Higher bitrates and shrink or lower bitrate and author?

[Blazey]

I can't seem to decide what is better as both seem to have their own faults in my testing. What is the general consensus here with encoding? Encode at a higher bitrate, author and then shrink to fit or use a lower bitrate and simply author? Just curious what everyone else is doing.

[greengate69]

Personally I prefer to encde directly using the appropriate bitrate to author directly.

Using 2-pass vbr encoding gives (in my opinion) the best results. Certainly for my camcorder DV projects this is the case.

Having said that, if an (already) encoded project is only just over the required target size I would shrink rather than re-encode.

I can't see how the other approach (encode at high bit-rate then transcode) would give better results, but all about opinion I suppose.

[mpiper]

Blazey,

quick question: What encoder are you using?

Here is why I am asking: CCE and a few others allow different quantization matrices. By using a specially designed low bitrate matrix, you get a better image at low bitrates than using the standard matrix. Quite a few people have been messing with this for DivX, and several of those (maybe all??) matrices work for Mpeg-2 encoding as well. I use a customer matrix that is designed for DivX to fit a 2 hour movie on 1 CD. Using this matrix, but at low Mpeg-2 bitrates (2500-3500), I get better details than the standard matrix at the same bitrates.

several encoders seem to create BAD low bitrate encodes, and do not allow custom matrices, so what you use decides what is best.

My priority for best images would be:
1. Multipass VBR encoding with custom/ideal matrix to exact size (CCE, etc.)

2. Multipass VBR encoding to exact size (CCE)

3. Author then shrink

4. Multipass vbr encoding to exact size (Mainconcept, Honestech, etc.)

[lenti_75]

here is the ideea...

1. encode at bitrate as high as you can.....you can't go back once encoded, and later you might need the extra bits.

2. IF you REALLY have to shhrink it (I preffer to use more discs and keep the high bitrate), then encode as necessary

3. use pass 2 when encode, the quality is better.

[bugster]

Encode at the right bitrate to start with. It will give better quality in most cases. The reason? A Transcoder like DVDShrink simply throws away data from the original encode in order to achieve the required size reduction. Admittedly it does it in quite a clever manner but it can't determine the best bitrate allocation in the way that an encoder, when using VBR can.

Get it right 1st time

[Blazey]

Encode at the right bitrate to start with. It will give better quality in most cases. The reason? A Transcoder like DVDShrink simply throws away data from the original encode in order to achieve the required size reduction. Admittedly it does it in quite a clever manner but it can't determine the best bitrate allocation in the way that an encoder, when using VBR can.

Get it right 1st time

Thanks, this is pretty much what I was looking for!