CBR vs 2-pass VBR

[slacker]

Read quick...

I read in a thread this week (can't find it now!) that if your bitrate is high enough (8 mbps or higher) you don't need to use vbr and that you would achieve better results using cbr for typical home videos. Any truth to this?

[JoRodd]

VBR is your best bet. VBR gives you a variable bitrate but consistent quality. CBR gives you a consistent bitrate but variable quality. Also, some DVD Players have difficulty with playing a constant bitrate.

Remember, not all aspects of your video will need the same bitrate. More motion needs higher bitrate thereby insuring better quality. Low motion or blank screens do not require high bitrate.

[44echo]

slacker,

I read the thread that you mention. The bottom line was that if space is not an issue, then using CBR is not a problem. Just make sure your bitrate isn't too high to cause some players to choke.

[slacker]

Most of the threads I have read deal with how to set your encoder settings when dealing with space issues, encoding time and sometimes the nature of the video. There is NOT much about what works BEST at higher bitrates, i.e. 8.5 mbps or higher. Does vbr and cbr output the same quality at these high rates?

My question really is...

At say 8.5 mbps or higher what yields the best results? Or is there even a difference? Since I'm only putting an hour of video on each dvd, space is NOT a problem. And encoding time is a neutral issue for me as well. I'm simply looking for the BEST quality encode using the MainConcept encoder included with Vegas.

My own empirical evidence says they are the same. But I was hoping for some other opinions out there.

[SLK001]

If you only have an hour of video, 8.5Mbps will give you pretty much the best result. Multiple pass VBR (with the same MAX BR as your CBR) will also give you pretty much the best result. In other words, the two methods yield pretty much equivalent results. But VBR has to pay an encoding time penalty for the same results.

So, go ahead and encode at 8.5Mbps CBR. You can't go wrong.

[dipstick]

When using bitrates that high, it's best to use CBR. Not only is it much faster, but it will control max bitrate much better than VBR. With VBR encoded @ 8.5 avr, the max bitrate could very well shoot up beyond 10 mbs causing playback issues with some players. The CBR encoded at the same 8.5 mbs, should keep the peaks below the danger zone.

For the record, I usually encode @ 8.5 mbs CBR. Quality is my main concern not file size. At those high bitrates, you may want to experiment encoding with little or no B frames and increase the P frames.

[edDV]

If you are talking about a high quality (e.g. DV input home videos) and can limit to ~60min per DVD then 8.0-8.5 Mb/s CBR gets you there with efficient encoding speed and predictable space usage.

VBR is used to get more minutes of video into the same target space. VBR requires peak, average and minimum data rate to be specified. Peak rate is limited to your (or target) DVD player's capability. The DVD spec allows this to go as high as 9.8 Mbps (including audio) but hardware decoders may struggle with video at that rate. Conservative maximum rates fall in the 7.5-8.5Mb/s range for peak or continuous decoding. So if you specify VBR 8.5 max and 8.5 average you get the same theoretical result as CBR. If you Spec something like 8.5 max, 7.5 average and 4.0 min, you might be able to squeeze in more minutes by reducing bitrate in low motion scenes but at a cost of encoding speed and space prediction. A 2 pass VBR algorithm will force fit the average data rate at the cost of a second pass and possibly quality.

Bottom line. CBR is the most direct route to high quality

[jimmalenko]

The DVD spec allows this to go as high as 9.8 Mbps (including audio)

9.8Mbps is video only - it's 10.08Mbps including audio.

[edDV]

The DVD spec allows this to go as high as 9.8 Mbps (including audio)

9.8Mbps is video only - it's 10.08Mbps including audio.

True.

[slacker]

Thanks all!

dipstick,

Is there a paper on how, when and why to juggle B and P frames? Using the DVD NTSC rendering profile in Vegas the I-frame value defaults to 15, the B-frame value defaults to 2, and no where can I find the P-frame value.

[SatStorm]

In Theory, a multipass CBR does a better pixel allocation per frame, even on higher bitrates.
IMO it is a waste of time.

VBR vs CBR:
- If the maximum value of the VBR file is set to equal or less bitrate than CBR's bitrate, then the CBR file looks better
- If the average value of VBR is the same with the CBR bitrate, the VBR may look better.

