed becomes a moot point...then its left up to the quality of both encoders

[Truman]

<TABLE BORDER=0 ALIGN=CENTER WIDTH=85%><TR><TD><font size=-1>Quote:</font><HR size=1 color=black></TD></TR><TR><TD><FONT SIZE=-1><BLOCKQUOTE>
On 2001-09-06 03:25:10, Kdiddy wrote:
well the 3+ VBR passes on CCE took the same amount of time as TMPG's 2 pass, so the speed becomes a moot point...then its left up to the quality of both encoders
</BLOCKQUOTE></FONT></TD></TR><TR><TD><HR size=1 color=black></TD></TR></TABLE>

I did some tests comparing TMPG's MPEG-2 to CCE's MPEG-2.
Source is DVD PAL, progressiv.
DVD2AVI (colorspace YUV 4:2:2, no crop)
Video encoding seperate to have a better comparison.
Resolution 480x576
Audio encoding with tooLAME, muxed with TMPG.
I used a 30min clip
In TMPG I set motion search precision: High quality(slow), 16:9->4:3 resize, VBR2pass, no additional filters
In CCE I used the default settings, but progressiv frames, zigzag scan and disabled the noise reduction filter, VBR3pass
To CCE I frameserved with AVIsynth / MPEG2DEC (resize + add borders, no additional filters)
The result @ min500 / avg 2260 / max 2500 is, that TMPG VBR 2pass gave me a slightly better quality than CCE VBR 3pass.
At lower bitrates the TMPG video get blocky.
@ 500 / 1500 / 2500 the CCE video is still good, of course worse than avg2260, but still good.
TMPG's video @ avg1500 isn't worth to watch.
Conversion time comparison: (30min clip)
TMPG VBR 2pass: 4:05
CCE VBR 3pass: 3:03

<font size=-1>[ This Message was edited by: Truman on 2001-10-09 04:57:12 ]</font>

[Kdiddy]

Ouch Truman, talk about an a$$load of information...LOL, I did some more test with method you sent about saving a .tpr project in TMPG, VFAPI wrapping it, and encode with CCE, and was quite surprised...I may have to change my stance..

now again, there were SLIGHT differences, to the common viewer, I pooled both my neighbors, and they said that they thought they were the same...but to me, I saw a lil bit better in CCE...

I did 3 different movie parts, of about 5 min clips...and on each clip, CCE's mpeg2 3pass, progressive, 1 complex, no noise, 4:3DAR, DC 10, supposely TMPG12i mattrices (min-1500, max 2000, avg-1800) VS. TMPG's 2 pass basically same settings with a lil brightening/saturation thrown in...same bitrate settings....and yes CCE was about 20% faster on all times...

So now I lean toward CCE purely based on speed since quality is about the same...I'll assume the fact that CCE is being framserved is what slowed it down some in my previous tests...I was even surprised with its 2pass & CBR quality...on my son's movies, I might be doing with those settings....now my big questions are...when I save this TMPG project file...EXACTLY what settings besides the resize, get passed on to CCE??..do the DC, YUV, & other advanced options get forwarded as well???..also, when I did the patch for the matrices, it never said ok or anything, so do I just assume that it is using the matrice I told it too??..cuz I reeally didnt see a difference when I changed it...tand alos, how much affect does packet size setting in CCE have??...thanks for all ya input..

[nsdn]

Sorry did not read this till now...

Kdiddy, sorry about missing your point...I guess my point was I really did not do anything special with dvd2svcd to get the results I reported. I was trying to think of why you were seeing differing results to mine.

My settings I use (from memory)
Using 90 min CDR
In bitrates select number of disks to 2 for all run times…
Min bit rate is 500 min average is 1000,max average is 3000 max bit rate is 3400 (my dvd player gags at 3500)
For the audio I do not resample to 44Khz ..leave it at 48….resampling takes forever and the amount of space u save is minimal…set audio bit rate to 128….

I sometimes play with the avisynth script but normally leave it alone…..only if I want to resize from 2.35:1 to 16:9.

for nflk probably the reason you still end up with 3 disks is that u have not lowered your MIN AVERAGE...i got caught with this at first as well.

