Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Tool for detecting transcoded mp3's (Read 9840 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Tool for detecting transcoded mp3's

Hello,

I wonder if there is a tool that can tell if an mp3 file has been encoded. I guess that it would be hard for the tool to tell, but maybe it could make some assumptions??? I mean if I have a file that is encoded with LAME but has only a frequency of 16 kHz then it would be most likely that it was actually encoded with Xing.
--alt-presets are there for a reason! These other switches DO NOT work better than it, trust me on this.
LAME + Joint Stereo doesn't destroy 'Stereo'

Tool for detecting transcoded mp3's

Reply #1
I have encoutered cases before where I downloaded a 320kbps MP3 and was able to determine that it had been transcoded from a lower bitrate by re-encoding it myself using alt-preset standard.  If, using an example I have seen, a 128kbps MP3 had been transcoded to 320kbps the new encoding would use 95% or more 128kbps frames.  I have also seen a case of a 160kbps MP3 transcoded to 320kbps and detected it in this manner.  Any time virtually all frames are at the same rate using alt-preset standard I would suspect transcoding.
Dave

Tool for detecting transcoded mp3's

Reply #2
Lame is storing the source type in the INFO tag. So if it's originating from mp3, you can know it.

Tool for detecting transcoded mp3's

Reply #3
Quote
Lame is storing the source type in the INFO tag. So if it's originating from mp3, you can know it.

Could you please give an example? I’m not sure what I have to look for. Also, not every Lame file comes with a Lame Tag! I encoded some mp3's a while ago (using Lame 3.92) and there has been no LAME Tag included. Then I have other files I encoded some weeks ago (Lame 3.92) and they do come with a LAME Tag...

Btw. what does 'unwise settings' indicate?

@dhdurgee
yes, that could be a method but it is pain in the butt to do that with all your files...also as you may know there are some groups that like to release 'fake files'...as long as they only cut something out or put in a loop it is fine, because it is obvious to everybody that it is a fake file; but what if those guys just encode a song with Blade or Xing and then encode it again (same or higher bitrate) with FhG or LAME (without the LAME TAG)? That could become to a serious problem...

[Edit]
I just transcoded a 192kbps CBR file (using --alt preset standard) and it gave me a even bigger file as the source where it was taken from...I think the aps settings do not work properly. How can something like that happen? I used Lame 3.92...very irritating
--alt-presets are there for a reason! These other switches DO NOT work better than it, trust me on this.
LAME + Joint Stereo doesn't destroy 'Stereo'

Tool for detecting transcoded mp3's

Reply #4
Transcoding to --aps will tell you nothing about whether an MP3 is transcoded from a lower bitrate or not.  Transcoding from CBR to --aps might make a larger file or a smaller file - the output file won't be the same bitrate as the input.

Tool for detecting transcoded mp3's

Reply #5
Quote
Transcoding to --aps will tell you nothing about whether an MP3 is transcoded from a lower bitrate or not.  Transcoding from CBR to --aps might make a larger file or a smaller file - the output file won't be the same bitrate as the input.

Could you explain why not?

To me it sounds reasonable; if you just beat all the difficult pieces out of the music by encoding it 128kbit cbr, then why should aps need high bitrates on such a file?

edit: why would we need tools to see this kind of things? Were all audiophiles here, right? So we'd hear if a file was badly made... 

Tool for detecting transcoded mp3's

Reply #6
Quote
Quote
Transcoding to --aps will tell you nothing about whether an MP3 is transcoded from a lower bitrate or not.  Transcoding from CBR to --aps might make a larger file or a smaller file - the output file won't be the same bitrate as the input.

Could you explain why not?

To me it sounds reasonable; if you just beat all the difficult pieces out of the music by encoding it 128kbit cbr, then why should aps need high bitrates on such a file?

edit: why would we need tools to see this kind of things? Were all audiophiles here, right? So we'd hear if a file was badly made... 

Think of it this way - encoding to 128 kbps does throw out part of the sound, but in exchange it adds a considerable amount of noise, so if you try to encode it again, the encoder will attempt to encode the signal+noise.  Anyone else have a better explaination?  My brain doesn't seem to be working right now (I've gone to too many classes), so I can't think of a better one.

