-Brading Software Message Board
+--Forum: Mp3/Tag Studio Support and Bug Reports
+---Topic: What a BASTARD FILE! Wrong lengh, kbps? started by Raysy
Posted by: Raysy on Sep. 30 2002,15:22
Hi Boys! (and Girls, if there are any in here :-))
My Problem is, that MP3TS shows the wrong bitrate with one MP3. It shows 416 kbps, while the real bitrate is 112 kbps. (according to Winamp and the filesize) and the wrong lengh. It's 8:53, not 2:23.
It's a rapsong dealing with the German Election, that took place last month. I tagged it with WinAmp, because originally it has no tags, but I don't think this caused the problem.
Data:
Artist: Toni-L, Spax, Flowin’ Immo, Maze & Bene (MB1000), Later (Burns Bros.), Kabelage & DJ Brisk Fingaz
Title: Wählt nicht Stoiber
Size: ~ 7,1MB (I tagged it)
MP3TS: 2:23 min, 416 kbps, Joint-Stereo, 32000Hz, all flags "Yes", Emphasis: CITT j.17
EncSpot 2 (www.guerillasoft.com): 8:53 min, 112 kbps, Joint-Stereo, 41000Hz, Encoder guess: FhG (fastenc or mp3enc), Scalefac: 17.2%, Emphasis: none, Sync Errors: none, Bad Last Frame: No, Xing Header: No, Original: No, Copyright: Yes
Summary:
When I presume that WinAmp 2.80 says the truth, then both tools are partly wrong:
MP3TS shows up with the wrong lengh and bitrate, EncSpot 2 with the wrong sampling frequency, Emphasis and original flag.
But WinAmp 3.0 tops it: 2:23 min, 416kbps at the file info menu, BUT at the main windows it shows 112 kbps.
And Windows Media Player 6.4 says 8:53 min, 44100Hz and 112 kbps.
So: 112 or 416 kbps, 32000 or 44100Hz, Emphasis: Yes or No, Original flag: Yes or No, 8:53 or 2:23 min?
My personal guess: WinAmp 2.80 is the right, except the sampling frequency.
But check it for yourself: The file is free promotion and everybody can download it:
< http://www.rap.de/bundestagswahl/antistoiberpopup.html >
and download"waehltnichtstoiber_lo.mp3" from one of the three servers. Please post your own experiences, if they differ.
Thanks for reading!
Raysy from Germany
Posted by: Magnus Brading on Oct. 01 2002,00:10
That mp3 contains 69 bytes of trash data in the beginning if the file. Parts of this trash data does also unfortunately have such a pattern that it is detected as a valid mp3 frame header, and hence the program interprets the wrong sound data stats from it.
The correct stats for this mp3 is:
MPEG version: 1
Layer: 3
Sample freq: 44100 Hz
Bitrate: 112 kbps
Private: No
CRC: No
Copyrighted: Yes
Original: No
Emphasis: None
You can repair the file manually like this:
Step 1. Overwrite the first 10 bytes of it with zeroes, with e.g. a hex editor (this will neutralize the problematic data in it, and should suffice for it to show up with the correct data in most players).
Step 2. With Mp3/Tag Studio, perform a repair of the type "Remove junk data before first mp3 frame" on it (this will completely remove all the junk data, for that extra aesthetic touch, although step 1 alone will solve most problems caused by it
).
Posted by: Raysy on Oct. 01 2002,15:25
Thanks for the fast reply! It worked, although it was not very comfortable, because I wasn't used to a hex-editor.
I would like to do it all the way with MP3TS repair-function. Perhaps this could be even more improved??
Posted by: Magnus Brading on Oct. 01 2002,16:25
If it weren't for the malignant composition of that trash data, step 2 would have been enough to repair the file, and this is the case for most files with trash data in them.
Errors like that aren't very common at all, so it should not be a problem anyway (i.e. you won't have to perform the procedure on many files). No matter how "good" I make the repair functions, there will practically always be a way to construct a file that will confuse it. As I said, the problem with this trash data was that it contained data that could be taken as a genuine mp3 frame header, and that's why we needed the neutralization step at all.
Anyway, being able to use a hex editor is always a good knowledge to possess, so it was probably not all bad anyway.
end