does high bitrate affect undersizing?

Will high bitrate cause a need for more than two passes to get to the right file size? Up to this point, I've almost always been able to get there w/ only two passes. I've just gone several rounds trying to back up a high bitrate source (5155kbs) at 2 pass VBR, and cannot get the file size to where it should be. The movie is 'the gray zone' , which has lots of dark scenes, not much fast action. It consistently comes in between 4.05 - 4.10 Gb, and that's with with CCE 2.70, and 2.50. At first I thought it might be RB, then CCE 2.70, but with the consistency of the undersizing with both encoders, I think it may have something to do with the higher bitrate? My VBR settings were at 25-35, and QP at 16. RB was set on 'movie only'. I think I'll go to 3 pass to see if that helps, but it just surprise me to see this stubborn of a source

You said that you are doing movie only, right?
Have you checked what's the original quot;movie onlyquot; size removing the audio stream that you are probably removing?

Maybe you movie just can get any bigger, know what I mean?

yeah, I checked that first Rockas, and it's at the 6 Gb size. So there's room for reduction.

Why not run DVD-RB w/ CCE 2.70 and do just the prepare phase (i.e., abort encode if you operate in One Click mode), copying and pasting the log here. It would be interesting to get a look at the High/Low/Average bitrate and the Q level that the OPV phase will detail. These data might help explain the quot;undersizingquot;.

here's the OPV log -the high/low/average is almost identical:

-----------------
[08:57:00] Phase I, PREPARATION started.
- CCE SP 2.70.1.5 encoder selected.
- quot;CCE Adaptive Quantizer Matricesquot; is enabled.
- quot;One Pass VBR (w/analysis)quot; mode is enabled.
- quot;Movie Onlyquot; mode is enabled.
- VTS_01: 3,418,007 sectors.
-- Scanning and writing .D2V file
-- Processed 156,260 frames.
-- Building .AVS and .ECL files
- Reduction Level for DVD-5: 63.8%
- Overall Bitrate : 5,161Kbs
- Space for Video : 4,106,156KB
- Analyzing VTS_01 for optimal Q factor. -- TargetSize (sectors):2,083,874 -- Sampling 1572 of 156260 frames. -- Predicted size (sectors) at Q=19: 827,562 -- Predicted size (sectors) at Q=3: 1,943,582 -- Predicted size (sectors) at Q=2: 2,113,189
- Q Value selected: 3
- HIGH/LOW/TYPICAL Bitrates: 5,165/5,124/5,161 Kbs
[09:12:49] Phase I, PREPARATION completed in 15 minutes.

Yes the Q is pretty low, but I'm doing a multipass(2) on these attempts. Shouldn't it get closer than what I've been getting?

*EDIT- with high/low/average being so close, would a CBR encode fare any better?

Originally posted by archaeo
here's the OPV log -the high/low/average is almost identical:Yes the Q is pretty low, but I'm doing a multipass(2) on these attempts. Shouldn't it get closer than what I've been getting?

*EDIT- with high/low/average being so close, would a CBR encode fare any better?

I don't know about the others, but I'm confused. You talk about having problems doing multiple passes, and then throw out an OPV(One-Pass Variable) log. Which one are you trying to do? If it's muti-pass, then please post the log from a multi-pass quot

repare Phasequot;. Do you have quot;One Pass w/analysisquot; enabled in your settings?

no, I'm doing multi pass. (I thought that it might get confusing - I only ran the OPV at aaron10's request to get the Q value. Not sure if Q value has anything to do with a multipass undersize, though...)

EDIT: I'll post the multi pass log in a moment, but I imagine the high/low/average will be the same.

multi pass log (the same):

[10:32:49] One Click encoding activated...
-----------------
[10:32:49] Phase I, PREPARATION started.
- CCE SP 2.70.1.5 encoder selected.
- quot;CCE Adaptive Quantizer Matricesquot; is enabled.
- quot;Movie Onlyquot; mode is enabled.
- VTS_01: 3,418,007 sectors.
-- Scanning and writing .D2V file
-- Processed 156,260 frames.
-- Building .AVS and .ECL files
- Reduction Level for DVD-5: 63.8%
- Overall Bitrate : 5,161Kbs
- Space for Video : 4,106,156KB
- HIGH/LOW/TYPICAL Bitrates: 5,165/5,124/5,161 Kbs
[10:44:07] Phase I, PREPARATION completed in 12 minutes.Just to reiterate, I'm getting a consistently undersized two pass encode with both CCE 2.70 and 2.50 on this source. Just wondering if the bitrate levels have anything to do with this. It looks almost like a CBR with the hi/lo/avg spread so close to each other.

If the thread is read in its entirety, I don't think it is confusing in the least.

There's a chance that with a Q level as close to perfect as this is, CCE won't artificially inflate the bitrate just to fill space, since raising it won't actually make the video any better.

It seems it might have something to do with quot;Movie Onlyquot; mode. Have you tried disabling it, and then running a Prepare, to see what the bitrates say? They should not be that close together. But other than that, 5161 kbs is more than enough on a 108 minute movie that was originally released only on a single-sided, single-layer disc anyway, even if it does come out a wee bit small.

spazzHH wrote:
Have you tried disabling it, and then running a Prepare, to see what the bitrates say? They should not be that close together.

Haven't yet tried disabling 'movie only' to see what I get in a whole disc prepare phase. Right now I'm running it in 3 pass to see if it closes the gap in file size. If this fails, then I'll try your suggestion.
aaron10 wrote:
There's a chance that with a Q level as close to perfect as this is, CCE won't artificially inflate the bitrate just to fill space, since raising it won't actually make the video any better.

I know that granularity can affect sizing accuracy for OPV, but does this apply to a multipass, too?.

Update:
Did a 3-pass VBR, and improved the file size to 4.27 Gb (about a 5% increase), which is acceptable to me. But it does surprise me to see something like this that has to be 3+ pass to get it to the desirable size

Anybody else run into a source that required 3 passes or more to get to the right size?

have you adjusted the target size in traget sectors or rb-opt?

dvdrb aims for 4.31, so 4.27 aint far off.

with rb-opt set to 4484mb i have only not gotten exactly 4.37gb once and that was when i used filters.  I use 4-5 pass

jptheripperwrote:have you adjusted the target size in traget sectors or rb-opt?

In RB I've had CCETargetSectors set at 2246000, and was getting outputs consistently sized at around 4.34gb, with 2 pass. It appears that a high bitrate low Q source will require several passes just to achieve the right range in target size.

I think aaron10 is probably right.  CCE will NOT artificially raise the bitrate when the calculated Q is 1...  it woudl be be much more likely to happen in movie-only modes -- only because they are more likely to have the higher bitrate and lower Qs...

The bottom line is:  Does the quality look good at the current size?

I have serious doubts as to whether more passes will change that... or more importantly, whether it should (if it does it sounds like a bug in the encoder to me).  A 5% increase in size isn't necessarily a 5% increase in quality...

The bottom line is: Does the quality look good at the current size?

In a word, yes.
It wasn't so much a matter of squeezing more quality out as it was of figuring out why I couldn't get it closer to the desired size... I was trying to understand the reasons why it wouldn't budge much.

With these explanations, I understand more of of this now. thanks to all who responded