FORUMS: list search recent posts

Apple Compressor. H264 50P the same filesize as 25P from 50i Source.

COW Forums : Compression Techniques

<< PREVIOUS   •   VIEW ALL   •   PRINT   •   NEXT >>
Scott Goddard
Apple Compressor. H264 50P the same filesize as 25P from 50i Source.
on Jun 1, 2013 at 1:14:21 pm

Converting 50i (interlaced) footage (XDCAM HD 422/ prores or any other interlaced source) in apple compressor to h264 50p renders the same filesize as a 25p h264 encode.

Same in Episode.

If you do a conversion to prores from 50i to 50p the filesize will be double that of a 25p version.

Any ideas why this is? I would have assumed the h264 50p to be twice the size of a 25p file.

50i looks great converting to h264 50p. As an intermediary transfer format it works very well for online transfer / delivery to maintain the correct temporal motion and still enabling 50i workflows to be maintained both in the field and to convert back for 50i broadcast.

One day interlacing will be over. One day..

Scott Goddard

Neo Verite Limited
http://www.neoverite.com


Return to posts index

Craig Seeman
Re: Apple Compressor. H264 50P the same filesize as 25P from 50i Source.
on Jun 3, 2013 at 4:35:02 pm

File size is a function of data rate times duration.
Data rate per second doesn't care how many frames there are in that second.
50i to 50p is converting fields to frames so the encoder is creating the extra information but not necessarily increasing the file size. It's using the bits (data rate) you targeted to do that.



Return to posts index

Scott Goddard
Re: Apple Compressor. H264 50P the same filesize as 25P from 50i Source.
on Jun 3, 2013 at 5:24:30 pm

I see where you are coming from with the target data rate and on that front you are absolutely correct. In regards to doing the same encode to a predefined codec, such as prores 50i to 50p this will double the file size. Can you explain the difference here in the same terms?

I guess it's because we cannot choose a data rate as that is predefined by the codec standards. So loosely speaking 50p will always be double 50i.

Do you see any downsides of converting 50i (source) to 50p h264 for FTP transfer? The useful aspect is that we retain temporal motion and have a low fielsize to transfer. We can then convert this back to 50i for broadcast with very good results. I see news agencies transferring video from 50i to 25p all day and I really question it as the none thing broadcasters like is to maintain the fluid motion. Seeing as the target bitrate is maintained and by eye the quality is excellent I see little reason to not do this.

It's all workflows, but I am surprised by the way many broadcasters are dealing with this.

Much appreciated.

Scott Goddard

Neo Verite Limited
http://www.neoverite.com


Return to posts index


Craig Seeman
Re: Apple Compressor. H264 50P the same filesize as 25P from 50i Source.
on Jun 3, 2013 at 6:07:02 pm

[Scott Goddard] "In regards to doing the same encode to a predefined codec, such as prores 50i to 50p this will double the file size. Can you explain the difference here in the same terms?"

ProRes codec is determined by the files parameters such as frame size and frame rate. It's part of the spec.

[Scott Goddard] " guess it's because we cannot choose a data rate as that is predefined by the codec standards."

That's it exactly.

[Scott Goddard] " I see news agencies transferring video from 50i to 25p all day and I really question it as the none thing broadcasters like is to maintain the fluid motion."

As you note, that would lose temporal motion. That's the advantage of 50i to 50p. There may be some impact on quality as you're creating information that didn't exist but that depends on the deinterlacing algorithm. My preference would be to keep the temporal motion as well.



Return to posts index

Scott Goddard
Re: Apple Compressor. H264 50P the same filesize as 25P from 50i Source.
on Jun 3, 2013 at 6:11:20 pm

I felt the quality hit would be a factor too when I theorized the maths, but like with anything compression related, seeing is believing. I round tripped some test files and on a broadcast monitor I was hard pressed to notice a significant loss. Certainly not anywhere near the loss of loosing the temporal motion..

Thanks for the discussion, it's good to talk this one over with others.

Scott Goddard

Neo Verite Limited
http://www.neoverite.com


Return to posts index

<< PREVIOUS   •   VIEW ALL   •   PRINT   •   NEXT >>
© 2017 CreativeCOW.net All Rights Reserved
[TOP]