INTERACT FORUM
More => Old Versions => JRiver Media Center 27 for Windows => Topic started by: d_pert on December 17, 2020, 03:35:10 pm
-
I'm wondering if there's something wrong with how waveform metadata is being measured and/or displayed for some M4A Apple Lossless files.
Across three different PCs, many -- not all -- waveforms displayed for M4A tracks bear no resemblance to the dynamic reality of the track. An ordinary pop song, for example, might show as almost flat-line for 1/3 of the way, then slope up perfectly, to a series of plateaus, or other strange, uniform 'shapes' with smooth slopes/edges. I've noticed this since the feature was originally introduced. Happens with 16 or 24 bit, and any sampling rate.
MP3 tracks appear as one would expect: a LOT of intricate, dynamic activity (little peaks/troughs), clearly representing the dynamic 'happenings' in the track, exactly like a DAW / wave editor would.
The attached images show an M4A followed by an MP3 of the same piece of music (different performances, but same approx. length and dynamic qualities).
There are also some M4As which don't have an obviously 'wrong' contour, per se, but appear extremely 'simple' as if the entire dynamic variety of the track has been represented by 5 to 10 peak sample locations.
I've tried re-analyzing tracks to see if they 'update' to a more normal/detailed 'look'; no dice.
Any ideas? Why is the waveform metadata / display so simple/weird for apparently only M4As?
Thank you.
P.S. -- I realize that tracks in general have wildly varied dynamic signatures; this is not about that. ;) This is about why some tracks are represented by a few slopes, plateaus, and blobs (no waves at all), where others appear with a normal detailed 'wave' display.
-
Could you email a file that seems wrong to matt at jriver? Thanks.
-
Done.
-
Thanks. I replied by email, so let's discuss there. If we figure it out, we can post here.
-
Initial investigation looks like M4A files are truncating the waveform data. We save good data, then update from the tags and shortened data comes back. Thanks for the help, and I'll post once we figure it out.
-
Wonderful!
Not sure it it's related, but ... when I re-compress to a lower-spec M4A, with DSP (e.g., peak normalization), using MC's Handheld Devices 'sync' facility, the source file's Analyze data (incl. Waveform string, truncated or not) is transferred unchanged into the new file. The output appears to not have been re-analyzed during a sync which includes DSP. I always thought this was just an oversight, but now I'm wondering if it might be related?
-
That was the problem and Yaobing is making a fix for a beta build later today. Thanks for the help!
-
Always a pleasure.