[FFmpeg-devel] [PATCH 03/13] lavc/videotoolboxenc: Remove redundant code
nfxjfg at googlemail.com
Sun Apr 10 15:36:02 CEST 2016
On Sun, 10 Apr 2016 08:57:11 -0400
Richard Kern <kernrj at gmail.com> wrote:
> > On Apr 10, 2016, at 3:47 AM, Hendrik Leppkes <h.leppkes at gmail.com> wrote:
> > On Sun, Apr 10, 2016 at 6:17 AM, Carl Eugen Hoyos <cehoyos at ag.or.at> wrote:
> >> Rick Kern <kernrj <at> gmail.com> writes:
> >>> Don't require hardware encoding and explicitly fallback to
> >>> software if it fails. Enabling it without requiring it
> >>> will use hardware encoding if available.
> >> Hardware encoding should be required unless the user
> >> specifies something else.
> > I agree, the goal of VT should be hardware encoding primarily.
> This commit message could have been worded better. It already falls back to software - this is just cleaner. The user has to explicitly request this encoder, so it shouldn’t fail if it’s capable of encoding.
> Another goal is to encode video without embedding the encoder in the user's application, so it doesn’t matter if it’s hardware accelerated. I’m not familiar with patent laws, so I can’t really speak to why this is important.
> Beyond that, a few other VT codecs have been requested, but none of these are hardware accelerated.
> In any case, I'll add a require-hw option, but I think the default should fall back to software.
Yes, such an option would be good.
More information about the ffmpeg-devel