[edDV]

Thanks all!

dipstick,

Is there a paper on how, when and why to juggle B and P frames? Using the DVD NTSC rendering profile in Vegas the I-frame value defaults to 15, the B-frame value defaults to 2, and no where can I find the P-frame value.

The Vegas 8-8-8_CBR_224 mode is optimized for DV to DVD MPeg2.

The Render As - Custom - Advanced Video settings gets you into micro adjustments.

[bugster]

but hardware decoders may struggle with video at that rate.

The hardware decoders in all DVD players will happily decode 9.8Mps all day long. They have to, its part of the spec. It is the fact that some (perhaps many) players get more read errors, and therefore have to do more read retries, on recordable media than pressed media that they may struggle to cope with higher bitrates on DVDr's.

But as to the actual question, if space/filesize is not an issue, high bitrate CBR is as good as VBR.

[slacker]

BTW,

is the 8.5 mbps video bitrate recommendation for lpcm or ac-3?

[edDV]

but hardware decoders may struggle with video at that rate.

The hardware decoders in all DVD players will happily decode 9.8Mps all day long. They have to, its part of the spec. It is the fact that some (perhaps many) players get more read errors, and therefore have to do more read retries, on recordable media than pressed media that they may struggle to cope with higher bitrates on DVDr's.

But as to the actual question, if space/filesize is not an issue, high bitrate CBR is as good as VBR.

It all depends on whether one focuses on the choppy input to the buffer (caused by read retrys) or the inability of the hardware decoder to empty the buffer with burst decode rates over 9.8 Mb/s the result is still gaps in playback.

[slacker]

edDV wrote

The Vegas 8-8-8_CBR_224 mode is optimized for DV to DVD MPeg2.

I can't find it as a template or referenced in any settings in Vegas 6.0c! I'm gambling that _CBR_ = Constant Bit Rate and _224 = AC-3 bitrate. What is 8-8-8_?

[edDV]

BTW,

is the 8.5 mbps video bitrate recommendation for lpcm or ac-3?

I think you would look at total (audio+video) bitrate for the DVDR issues bugster describes. The issue is allowing for recovery tolerance for bad read retries. Both the MPeg2 and AC-3 hardware decoders will keep up within spec.

Key issue is which player will be playing the DVD. If you have the player you can test its limits. If you are sending the DVD to someone else, you should be more conservative.

[edDV]

edDV wrote

The Vegas 8-8-8_CBR_224 mode is optimized for DV to DVD MPeg2.

I can't find it as a template or referenced in any settings in Vegas 6.0c! I'm gambling that _CBR_ = Constant Bit Rate and _224 = AC-3 bitrate. What is 8-8-8_?

Maybe I created that one or it was updated from v5.

my settings are 720x480 15 I frames 2 B frames
8Mb/S CBR

Added.
Just looked a "DVD Architect NTSC video stream" default. That is VBR (8M max, 6M ave, 192K min).

CBR mode is 6Mb/s default.
Just up that to 8Mb/s and select AC-3 224 and you have my "8-8-8 CBR".

[jimmalenko]

Some extra reading:

https://www.videohelp.com/forum/viewtopic.php?t=265706&highlight=cbr+vbr
https://www.videohelp.com/forum/viewtopic.php?t=265820&highlight=cbr+vbr
https://www.videohelp.com/forum/viewtopic.php?t=272583&highlight=cbr+vbr
https://www.videohelp.com/forum/viewtopic.php?t=273448&highlight=cbr+vbr
https://www.videohelp.com/forum/viewtopic.php?t=274197&highlight=cbr+vbr
https://www.videohelp.com/forum/viewtopic.php?t=274222&highlight=cbr+vbr
https://www.videohelp.com/forum/viewtopic.php?t=274906&highlight=cbr+vbr
https://www.videohelp.com/forum/viewtopic.php?t=275934&highlight=cbr+vbr
https://www.videohelp.com/forum/viewtopic.php?t=278050&highlight=cbr+vbr

[SLK001]

Thanks all!

dipstick,

Is there a paper on how, when and why to juggle B and P frames? Using the DVD NTSC rendering profile in Vegas the I-frame value defaults to 15, the B-frame value defaults to 2, and no where can I find the P-frame value.

I'm not a VEGAS user, but I would bet that 15 refers to the GOP size (or 1 "I" frame for every 15 frames). The number of "B" frames usually refers for each "P" frame (so, 2 "B" frames for every 1 "P" frame).

