Sadly, there is no ETA on when/if this would ever be fixed in ExoPlayer. and I did for a week, and then I threw in the towel this weekend, and switched back to IJK. I hit the next commercial and skip forward, and it thinks I'm way back in time, so then I skip past the next commercial block, and the video plays correctly. Start watching, get to commercial, I skip forward once to skip the commercialĮxoPlayer starts playing at the right spot, but the time bar "loses" the information about where i am at, and then resets the timebar back to original position., as if I never skipped. but a typical ExoPlayer session is like this.
I wouldn't enable auto-comskip on exoplayer. the Seeking is annoying, but things generally still work. I also get some lip sync problems with IJKPlayer presumably due to the lack of passthrough.ĮxoPlayer works.
The exception is that with the release I did on the weekend, it can use ffmpeg to software decode ac3 audio if you don't have an audio-passthrough setup (which I don't). I've watched about 10 shows with it so far, and I haven't found any issues, yet.ĮxoPlayer is a pure java player with no native parts (makes it easy to build and work with). I the 2.xx release of ExoPlayer is pretty stable, and as of the release I did on on the weekend, I'm going to try to use ExoPlayer as the default player (in my setup), to see other issues I encounter. but I also would run into some issues on some files, and so I tended to stick with IJKPlayer. Initially it didn't support mpeg2-ps, so Narflex added that to ExoPlayer. Narflex later suggested that I look at ExoPlayer, and so I did. Initially I chose IJKPlayer because it's a ffmpeg based player and it could software decode formats if the hardware didn't have hardware decoders (more useful on a phone/tablet, since those generally are lacking on the hardware decoding side). Also it's nice to have options which is another reason why ExoPlayer is there. Is there any advantage in using Exoplayer over ijkplayer?Īudio passthrough appears to be the only big one that comes to mind. However, because the new decoder isn't that well tested yet, for all other channel configurations ArcSoft is currently still the default decoder option (if it's available, otherwise dcadec becomes default).The new version is working great with Exoplayer on my Nvidia Shield and Comcast MPEG2 channels. The latest eac3to build now automatically switches to dcadec for all DTS 7.x tracks. it doesn't support XSA / LBR low bitrate tracks yet + eac3to can run multiple decoding tasks in parallel + it can decode all three 7.1 configurations perfectly + it never post-processes, for any channel/speaker configs The good news is that there's a new open source DTS decoder (dcadec) available, which in its current form has the following advantages and disadvantages compared to ArcSoft: I'm sorry about that, and not very happy myself. Basically this means, to be safe I'd recommend to redo all DTS-HD Master Audio 7.1 tracks. And one speaker config seems to have been decoded by ArcSoft correctly, but then post-processed, with some mixing going on. One of them was decoded perfectly by ArcSoft - but only with some dtsdecoder.dll versions, not with all (with some old 1.1.0.0 versions the surround and back channels were swapped, but otherwise lossless). Basically there are 3 possible speaker configurations for 7.1 MA, one of them is the "strange setup", the other two eac3to has not specifically marked. Recently I learned that also some other DTS-HD Master Audio 7.1 tracks may not have been decoded correctly. We already knew that "7.1 (strange setup)" DTS-HD Master Audio tracks were not decoded correctly by the ArcSoft decoder. * fixed: #264: using "-float32 -normalize" didn't work in all casesThere's good and bad news. * fixed: #263: decoding TrueHD Atmos with active dialnorm information failed * fixed: #086: left/right eye information was inverted in some 3D Blu-Rays Code: * added libDcaDec decoder for DTS decoding, new default for 7.x tracks