The only thing is that for lower bit rates like for getting the matrix on two disks you really need to use 4pass or more. With 3 pass you can see ghosts and small amounts or artifacts...4 pass they are virtually not visible. For shorter movies on 2 disks u can get away with 3pass.

to Truman, your technical explanations are impressive, however, having worked in television, both in master control and broadcast camera support, I learnt that there is a balance between what is subjectively acceptable/same and what is technically better. Looking at the SVCD coming out of CCE 4 pass, subjectively there is little difference with DVD…at least on my TV and DVD player…that is not to say that there might be a more visible difference on someone elses equipment, however it looks damm close to me on mine.

I will reiterate, particularly for longer movies …fitting onto 2 disks, CCE eats TMPGE subjectively speaking, PROVIDED you use MORE than 3pass VBR however as soon at you use 4-5pass VBR, of course u loose your speed advantage with CCE.

To be honest I do not find DVD particularly good anyway since there is a real lot of variety in the quality of encodes….you can “see” the difference…..though admittedly if you get a DVD that has been encoded well…..then the quality is fantastic, regrettably it seems that as often as not the encode is crap in the first place in spite of all the it rate they have to play with

[Kdiddy]

No prob man...stil find it weird that framserve was a slower & worse process than using VFAPI to make an fake .avi out of TMPG project file..then encoding using CCE.

[Truman]

@nsdn, I agree with you in all points. The only thing I did was to compare TMPG VBR 2pass to CCE VBR 3pass. I posted the .mcf and .ecl files (well it isn't a technical explanation), because I was hoping, someone will correct my faults (there are always some).

@Kdiddy, First I want to say, That TMPG method isn't the fastet.
Here a time comparison, in all cases I had to resize 16:9 and I didn't used any other filters.
VirtualDub frameserving to cce: 0.320 (precise bicubic resize)
TMPG->VFAPI->CCE: 0.430
AVIsynth/MPEG2DEC frameserving to cce:0.640 (precise bicubic resize)(keep the video in YUV2 colourspace)

The following TMPG settings get passed to CCE. Resolution, aspect ratio, frame rate and all filters, nothing else.
When you are creating the TMPG project, save the template and compare it with cce's encoder control list.
<TABLE BORDER=0 ALIGN=CENTER WIDTH=85%><TR><TD><font size=-1>Quote:</font><HR size=1 color=black></TD></TR><TR><TD><FONT SIZE=-1><BLOCKQUOTE>
when I did the patch for the matrices, it never said ok or
anything, so do I just assume that it is using the matrice I told it too??
</BLOCKQUOTE></FONT></TD></TR><TR><TD><HR size=1 color=black></TD></TR></TABLE>
yes
<TABLE BORDER=0 ALIGN=CENTER WIDTH=85%><TR><TD><font size=-1>Quote:</font><HR size=1 color=black></TD></TR><TR><TD><FONT SIZE=-1><BLOCKQUOTE>
I reeally didnt see a difference when I changed it.
</BLOCKQUOTE></FONT></TD></TR><TR><TD><HR size=1 color=black></TD></TR></TABLE>
well, as I used cce 2.62 I played around a bit with the matrices, but also didn't see a difference. Maybe anyone else do.
Actually I am hoping someone else will figure it out for me.

I never touch the packed size in cce, I leave it always 4096 without probs.

[Kdiddy]

Well I never use avisynth so I cant speak for it...Vdub on was slightly faster maybe a min or 2 for me...but when you are talking about hours of encoding time..whats a few minutes....also the quality of what Vdub & Premiere gave me is what I most disappointed in, not time...and quality is what Im after more than time...so I think I gonna stick with the TMPG->VFAPI->CCE format...because this indeed does produce a VERY "dvd like" quality, for me....so again, the settings from all TMPG tabs, ie, "Video, advanced, GOP, & Q matrix" get pass through or just the video tab??

[Truman]

<TABLE BORDER=0 ALIGN=CENTER WIDTH=85%><TR><TD><font size=-1>Quote:</font><HR size=1 color=black></TD></TR><TR><TD><FONT SIZE=-1><BLOCKQUOTE>
....so again, the settings from all TMPG tabs, ie, "Video, advanced, GOP, & Q matrix" get pass through or just the video tab??
</BLOCKQUOTE></FONT></TD></TR><TR><TD><HR size=1 color=black></TD></TR></TABLE>
Size, aspect ratio and frame rate (video tab)
All tha filters like source range, noise reduction etc.(advanced tab)
That's all.
Bitrates, GOPs or Qmatrices wont get pass through.

[Kdiddy]

cool, thanks a lot man!