Most commercial DVDs use a nominal 12 frame GOP, which looks something like this:
IBBPBBPBBPBB

[bugster]

Most commercial DVDs use a nominal 12 frame GOP, which looks something like this:
IBBPBBPBBPBB

A typical GOP for NTSC is 18 frames long and looks like this: IBBPBBPBBPBBPBBPBB

For PAL it is typically 15 frames and looks like this: IBBPBBPBBPBBPBB

Though obviously this will vary throughout the movie. Extra I frames are often inserted at scene changes and chapter points.

[SLK001]

The MAXIMUM GOP structure for NTSC is 18 frames long. Not many DVDs use that structure, but it does vary from DVD to DVD. I haven't looked at the GOP structures of DVDs in a while, but I don't ever recall running across one with 18 frames. When I was doing research as to what the major studios used for GOP length, a length of 12 came up quite often.

But times do change, and I will peruse some of my DVDs tonight for an up-to-date number (as examined by BitRate Viewer).

[slacker]

edDV,

Unfortunately I can't predict which dvd player will be playing these dvds five years from now. I'm assuming players are going to improve, NOT degrade, in quality.

However, I DO know that my current Panny es30vs dvd recorder encodes at 9.558 average mbps vbr with 256 kbps ac-3 in xp mode from cable, hi-8 and vhs, and they all play back flawlessly. So...

[slacker]

Okay, based on inputs, I'm going to go with 8.5 mbps cbr along with 0.448 mbps AC3 for my settings. And somehow will pass around the dvd to friends and neighbors to see how it holds up on their machines.

Thanks all.

[vhelp]

Evening everyone

yea.. sorry for the long post. Anyways..

I didn't see that thread you referenced, but..

First, everything I do operates around a 1-hour video length.
That is the *key* to successful quality.. whatever that is to
each given user.

And Two, I think that everyone is soo per-dvd'r contious, that they
ultimately believe that VBR is the only way to acheave maximum
quality. The problem I have with this, is the length of the video
source being transfered to DVD'r media, and the thinking/planning
that is done to arrive there. But I won't go into it here at this
time. Instead, I'll only say that it would be wise to re-look at
your process and final goal for your video-to-dvd'r. In other words..
think 1-hour videos

One day, you might get it. I have

Anyways..

Personally, I don't necessarily believe in the numbers, when bitrate
preference/settings are mentioned.

One the things that a user has to look out for (or, consider) is the
acutaly MPEG encoder doing the encoding. I don't think there is such
an MPEG encoder that acts and encodes the same way as another given
one.

In my many years experinece with encoding video, I have come to one
final conclusion, when I want the most quality in my transfer of
video from one medium to MPEG.

CBR, and highest bitrate *my* dvd player will allow.
(ie, I prefer a starting setup as, CBR / 9000 bitrate)

Something to consider ...

With differences between various MPEG encoders, the same setup as the
above, when used with more than one MPEG encoder (say, you are testing
which yields greater quality from each trial run) once you author and
burn to (test) DVD, that your given dvd player may not play the same
setup. That is something that *the user* is responsible for, not the
MPEG encoder. Sure, you could blame the dvd player. But the bottom
line here, is user knowledge, (and wisdom) here. With that in mind,
your wisdom move, would be to work around such limitation per dvd player,
and use what is best for *each* dvd player. After all, you could be
sending/sharing your home-made dvd's with other friends that may have
quirky players. So, you work accordingly, and encode per dvd player
Now, you have *maximum* quality when using CBR and high bitrates.

My personal notes regarding VBR ...

I don't incorporate VBR in any of my encodings. After many years of
trial n error tests, and making actual xCD's and DVD's, I have come
to my own final conclusions, and as such, I incorporate CBR / 9000
bitrates in my projects as a minimum. And, the MPEG encoder that I
use with this decision is TMPGenc. Have I use other MPEG encoders
in the past ?? ..yes.

Personally, I don't see the gain (or numbers) that give me the video
quality that justifies VBR over CBR. The source is just not good
enough for my VBR setups (IMHO) when used with the many MPEG encoders
that I have tested. Mind you, this preference with CBR did not come
over night. It took many years for me to come to this conclusion.

