avi.NET single pass vs multi pass question

[SonofPhatty2x4]

I use avi.Net 2.2.00 to convert all my movies and I love the product.

I don't understand something about the conversion process.

Let me give an example.

I am converting Halloween H20 over to xvid.

For both tests I use the deinterlacing and sharpen filter options (I find that these make the converted movie more pleasing to my eye)

If I use the quality pass conversion set to 2.5 while keeping the original AC3, I will get a xvid file roughly 800 mes in size and bit rate 1000.

If I force bitrate to 1350 while using the AC3 file, I will get and xvid rougly 1100 megs (I don't care about size, just as long as the bit rate remains a constant 1350).

Now here's the confusion:
The xvid created through the multipass conversion that has the higher bit rate (1350) is not as pleasing to the eye as the single pass quality option.

Why? I am confused by this. Shouldn't the xvid with the higher bit rate be the one most pleasing to my eye?

[CWR03]

If the one with the lower bitrate was a multipass, it's quite possible for it to look better. A multi-pass encode can spread the bitrate out where it's needed, giving less to still or slow-moving as well as darker areas. With single pass, each frame gets the same treatment. You'll probably notice that still parts of your single-pass encode look good, but camera pans or high-motion parts (especially if there's smoke, fire or fog) will be very blocky.

[SonofPhatty2x4]

If the one with the lower bitrate was a multipass, it's quite possible for it to look better.

No. It's the single pass option that looks better.

That's what confuses me. How does the single pass with less bitrate look better (to me that is) than the 2 pass with a higher bitrate?

[weaver4]

Single Pass constant quality will spread bitrate around as well. It is not a constant bitrate but a constant quality. If it is working with fast moving action it will use more bits for those frames.

On a two pass the encoder needs to determine what an average bitrate needs to be to achieve the size that you want then make a decision where in the content it needs to add bits and where it needs to take them away. Sometimes that algorithm is good and sometime it is not so good.

I always use a constant bitrate to get the best overall picture. Let the coder figure out how many bits it needs for each frame to get the quality that you want.

I have seen some movies that are excellent quality at 700kbs and others that need 1600kbs for no obvious reason. I don't know how one would make a decision on what they should use; so I let the coder figure it out by using single pass.

[SonofPhatty2x4]

On a two pass the encoder needs to determine what an average bitrate needs to be to achieve the size that you want then make a decision where in the content it needs to add bits and where it needs to take them away. Sometimes that algorithm is good and sometime it is not so good.

:thanks:
Yeah - I think understand what you have said.

So basically, even though I have selected 2pass with no size limit, parts of the movie can get garfinkled (notice the teknikal talk) with the encoder when it attempts to balance all the bits to keep the rate constant.

Conversly, the single pass will use what it needs to maintain a given frame with in the context of the quality ruleset.

Now that I write that out, that does seem to be a more logical choice, especially if you are not to worried about file size.