VBR vs CBR

[Huit]

I know this question has been asked before but it never really answered the question I have about the difference between VBR and CBR. My question is this: if I would encode an SVCD (in CCE) at 3pass VBR with the avg bitrate and the max bitrate BOTH at 2496 (so the bitrate is actually constant isn't it), would that give the same result as when I would encode it at 2496 CBR?

[Truman]

Take a short clip and encode it with both settings. After that view the result with bitrate viewer, compare GOP by GOP. I guess, there is only a slight difference, probably the bitrate of the VBR3pass encoded file will be more constant than the CBR encoded file.
Then please tell us the result.

[Huit]

Unfortunately my DVD-drive is away for repairs . A friend of mine asked me to help him rip a DVD and then this issuw came up. He wants to rip an episode of Fawlty Towers and it's about 35mins. He only wanted 1 episode so if he used VBR he could use the max bitrate to fit it on 1 CD. I told him I thought he could just as well use CBR if the avg bitrate = max bitrate. When I have my drive back I'll try and let you all know.

However, maybe someone already knows the answer to this. I reckon the bitrate per frame will be (almost) equal for both methods. But is it possible that more bits are allocated to parts of a frame that are part of a (fast) movement? And would this be done better if a multi pass VBR was used? I guess if that's the case then a multi pass VBR could be better than a CBR.

[vitualis]

<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 2002-01-14 03:26:52, Huit wrote:
if I would encode an SVCD (in CCE) at 3pass VBR with the avg bitrate and the max bitrate BOTH at 2496 (so the bitrate is actually constant isn't it), would that give the same result as when I would encode it at 2496 CBR?
</BLOCKQUOTE></FONT></TD></TR><TR><TD><HR size=1 color=black></TD></TR></TABLE>

Theoretically yes. If the max and avg bitrate of a VBR stream are the same, then it is equivalent to CBR at that bitrate.

The results should be the same except that it would take 3 times as long to encode.

Regards.

[Huit]

Thx for your replies. As soon as I have my drive back I will do some tests as Truman suggested.

[KoalaBear]

My question is 'why'?

What possible benefit can be derived from a 3-pass CBR?

<font size=-1>[ This Message was edited by: KoalaBear on 2002-01-15 11:13:58 ]</font>

[Huit]

no no, it was a 3pass (or whatever number) VBR vs a CBR

<font size=-1>[ This Message was edited by: Huit on 2002-01-15 12:10:09 ]</font>

[KoalaBear]

I guess I don't understand the question. Are you asking what the difference is between CBR and VBR in general, or within the context of your example?

[Huit]

Within the context of my example. The 1st part was answered: if for a (multipass) VBR the avg bitrate = max bitrate, this would be the same as a CBR with that same bitrate. However, I still wonder if CCE (or any other encoder) can allocate more bits within a fram to parts that are part of a fast movement. And if this is the case, if this allocation of bits based on movement detection works better when doing a multipass VBR instead of a CBR.

[kinneera]

Perhaps its just wishful thinking, but wouldn't VBR at a constant birate allow the encoder to do smarter motion searching and thus code more motion vectors, leaving more data for iamge information with each successive pass? That is to say, by improving the efficiency of the highly compressed P and B frames, more information is available for describing the actual image?

I more or less did this experiment with a short movie, and the CBR output exhibited more blocks than the VBR output. My question, though, is whether a decoder will interpret a VBR stream with the same min/max/avg values as a CBR stream, or whether it will still treat is as VBR, causing compliance issues.