Strange file format?

I downloaded a file from Youtube that plays fine on my computer, sounds ok with computer buil-in speakers and with an ordinary headphone, but sounds distorted and lousy through my Bose system. OK I know this is bad recording, but some bizzare stuff with this file are as followes:

  1. This is an album with may songs all recorded in single track. So to separate the tracks when I run the Audacity’s silence finder looking for silence when db < -51 and minimum duration = 1 sec. it cannot find the silences where it should. Instead all found silences are near the very end of the last song of the album (and all labels are lumped together overlapping with each other. I tried undoing and re-running the silence finder many times and got the same strange result. Eventualy I ended up marking the silences manually and that worked fine; I was able to put each song in different tracks. But the question is why did the silence finder did not find the silences correctly? I know how to run this, and have done this many times for other sound files that worked perfectly - but not for this file.

  2. Now then when I run the Audacity’s spectrum analyzer (Analyze → Plot spectrum …) choosing random areas of the track I see “no data” ploted - just an empty graph whose vertical scale ranges from -90 dB to -80 dB. Then when I move the cursor over the same graph the data values shown in the boxes below the graph shows values like -114 dB or -120 dB or -192 dB, etc. that are obviously outside the normal range of the graph. Strange??? Now then when I select a range at the very end of the track (remember: that is at the very end of the last song) I see some spectrum plotted in the graph that are in normal graph area. Also as I described in #1 above this is the same area where the silence finder “thinks” it found all the silences.

  3. The whole track however plays normaly in Audacity, the level meter indicates normal dB always peaking at or slightly past the 0 dB where the thermometer color transitions from green to yellow to red and back to green.

I am puzzled here. Why is the file this way? Is it safe to play this file on any device? Can it mess up the sound decoder cicuit in my player in any way? Ofcourse, my question here is academic - this is not a serious production that has to be corrected; I can easily discard the file.

Did you try right-click or control-click > Get Info? That might go a very long way to finding out what you really have. It might simply be a video file that Bose doesn’t understand.

That’s not a wacky idea. You can create a Data CD and an Audio CD with the same music on it and they will both play famously on any computer and many personal music players, but the Data CD drops dead if I try to play it in my old pickup truck which doesn’t understand data streams. It’s good to have a clear idea of the format and application.

Koz

If the “thermometer” goes red like this …
Red bits mean clipping = distortion.png
it means the waveform has gone off the scale and will be clipped , which is a form of distortion.
If you normalize the waveform to “0dB” the sound will be made as loud as possible without clipping.
[ If you are making an MP3, normalize to -1dB as the conversion-process can increase the volume ].

To clarify my question I am less concerned about sound cliping and poor recording than the missing frequency spectrum for the track, and the silence finder not being able to find the silence correctly as a result of this; and yet even with missing frequency spectrum the players (both Audacity and my Bose system) is able to correctly play the track. How is this possible? I don’t know how to embed image here, so I am attaching them here. There are four images:
one - main Audacity project window with the bad track where I show a 1 minute segment selected.
two - a frequency spectrum plot from the 1 minute segment selected above; nothing is plotted - why?
three - main Audacity project window of a good track where I show a 1 minute segment selected like in “one” above.
four - a frequency spectrum plot from the 1 minute segment selected in “three”; note the presence of a perfectly good spectrum in this case.

Both the tracks in my examples are imported from Youtube video MP4 file.
one.png
two.png
three.png
four.png

If the two tracks are identical , the only difference being one has been inverted, ( i.e. in anti-phase ), then they cancel each other out : as far as frequency-analysis is concerned it’s silent , no frequency content.

If you only select one of the two anti-phase tracks, ( by temporarily splitting the stereo pair ), then a frequency analysis will appear.