Encoding Quality affected by interlacing?

[puntloos]

Just wondering.. to an encoder a frame inside an interlaced source file would probably look like something with a lot of small fine lines (hard to encode).

Is an encoder usually smart about this or is it better to provide an encoder (like TMPGEnc or CCE) with a non-interlaced (progressive) source?

And would it be more optimal to output a progressive or interlaced MPEG-2 file? My DVD player can handle both progressive and interlaced.

[Arky]

First of all, you obviously have a decent player (a KISS, perhaps?) if it has progressive playback mode. Mine does (I just bought a Limit 9900SE), but beware that a standard TV will not cope with progressive playback, in the hardware-output sense.

Anyway, that is by-the-by. To address your main question, you are correct in your suspicion that a progressive source will be better-encoded by your encoder of choice. Bitrate is used much more effectively when encoding a progressive source than when it is used to encode an interlaced source. Of course, many people will claim that you get subjectively-smoother motion with interlaced material, but I would argue that this might be traded-off successfully if you get a higher-quality end result (with less compression artifacts, due to more efficient bitrate distribution) with progressive sources.

I encode both formats, and don't really discriminate between the two to any great extent, since I can still get great results with interlaced footage (DV-cam firewire captures) with the allowable bitrates which DVD-R can handle, owing to it's large capacity (I rarely exceed 90mins for my own projects). If you're going to back-up commercial Hollywood films onto DVD-R, however, then the duration of these are pushing the limits of the DVD-R capacity, so there is a legitimate (if inconclusive) argument for favouring progressive encodes in such instances, where available bitrate is limited.


Arky ;o)

[puntloos]

Aye, you're right.. but the thing is that actually I have exactly what you described.. an interlaced DV data source.

My problem is that I only have SVCD (703Mb mpeg2) to play with, no DVD burners.. so every last bit counts..

And I was toying with the idea wether or not de-interlacing the DV source before mpeg-encoding is the best way to preserve as much quality as humanly possible.

(And yea I have a Pioneer 444 DVD player.. plays mp3 too so Im encoding my footage to MPEG-2 + MP3 audio on 1650 kbit/sec total it makes a diff to go from 256kbit audio to 128kbit for acceptable quality

[Arky]

Well, to be honest, I think the simplest solution (not the only one, but the simplest!) would be to use TMPGEnc or ProCoder, because they naturally deal with field-order correctly. There are many people who succeed with CCE for SVCDs, but because I have never been into SVCD authoring, I don't know what solution they use for overcoming the field-order issue in CCE. I suspect they swap field order using AVIsynth, but this in itself would not be an ideal solution either. The man you should talk to about this is German, and you may have heard of him before. His forum nickname is "mb1", and he has a website dedicated to the topic, but sod's law means that I can't find the relevant link!!! Sorry. He can often be found on Doom9. Because I am new here on VCDhelp, I don't know if mb1 hangs around the VCDhelp forums.


Arky ;o)

[puntloos]

Thanks for your help. I appreciate it!

Anyway I compared TMPGEnc's output (at every quality setting maxed out) with CCE and CCE's output is slightly superior. I haven't noticed any field order problems but I think my player is kinda flexible...

But if you look at output windows (for example with TMPGEnc) you see the interlace jagginess so I'm kinda assuming that that is what it sends to the codec. And if that's the case then it is wasting a ton of bandwidth trying to encode those teensy jaggies.

[Arky]

If you are subjectively dissatisfied with TMPEGEnc, have you tried Canopus ProCoder instead?


Arky ;o)