Compress a video for Web: starting from Interlaced or Progressive?

[Daniel Low]

Progressive source is always better than interlaced source when delivering to a progress output as there’s no need to deinterlace.

__________________________________________________________________
Please post back saying what solved your problem. It could help others, and saying ‘thanks’ is free!

[Alessio Gemma]

Thanks Daniel, I suspected this..
So, for example… I edit a DV PAL video in FCP and then I export in a ProRes 422 Progressive’s self contained clip: starting from this, I use Compressor to get the Web Clips I need. Do you think is the right workflow?

Do you think it’d be better to resize the clip directly from the first export or let the Compressor do the resize? I know that some codecs suffer and loose quality when resizing clips.

Regards,
Alessio

[Daniel Low]

Why do you export to ProRes? – Exporting out to ProRes nearly doubles the bitrate compared to DV (42Mb/s vs. 25Mb/s of DV) Check out this Whitepaper on ProRes
https://images.apple.com/finalcutstudio/resources/white_papers/L342568A_ProRes_WP.pdf

It’s best to leave the resizing (and deinterlacing) to Compressor.

When exporting a finished program/sequence to a transcoder (compressor in your case) on the same machine I generally export as a non-self contained clip at the current sequence settings (in your case DV PAL). This avoids any kind of format conversion, it’s fast and the transcoder gets exactly what it needs, no more, no less.

__________________________________________________________________
Please post back saying what solved your problem. It could help others, and saying ‘thanks’ is free!

[Alessio Gemma]

Ok, I’ll try your settings.
But.. if I directly export from FCP to Compressor, without creating any QT clip.. do you think it’d be better?

I told tou about the progressive source, because I read an article on Adobe Web Site about FLV compression, where the authors wrote to start from a progressive source to encode in FLV video format.
Here the link:

https://www.adobe.com/designcenter/dialogbox/encode_video_02.html

Regards,
Alessio

[Daniel Low]

I always prefer to have the reference clip as going straight to compressor has caused problems in the past.

The Adobe document quotes this “Always render a deinterlaced (progressive) version of your video before encoding to FLV”

I’m guessing the reason they said that is that not all encoders have a deinterlacing filter?

The bottom line is that if you are delivering to progressive output then you need to deinterlace if you are using interlaced source, if you are using progressive source there is no need to deinterlace.

__________________________________________________________________
Please post back saying what solved your problem. It could help others, and saying ‘thanks’ is free!

[Ben Waggoner]

If I’m doing any graphics at all in my NLE, I render out as progressive from the NLE. That way all the fine details in my font are at least rendered in progressive, even though the underlying video will be deinterlaced. Most editors these days will automatically deinterlace interlaced source when rendering out as progressive.

Ben Waggoner
Principal Video Strategist, Silverlight
Microsoft Corporation

Compression Blog: on10.net/blogs/benwagg/
Compression Classes at Stanford and PSU: on10.net/blogs/benwagg/21622/