VCD vs low bitrate XSVCD

[Azathoth]

Greetings all:

I was hoping someone could lend their opinion on this matter. Would encoding an XSVCD with the standard SVCD resolution but a bitrate equal to VCD actually produce a better looking output then a standard VCD?

[Mirror_Image]

Try it with a small file and let your own eyes be your judge - you can trust them.

[homerpez]

Well, IF you use CCE (Cinemacraft Encoder) to encode the SVCD as MPEG-2... then maybe you can try it.

I can personally use CCE and with everything going for it (not an action film, letterbox, pulldown), I can use AVG bitrates as low as 700-800, and 128k audio, and it looks MUCH better than a VCD at the same bitrate (in some cases, almost as good as the DVD).

But again, this is with EVERY factor going for it - CCE, lower # of frames/sec, less info to encode because of black bars, and not as much motion to get blocky/grainy...

The difference you will notice between a VCD and an "X"SVCD of the same bitrate is that the SVCD will look "better" in most places, because of the extra vertical resolution (makes it sharper). But, ultimately, you HAVE to use CCE and you MUST use enough bitrate, or SVCD will look crappy.

So if you only have TMPGEnc... stick to VCD.

[homerpez]

Oh, and the main reasons for using CCE, if you're not familiar... Multipass encoding (3,4 or even more pass, vs. only 2-pass for TMPGEnc), superior VBR, and accurate bitrates...

I set mine at ~ 400 MIN, ~700-1100 AVG (depending on the size per disk I need), and 2520 MAX (to make it stay a "compliant" SVCD).

Hope something there was helpful..

Homerpez wrote So if you only have TMPGEnc... stick to VCD

bullshit, Tmpgenc also does excellent quality X/svcd`s

cheers

[energy80s]

MPEG2 is crap below 2 Mbps. If you want good quality low bitrate files use MPEG1. That's what Kwag has done with his templates. MPEG2 is only meant for bitrates of 2Mbps and above.

[Olli]

I get excellent results creating XVCD's

- 480 x 576 resolution
- Bitrate of average 2000, max 2300

My standalone DVD does not play SVCD's so my XVCD is virtually an SVCD except for the slightly lower bitrate and the fact that it is MPEG1 rather than MPEG2

TMPGEnec does a great job with the s pass VBR so I don't feel a need to change to anything else

[SatStorm]

The difference between CCE and TMPGenc in quality ain't that big anymore...
Less than 8% I would say....

The problem with TMPGenc is the speed....

[sherman]

I have to agree with Olli I also always use XVCD (480X480 for me in Canada). Bitrate Max 2500 Ave 1800.

I get excellent XVCD from Satalite TV captures.

Standard VCD is crap. XVCD or SVCD is the way to go. One advantage of using XVCD with MPEG 1 is that the TMPGEnec trial period doesn't expire on MPEG 1.

Sherman

[SatStorm]

The basic minus of mpeg 1, is that is only progressive. So, you have to de-interlace your source (if it is interlace, much common in europe)
That means you loose quality, wanted or not!

About the mpeg 2 trial fuction of TMPGenc, there are semi-legal ways to keep the limit far away! Living in the future for example
Don't mention the updates! (once a month at least!)

[banjazzer]

The problem with TMPGenc is the speed....

Only really a problem if you insist on using 2-pass VBR. CQ gives equal or better results in less than half the time, but you do need to get to know how best to predict your filesize.

[banjazzer]

So if you only have TMPGEnc... stick to VCD.

Absolute nonsense. I'll bet my SVCDs using TMPGEnc's CQ mode are every bit as good as yours - maybe even better. 8)

The original ? from Azathoth

Would encoding an XSVCD with the standard SVCD resolution but a bitrate equal to VCD actually produce a better looking output then a standard VCD?

