•  Previous
  • 1
  • 2(current)
  • 3
  • 4
  • 5
  • 8
  • Next 
 
Thread Rating:
  • 1 Vote(s) - 5 Average
ACCEPTED: On-the-Fly Transcoding
#16
I'm sure, but let's do it step-by-step.
A fixed transcoding of video to, e.g., 720p h264 + 5.1/2.0 aac, would be a tremendous achievement already, and would already please a lot of users.

From there, all the rest would be quite easy to develop.

I'm personally not convinced a FLAC to mp3, or even audio-only, transcoding should be high-priority. FLAC is niche, imho.
Reply
#17
(2015-03-20, 12:34)Tolriq Wrote: Actually from lot's of user feedback the transcoding needs are 33%/33%/33% :

- Video to local devices or limited devices so h264 / mp3 (or aac) (aac is not supported correctly by some old devices)
- Video to WAN : So main need is to limit the bw so lower the video profile and mp3 for audio (mostly watch tv shows on the go)
- Audio to WAN so flac / high quality mp3 to lower quality mp3 for all low bandwith users.

This should cover 90% of actual needs and be better for the future. (And for WAN need UPNP is not good)

But as stated, if the goal is to make everything as simple and limited as possible on first run, h264 and aac seem like as good a solution as any. And the question of UPnP and WAN are things to worry about only if transcoding gets done fairly early on.
Reply
#18
I was just pushing the feedback I got from about 1.2 million users now (800K active) so no Flac and audio only transcoding is far from a niche Smile

But yes step by step is the way to go as always just keep users needs in mind for future target, as dev needs are often far away from users needs.

And the WAN need being big it's just a big +1 on doing the transcoding before serving and not only as an UPNP extension.

So as I said i just push information.
Reply
#19
Re aac, it's mostly because it's widely supported, supports multi-channel, and is license-free, btw...
Vorbis would be an alternative, but it's far less well supported in consumer products, for reasons I still can't grasp Wink

[EDIT] Well, not so license-free, apparently http://en.wikipedia.org/wiki/Advanced_Au...nd_patents
But hey Wink
Reply
#20
(2015-03-20, 09:07)Koying Wrote: Yep. In a first instance, I'd suggest to agree on a single, hardcoded, profile (e.g h264 + aac) to maximize the chances to reach the transcoding goal for most of potential clients.
In that case then might to just copy the Built-In Presets from example HandBrake for specific player clients. Also FFmpeg have have their own iOS compatibility matrix for x264 encoding presets. These preset profiles are pre-configured with resolution, bitrates, and video/audio profiles.

https://trac.handbrake.fr/wiki/BuiltInPresets
https://trac.ffmpeg.org/wiki/Encode/H.264
http://rodrigopolo.com/ffmpeg/cheats.php

Have a drop-down list in Kodi's GUI where the user can manually select one out of ~10 preset profiles, as one preset could use 1080p x264 with AAC and another 720p x264 with MP3. At least then you will meet more people requirements and still being relatively user-friendly without having to automate the process.

So no need to even display the specific encoding specifications in Kodi's GUI, instead only have a simple drop-down list with a few per-configured preset profiles for the most probable commonly occurring clients.

Example of devices as preset profiles that user can select in Kodi's GUI settings
  • Universal Low-End Profile
  • Universal Midrange-End Profile
  • Universal High-End Profile
  • iPod
  • iPhone & iPod Touch
  • iPad
  • AppleTV 2
  • AppleTV 3
  • Android Phone (Midrange-End)
  • Android Tablet (High-End)
Reply
#21
(2015-03-20, 12:43)Koying Wrote: I'm sure, but let's do it step-by-step.
A fixed transcoding of video to, e.g., 720p h264 + 5.1/2.0 aac, would be a tremendous achievement already, and would already please a lot of users.

From there, all the rest would be quite easy to develop.

I'm personally not convinced a FLAC to mp3, or even audio-only, transcoding should be high-priority. FLAC is niche, imho.

+1
"Your most vocal users are not reflective of your userbase" J.M.
"Of course, they [XP Users and people with outdated hardware] need to tell the world about the kind of hero they are, and block innovation for everyone else because their decades old OS / hardware needs to work =p" nevcairiel (ffmpeg)
No Debug Log no issue.
Reply
#22
(2015-03-20, 12:51)Tolriq Wrote: I was just pushing the feedback I got from about 1.2 million users now (800K active) so no Flac and audio only transcoding is far from a niche Smile

You have that many cases of user feedback or are those just anon-stat numbers?

