I have been noticing this for a while and thought I would go ahead and see if any other users are noticing this as well?
I bring a lot of my recordings along with me when I travel. I use Tablo Ripper to save them and just transfer them to my iPad Air2 through iTunes.
Listening on the plane I use a pair of Westone UM2 in-ear monitors. Most of the time things sound fine but when I am watching The Flash I hear a lot of swimming and artifacts in the audio at times.
I also work as a music arranger/producer and have some experience with encoding audio and that is typical of something going on with the audio encoding profile.
I have also noticed lots of conversations in going from a 5.1 source to a stereo mix and somehow Tablo adjusts the audio down to just the stereo.
I am wondering is this something that Tablo is doing when recording or is Tablo Ripper possibly doing something with the audio when it is pulling something in?
I would love to hear more of what I am possibly missing ;^)
Thanks to the Tablo team and the great community of users here that ultimately help us get the most rewarding experience from using Tablo.
Tablo Ripper uses ffmpeg and from my understanding it is re-encoding the video and audio from the Tablo.
It is not pure export leaving it in the native stream that Tablo reencoded. So that could be part of the issue.
I’m pretty sure CycleJ has Tablo Ripper using ffmpeg to concatenate the TS segments into an MP4 container. That’s what I’d be doing anyway. There would be no problem doing this using ffmepg.
I know from discussions with Jestep that his Tablo Exporter that uses ffmpeg even in the “copy” mode is actually a re-encoding of the video and audio. It is not a simple re-mux of the source streams from .ts to .mp4.
Maybe the two authors can chime in and correct me if I’m wrong?
Well, remember even Tablo is encoding from mpeg-2 which is the standard OTA stream and converting to mp4 in order to support the HLS stream. But I would assume when you convert from one format to another, regardless if they are "technically the same (HLS to mp4), would cause some loss to some extent.
Yeah the Tablo does 5.1 AC3 audio to 2.0 AAC. But ffmpeg is doing AAC to AAC when exporting no?
Well yea, but unless you want to go look at every line of code for ffmpeg, I am just making an assumption it is not bit for bit. I could be wrong, but just my guess.
No, the encode is h.264 in both the .ts and .mp4, so there isn’t any need to re-encode. ffmpeg will just concatenate the segments into an MP4 contain if asked to.
And the other thing is @simeon listening the same way on the non exported stream Maybe it is in the original conversion
If ffmpeg is concatenating the .ts, not encoding, it would have to be in an earlier conversion I guess.
I might try that by trying to find a spot where I hear the artifacts and listen here at home on the Tablo app.
I guess this just brings up the Tablo converting the original broadcast audio format into something else. I agree with you @Jestep that it probably is in the original conversion.
Thanks for the feedback,
Yea, I just don’t think Tablo can keep the original and then convert/not convert on viewing, which may be the “right” way to do things. But that is not what they were designed for originally