If your using Tmpgenc and VBR 2pass XSVXD(Mpeg2) I think it does look better than VCD, try for yourself. Also if you insist on such low bitrates try changing the res to 352x576.

cheers

[SatStorm]

@ banjazzer
My tests the last 2 and a half years show me the opposite. 2 Pass VBR (multipass VBR) is much better any 1 pass VBR encoding mode (including CQ modes), if you know what to do. So, I am not a fan of any 1 pass VBR mode...
Also, if I want speed AND top quality, I go CBR. Faster and better 1 pass VBR!

[adam]

Quality is relative. If you encode a SVCD at VCD bitrates then you will have significantly less bits per pixel, even with the benefit of VBR encoding. Picture a VCD encoded at like 600kbits and that's what it will look like. However, due to the higher resolution the picture will be sharper so which one looks better will be up to you and your eyes.

Now if your willing to pit XVCD against SVCD at low bitrate then that is different. Theoretically at the same settings mpeg2 and mpeg1 should look identical. Realistically, at least from my experience, mpeg encoders are optimized according to the typical encoding methods of most users ex: mpeg1 for 1.5mbits and lower and mpeg2 for 1.6mbits and higher.

So at lower bitrates mpeg1 will actually perform better in most cases.

Mpeg2 has some inherant advantages over mpeg1. Most of them are minor but the main ones are support for interlacing,support for overlay graphics such as subtitles, and finally support for a second audio track.

If you don't need these and your willing to raise your bitrate above the 1150kbits required for compliant VCDS than your better off sticking to mpeg1 and making an XVCD. Actually, in my opinion, your better off making a standard vcd then making a SVCD at 1150kbits avg. I just don't think thats nearly enough bitrate to use for a resolution that large.

[banjazzer]

SatStorm
With TMPGEnc, my tests, (admittedly not over 2 1/2 years) show me that CQ is every bit as good as 2-pass VBR given equivalent filesizes and equal min and max bitrates. And a damn site faster. Some would even go further than this. 8)

[SatStorm]

Your opinion, my friend, and I truly respect it!
It doesn't mean that I have to adapt it!
I have my opinion, and I think is well respected also!
For me, multipassVBR is better and nothing can change my mind after 2 and a half years of tests. I don?t care to test further for this subject, I am very pleased with my results. I only update my tests once in a while, when I notice that TMPGEnc has dramatically change in quality reasons. Last time I did that was with tmpgenc 2.52 plus, I?ll do some test this weekend also with 2.57plus (which I find some wierd quality issues again...)
Once in a while I also try and test alternatives like Kwag?s templates for example. Next week I?ll try also that new KDVD template to see the results. Kwag have good ideas and loves the subject, so I am very positive to his job.

About the 1pass vs multipass ?fight?
The only thing we - the 2 different parties- can do, is to post our thought/results/methods/conclusions so other users choose what suits them most. Nothing more...

The conclusions in this subject, is like adam say: " relative"

For me (and others), multipass VBR is better onepass VBR.
For you (and others), onepass VBR is better multipass VBR

Also, I suggest to start using terms like "In my opinion", "My conclusions", "My advice" and "I belive", any time we posting something for this subject! That way we don?t start fights all the time for no reason.

Finally, there are a great other post for this subject, the "battle of CQ and VBR modes?. I think is better posting those things there and not here!

[banjazzer]

Your opinion, my friend, and I truly respect it!
It doesn't mean that I have to adapt it!

Absolutely. We all do what works best for us. 8)

For me (and others), multipass VBR is better onepass VBR.
For you (and others), onepass VBR is better multipass VBR

I think it's important to specify which onepass VBR we mean - there are 3! I am talking about CQ versus 2-pass VBR. Also, after weeks on "battle of CQ and VBR modes?" thread, I seem to recall very few claiming that 2-pass VBR was actually superior (quality-wise) to CQ, at least to far as TMPGEnc is concerned. The consensus (if there can be such a thing) seemed to be that CCE multipass VBR was better than any TMPGEnc mode, and that possibly TMPGEnc's CQ was better than its 2-pass VBR. I cannot find fault with either, used properly. I favour CQ because of the lengthy time involved with 2-pass VBR, and I've got reasonably adept at predicting filesize.