Tool for detecting transcoded mp3's

Reply #7
Everytime you enocde quantization noise is added. The more you times you encode a file the noise is added.  If this were not the case you could go from 320 - 128 without any problems it would just remove extra things, but instead it removes AND adds this noise, creating artifacts.
r3mix zealot.

Tool for detecting transcoded mp3's

Reply #8
Thanks for explaining, Chun-Yu. I understand what you mean.

Lord of Stars: your statement is the other way around it seems. I can see that for a 128kbit file (which will have lots of quantization noise), re-encoding to aps will not save bits. Aps aims at a lower noise floor and will thus not discard the noise that got into the file when encoding to 128kbit. It will do its best to encode all that.
But going from 320 to 128, you are actually raising the acceptable noise level for the encoder. The noise in the 320kbit file is lots lower than the 128bit encoding process will care about. So there shouldn't be artifacting I'd think. (Ofcourse you're still saver using the original wave...)

Tool for detecting transcoded mp3's

Reply #9
Quote
Quote
Lame is storing the source type in the INFO tag. So if it's originating from mp3, you can know it.


How do I find that Tag? Could you please give an example? I’m not sure what I have to look for. Also, not every Lame file comes with a Lame Tag! I encoded some mp3's a while ago (using Lame 3.92) and there has been no LAME Tag included. Then I have other files I encoded some weeks ago (Lame 3.92) and they do come with a LAME Tag...

Btw. what does 'unwise settings' indicate?

???
--alt-presets are there for a reason! These other switches DO NOT work better than it, trust me on this.
LAME + Joint Stereo doesn't destroy 'Stereo'

Tool for detecting transcoded mp3's

Reply #10
"unwise settings" is indicating that some unwise settings were used.

Tool for detecting transcoded mp3's

Reply #11
Quote
"unwise settings" is indicating that some unwise settings were used.

Well, that is pretty obvious...but what does that mean? I remember when I encoded a *.wav file (32khz) in an mp3 file (also 32khz) it said that I used 'unwise settings'...well I think in such case it would have been 'unwise' to encode it to 41,1khz...since the source was only 32khz...however I'm still not sure how to find a transcoded file...so I ask you one again, how can I read the Tag that tells me from what source the mp3 was encoded from . I transcoded some files but the Lame Tag doesn't look any different than the original (viewed with EncSpot).
--alt-presets are there for a reason! These other switches DO NOT work better than it, trust me on this.
LAME + Joint Stereo doesn't destroy 'Stereo'

Tool for detecting transcoded mp3's

Reply #12
Encspot doesn't show the full information.

Tool for detecting transcoded mp3's

Reply #13
Quote
Quote
Quote
Transcoding to --aps will tell you nothing about whether an MP3 is transcoded from a lower bitrate or not.  Transcoding from CBR to --aps might make a larger file or a smaller file - the output file won't be the same bitrate as the input.

Could you explain why not?

To me it sounds reasonable; if you just beat all the difficult pieces out of the music by encoding it 128kbit cbr, then why should aps need high bitrates on such a file?

edit: why would we need tools to see this kind of things? Were all audiophiles here, right? So we'd hear if a file was badly made... 

Think of it this way - encoding to 128 kbps does throw out part of the sound, but in exchange it adds a considerable amount of noise, so if you try to encode it again, the encoder will attempt to encode the signal+noise.  Anyone else have a better explaination?  My brain doesn't seem to be working right now (I've gone to too many classes), so I can't think of a better one.

So far so good - I did a test

I compressed a music sample with lame 3.92 in several ways:
-b 128; --alt-preset 128; -b192; --r3mix; --alt preset audiophile (here)
Then re-compressed using --alt-preset insane, re-compressed again using --alt-preset standard
I compared the results to a direct aps compressed mp3 with encspot.
I found two differences worth mentioning:

1. On recompressed files from low bitrate the number of 128 and 160 kbps frames was increased (from ~1/4 to ~1/2)

2. The number of short blocks was always much lower (decreased from 11% to 2% for "audiophile" that uses --noshort)

I think using this method it's hard to tell if something's reencoded if you don't have anything to compare with, but possible in some cases (at least to make a good guess), e.g. <128kbps files re-encoded to 320kbps and/or files created by encoders that don't use short blocks. It would be helpful to have experience on encoding behaviour of similar music to make a better guess. Maybe some more basic commandline than aps works better on this task.
Let's suppose that rain washes out a picnic. Who is feeling negative? The rain? Or YOU? What's causing the negative feeling? The rain or your reaction? - Anthony De Mello

Tool for detecting transcoded mp3's

Reply #14
Quote
Encspot doesn't show the full information.