FWIW, Google music already accepts personal FLAC uploads and plays them back to your device at 320k
Reply
#23
Ok, let's assume it's not a niche, but I surely don't see it as a priority for Kodi.
I really, really think there is a basic incompatibility between users willing the quality of FLAC and transcoding Wink
Reply
#24
I obviously do not have received 1.2 million feedback Wink but enough about the use cases to draw the percentage for the numbers to be representative.

@Koying
This is where you are wrong Smile

I do use FLAC a lot at home, I have put quite some money in good speakers and everything needed to enjoy full quality music when in my living room, but as now a very very large user base of kodi I also have tablets and phone and I like to listen to my music too when on the move.

And syncing flac on device takes way too much places, and streaming flac can take too much bandwidth for most users (I'm lucky I have 200Mb upload link from home so not my case).

The need to off home media on devices is big, this is the future this is what users want, watch their tv shows in the bus, listen music everywhere, stream their movies to friends tv when at friends house, with bluetooth you can throw your music everywhere instantly.

You should start some public pools to really understand users need and what the future should be for Kodi and this future could be quite soon is enough people understand this Wink

One box with one hard drive under the main tv is the past.

Anyway I know your point of view from the discussion on Kodi name thread, f...k users we do what we want and what we need each Smile (This part is not to start a debate it's just a reminder of a known situation).
Reply
#25
I would not discuss too much about the "profiles". When this feature is implemented correctly, e.g. the extension point is "fully ffmpeg aware" ...no format, that ffmpeg suport is a problem at all. Be it flac (totally sensless if mp3 is input or be it ogg, mp3 or whatever).
"Your most vocal users are not reflective of your userbase" J.M.
"Of course, they [XP Users and people with outdated hardware] need to tell the world about the kind of hero they are, and block innovation for everyone else because their decades old OS / hardware needs to work =p" nevcairiel (ffmpeg)
No Debug Log no issue.
Reply
#26
(2015-03-26, 21:27)Tolriq Wrote: I obviously do not have received 1.2 million feedback Wink but enough about the use cases to draw the percentage for the numbers to be representative.
Man. Assuming the 1.2 millions you speak about are all your users, I don't believe one bit they are all FLAC users (out of about ~6-7 millions active Kodi users, most of them not giving a f....ck about our music part)

Anyway, don't be deaf. I never said it should never be implemented, I just say it should not be the priority of this GSOC project.
I would be very, very disappointed if this project would be endangered by f...ing FLAC.
Reply
#27
(2015-03-26, 21:39)Koying Wrote: I would be very, very disappointed if this project would be endangered by f...ing FLAC.

lol. I think we're all on the same page here. The important thing is the transcoding. Everyone agrees FLAC would be nice to have maybe, but that can happen at any time after the initial work is done. For now, keeping it simple is absolutely the most important thing.
Reply
#28
Might be good to know that FFmpeg version 2.6 which was just released now supports NVENC - Nvidia Video Encoder API interface for H.264 hardware encoding on GeForce based GPUs

http://git.videolan.org/?p=ffmpeg.git;a=...elease/2.6

And I believe that at least Libav already supports H.264 / x264 hardware encoder on Intel GPUs via VA-API, so adding hardware encoders via FFmpeg might to be too much work?

http://www.intel.com/content/dam/www/pub...-paper.pdf
Reply
#29
vaapi wanted to merge a new middle ware into ffmpeg and that was not accepted. So VAAPI encoding only works with some shitty middleware. Libav is not an option for us.

ffmpeg 2.6.2 was merged into our master yesterday.
"Your most vocal users are not reflective of your userbase" J.M.
"Of course, they [XP Users and people with outdated hardware] need to tell the world about the kind of hero they are, and block innovation for everyone else because their decades old OS / hardware needs to work =p" nevcairiel (ffmpeg)
No Debug Log no issue.
Reply
#30
If this project happens I have only one thing that I think should be taken in mind.

Please allow the transcoder parameters to be modified very granular by the user.

Mainly allow the user to use any custom bitrates and any custom resolutions. This should be trivially easy to provide options for. People who are streaming remotely with limited upload bandwidth or limited cell phone internet plans would greatly appreciate being able to fine tune quality and bandwidth.

Also allow the user to override/add other parameters like x264 preset, and advanced parameters like cabac, b-frames, etc. This would make all the people with lots of extra CPU power to spare happy as they can squeeze more quality into less bitrate.

These features/ability are missing from all the competitors like Plex for instance.
Reply
  •  Previous
  • 1
  • 2(current)
  • 3
  • 4
  • 5
  • 8
  • Next 



ACCEPTED: On-the-Fly Transcoding51