Finally, there are a great other post for this subject, the "battle of CQ and VBR modes?. I think is better posting those things there and not here!

Well, that thread did a get a little bit long, and became confused because people started talking about different encoders, and non-standard templates! However, I think when idiotic comments such as "So if you only have TMPGEnc... stick to VCD" are made, it is important not to let such misinformation go unchallenged. 8)

[SatStorm]

Yeah, that's the point!
In Between advance users, there is no need for challenges, only for propositions! Just say your opinion and your conclusions and let the judge come from the user who interest to test!

About CQ/multiVBR
In my opinion, the CQ mode is a good one. It is like CBR in quality most of the times in smaller filesize. It is good when you use it with movies, but there is not the best option for VHS source and Music Videoclips in general.
For that situation helps TMPGenc: The CQ fuctions don't have any kind of bugs. Why? Ask Hori San! Probably he believes that the great challenge for a programmer, is to success the best possible quality in 1 pass, and in a way that is right. So, he implements this fuction all the time, while with multipass well.... Let say he don't like multipass VBR himself (as he don't like AMD CPUs in general!)

The true is that TMPGenc's 2 Pass is not yet perfect . Needs work and on many tmpgenc versions, is broken. This new tmpgenc version, the 2.57 plus, has some implement on this. My first tests didn't show me something positive or negative. I find the results equal as before. I have to test further to see if something really changed.

The experts which state that multipass VBR is in theory and praxis better any onePass VBR, are based on media and lab equipment we don't have. Judging that stuff based on encoders like TMPGenc and CCE is silly IMO. Just try to succeed the best possible with any method and share your success with other users! It is the only thing it counts!

[homerpez]

I guess I have to bring a few dolts up to speed (your responses earn you the name, guys)...

I find, for my purposes, TMPGENc is useful to me ONLY to make a project file, resize. ONLY Cinemacraft Encoder gets results.

First, I can calculate the time/size of a movie, and fit it EXACTLY onto ONE disk, within 1-2 MB space. Try that with CQ!

Second, How many passes can you do with TMPGEnc? Two?

Cinemacraft lets you do up to 9 I think, only 3 is necessary (I do 4 just to be sure). It completes a "4-pass" in the same time as it takes TMPGEnc to do 2.

And, when you do "4-pass" in CCE, it's really doing 5 passes (1 info pass, 4 encode passes)! TMPGEnc's "2-pass" is really only doing ONE excode pass, one is the info.

CQ to me means "Constantly Bad Quality". Multipass definitely does a better job at making a whole movie use space efficiently, while also maintaining quality.

Plus, CCE is STILL better quality than ANYTHING TMPGEnc has put out - believe me, I've been hoping they'd get on the ball...

I will concede ONLY that if you DO use TMPGEnc, that CQ will LOOK better that it's 2-Pass. I did try this a couple times, and TMPGENc does make a decent CQ, but NOTHING compares to CCE...

Sorry guys...

I guess so long as you're happy...

[homerpez]

And I only post this as STRONG hints that people change their evil ways.

But, then again, if people are happy...

[next]

poor homer is still at it...extremely passionate with his view of the moment (which i admire). only problem is that the moments change quite often.

[banjazzer]

Never argue with deeply religious people. When their blind faith finally goes - they go nuts. 8)

[bilbogod]

I was just curious Homer............What Res are u using?? Cause unless you are using 720 x 480 (576) I have never seen the CCE multipass(DVD2SVCD) match TMPGenc's CQ at DVD RES (2cds- KVCDx2). I have over 80 movies used via CCE that look excellent, surpassed only recently by the last 8 that I have done using CQ + dvd Res + resizing etc via TMPGenc. Oh and yeah.....as mentioned in another thread using DVD2SVCD's 4 test encodes only 1 out of 8 ever went oversize. The most inefficient encode leaving 30 megs of free space between 2 cds. A very small trade off ..............