Please don't take my comments/notes above as personal injury. I am
not trying to sway you from your current position in this area.
(if you were to ask me, then I would *suggest* to you my preference)

My last closing notes ...

I think that in time, some of you may agree with my comments above,
and practice the setup I noted above, in your on-going MPEG encodes.
I know this day, I would never go back to VBR, when using TMPGenc.
I just won't.

@ slacker

Regarding your last trial test run ...

Perhaps it would be wise to use standard audio setup, before you
incorporate other audio formats. I know that AC3 is a good choice,
but with todays *freeware* I see (through reading) how many users
have problems playing there home-made DVD's on their given dvd players.
I think (theory) that those that tests high bitrates, in the end,
blamed their dvd choking (or whatever test scenario took place) on
account of the Audio. I think that it is worth factoring that into
the *dvd chocking* checks to consider.

In other words, try various audio setups.. ie, MP2 vs. AC3 etc.
If your test dvd players (yours and friends) choke, consider trying
MP2 instead. It may work better with MP2 for instance. And, if
anyone tells you, NO.. you gotta use AC3 or AC3 is the new standard
or whatever, then just tell them to take a hike. Naa.. just kidding

Anyways..

My dvd player is the Apex AD-1500 and so far, I have never had any
trouble with high bitrates. In addition, I use MP2 audio.

To Recap:

** Apex AD-1500 dvd player: I use CBR / 9000 bitrate / MP2 audio @ 192k

-vhelp 3634

[edDV]

edDV,

Unfortunately I can't predict which dvd player will be playing these dvds five years from now. I'm assuming players are going to improve, NOT degrade, in quality.

However, I DO know that my current Panny es30vs dvd recorder encodes at 9.558 average mbps vbr with 256 kbps ac-3 in xp mode from cable, hi-8 and vhs, and they all play back flawlessly. So...

I archive all my DV source and edit material to DV tape for the future.

I also do a DVD MPeg2 8.0-8.5Mb/s w/PCM audio DVD backup of the edit. Distribution DVD copies have AC-3 audio.

[slacker]

vhelp,

That's a post that needs a cup of coffee to go with it! Your logic seems very sound. Makes sense.

The one part I'm not sure about is your take on the audio. I thought mp2 was going the way of the buffalo. I thought the only REAL two standards out there for NTSC were lpcm and ac-3. Am I wrong here?

And if lpcm and ac-3, you would want to go with the highest practical ac-3 bitrate, 448 kbps. No?

[edDV]

And if lpcm and ac-3, you would want to go with the highest practical ac-3 bitrate, 448 kbps. No?

That depends how many tracks and the quality of the source. 224Kb/s is usually fine for 2 channel distribution. I keep masters in raw PCM and enhanced tracks in wav.

[vhelp]

Sorry if I mislead you on the audio part. I didn't mean too.

I was just throwing out audio formats that were at the top of
my mind while writing (typing) the post. And, I was just maying
a suggestion, towards ruling out other factors, when testing
dvd players and putting the blame on the bitrate for instance.

But weather MP2 going the way of the buffalo, is just that, if
you want to rule out using MP2 ..but when you could use
it as *an alternative* and with great results. You could alway
transfer/convert it to another format. But, just remember that
you don't have superiour audio to begin with (not talking about
dvd rips here) when you source is (I assuming) is based on video
capture, and not dvd rips.

The reason why I included (as an example) 192k for my audio, is
because (to be quite honest w/ you here) I don't really notice
much difference with my current dvd player and audio setup.
My setup is a bit cheap.. 20" (sometimes 13" in bedroom) and
tv speaker. I don't have my surround sound system connected, as
that gives my tenats excuse to complain, unfortunately.

I realize that I am short-changing my audio area, but at this time
in my video endeavors, I am satisfied with my audio.

For the time being, I have comprimised with the setup I use in
my video processing and my home-entertainment center setup.
(I hope someday, that will change)

-vhelp 3635

[slacker]

edDV,

That is EXACTLY the "frosting on the cake" COMMENT I needed, and pretty much what I was thinking. Along those same lines, I archived my video to DV tape as well, and was looking for the BEST dvd VERSION of that tape archive. I had originally selected lpcm but have been unsure as to the accompanying video bitrate. Is your DV backup 8-8.5 mbps vbr or cbr?

So really, the end goal is
1. DV tape archive
2. DVD backup of that DV tape
3. DVD distribution copy