So, what program does?

In addition, could you please give me some more information about the 'unwise settings'...I'm getting really curious about that now...I got a file (Lame 3.92) 192kbps. The source was 44,1 kHz as well as the output and it still stated that unwise settings were used...so I would like to know more about that

thank you
--alt-presets are there for a reason! These other switches DO NOT work better than it, trust me on this.
LAME + Joint Stereo doesn't destroy 'Stereo'

Tool for detecting transcoded mp3's

Reply #15
Those are trigerring the "unwise" flag:
*shortonly
*noshort
*-k
*scale left different from scale right
*nores
*noath
*athonly
*ath 0
*input sampling rate <=32kHz

Tool for detecting transcoded mp3's

Reply #16
thank you, Gabriel! So, what about my original question...how can I display the entire Lame Tag??? There must be some way, otherwise it would be pretty useless though...
--alt-presets are there for a reason! These other switches DO NOT work better than it, trust me on this.
LAME + Joint Stereo doesn't destroy 'Stereo'

Tool for detecting transcoded mp3's

Reply #17
Just write a tool that is reading the whole information. It is that easy...

More serisously, I am not aware of such a tool, even if it is quite easy to read the tag.

Tool for detecting transcoded mp3's

Reply #18
Quote
Just write a tool that is reading the whole information. It is that easy...

More serisously, I am not aware of such a tool, even if it is quite easy to read the tag.

there is really no tool out there...? I can't believe that! Well, we have to find someone that writes such a tool...would be pretty useful...
--alt-presets are there for a reason! These other switches DO NOT work better than it, trust me on this.
LAME + Joint Stereo doesn't destroy 'Stereo'

Tool for detecting transcoded mp3's

Reply #19
Quote
Just write a tool that is reading the whole information. It is that easy...

More serisously, I am not aware of such a tool, even if it is quite easy to read the tag.

Could you please check out the following thread: http://www.hydrogenaudio.org/forums/index....=ST&f=15&t=6709

thanks
--alt-presets are there for a reason! These other switches DO NOT work better than it, trust me on this.
LAME + Joint Stereo doesn't destroy 'Stereo'

Tool for detecting transcoded mp3's

Reply #20
I think that a statistical analysis of the spectrum of an encoded file could tell the difference between them...

Tool for detecting transcoded mp3's

Reply #21
I mean... a 320 k mp3 spectrum has certain characteristics, different from a 192k mp3 and from other bitrates, I myself have already downloaded mpc's which i think were transcoded from mp3's because i'm a spectrum fanatic and of course have good ears...

Tool for detecting transcoded mp3's

Reply #22
Quote
I mean... a 320 k mp3 spectrum has certain characteristics, different from a 192k mp3 and from other bitrates, I myself have already downloaded mpc's which i think were transcoded from mp3's because i'm a spectrum fanatic and of course have good ears...

well, any program would be a good...I mean it doesn't have to tell you 100% if the file has been transcoded but at least a guess would be nice (like EncSpot does for detecting the used encoder).
--alt-presets are there for a reason! These other switches DO NOT work better than it, trust me on this.
LAME + Joint Stereo doesn't destroy 'Stereo'

Tool for detecting transcoded mp3's

Reply #23
The problem is that spectrum is modified by second encoder.
Thus any combination will have different characteristics.
That's not even counting different types of music...

Can't you just use your ears to detect? (like j_j_moura?)
I've changed only because of myself.
Remember, when you quote me, you're quoting AstralStorm.
(read: this account is dead)

Tool for detecting transcoded mp3's

Reply #24
That's the worst part about MP3. That there actually are people out there who would transcode from 128 to 320.