[mythtv-users] The Final Nail in MythTranscode?

[Karl Dietz]

On 05.03.2013 12:47, Anthony Giggins wrote:
> On 05/03/2013, at 9:20 PM, Paul Gardiner<lists at glidos.net>  wrote:
>> On 25/02/2013 08:11, Ian Evans wrote:
>>> just posted a thread about the .25 copy over trick no longer working
>>> either due to a missing libmythtv-0.25.so.0
>>>
>>> Did you get that issue too? Really screwed with mythtranscode and space now.
>>
>> Not sure this is very useful since from what I could see reading back
>> you want lossless transcode, but I've been using mythtranscode in
>> --fifodir mode running it into ffmpeg. That hasn't been broken in
>> 0.26. If anything it works better. If you are coming from mpeg2
>> recordings you get the advantage of the smaller files that h264 allows.
>> You can preserve interlacing, whatever you want. My script is linked to
>> from http://www.mythtv.org/wiki/Mythtranscode. You can just set it up
>> as a user job.
>>
>> I don't know. Maybe I shouldn't bother and just buy more disc space
>> instead, but it is nice to be able to stick a series on a DVD to give
>> to a friend who missed it. And I'm often taking recordings over to my
>> girlfriend's and it's much faster to copy 1G than 4G to a USB
>
>
> My theory is you can fit even more content onto these ever cheaper drives if you transcode, especially due to channel over runs and having to record upto 22min (20 mins hard pad&  2 mins soft pad) just to make certain you get entire shows in my country that adds up to a lot of extra space which would otherwise be wasted.
>
> Ie. I'd be down almost half of the 1486 recording I currently have and I'd require more Sata ports also.

It appears as if you (the group, not the person) are talking about
something different, but using (almost) the same wording. That is quite
confusing for me (and likely for others, too)

I see the following opinions
  * transcoding is bad - as in you should not encode the same video twice
    as you get generational artifacts, even worse if you mix codecs like
    first round MPEG-2 and second round H.264
  * transcoding is bad - as in it burns more money in electricity for
    the CPU then it saves money for disks and running the disks.
  * transcoding is good - as in you can cut off excess pre-/post-roll and
    advertising without reencoding (lossless cut)
  * transcoding is good - as in it makes smaller files that are easier to
    carry around (could be lossless cut or "the real thing")
  * transcoding is good - as in a good implementation of a modern codec
    makes a smaller file than a bad implementation of an old codec
  * transcoding is good - as in you drop all excess audio and data
    streams and repackage in a more efficient container
  * I'll leave scraping bits of audio/video streams without reencoding
    out of the list as its not really supported well. (requantisation,
    stripping the E- from E-AC3, etc)
  * some also factor in the value of personal time to get "it" running
  * some think transcoding is needed to support third party devices with
    restricted codec support

May I suggest to use different words for different things, maybe even
explain which thing you are talking about?

Some wordings that I've seen in the past are:
  * lossless cut
  * reencode to codec X (e.g. H.264)
  * export to container X (e.g. Matroska)
  * export to device (e.g. iPad)

Regards,
Karl

btw, over here I've got the following situation
a) stations with ads have around 3:1 to 2:1 content to ad ratio
b) I need up to ~45 minutes end-late for some shows that run at night
c) I record some short kids show of 5 minutes with 10 minutes
    start-early/end-late
I do think lossless cut with switching to a more efficient container
while dropping unwanted streams is a good thing.