Weed out the BAD MP3's

[Snarkridden]

Hi,

So many of the downloaded MP3's are of variable quality, due to many factors, one of which is the encoder used, a utility I have used for some time now is ENCSPOTwhich produces an Explorer like listing showing many features of each MP3 in any specific folder.

Interested then try...

http://www.guerillasoft.com/Encspot/

Sorry I'd post it as a link if I knew how...

Snark.

[goldie]

Originally posted by Snarkridden
Hi,

So many of the downloaded MP3's are of variable quality, due to many factors, one of which is the encoder used, a utility I have used for some time now is ENCSPOTwhich produces an Explorer like listing showing many features of each MP3 in any specific folder.

Interested then try...

http://www.guerillasoft.com/Encspot/

Sorry I'd post it as a link if I knew how...

Snark.

Hmmmm........veeeeerrryy innnnterrrrresting Snarkridden.

Thanks!

[theknife]

this is indeed very interesting...getting all this info on the mp3's you know and love is a little disconcerting - like finding your girlfriend's diary or something

now I'll have to listen to some of the ones I've checked and see if the quality is as bad or as good as this little proggy says it is...

[Snarkridden]

Yes, it makes you listen again, especially when you see the old Xing encoder shown, a notorious distortion device, but the new one seems OK..

Hi GR, off you go... tell us all about your bad ones eh?

Snark.

[goldie]

Originally posted by Snarkridden
Yes, it makes you listen again, especially when you see the old Xing encoder shown, a notorious distortion device, but the new one seems OK..

Hi GR, off you go... tell us all about your bad ones eh?

Snark.

Hey Snark - the only bad ones are the one's I've dl'd off of other users. Of course, not you or Pod LOL!!

I've burned mostly all my mp3's into a library so not much left to play with.

I don't have a clue on how to interpret this data and I'm too lazy to do the print screen thingy. The qualities of all three are in the green. Here's some text though.....

Pod_away's Enya (love that Enya)

Enya - Only One Time [remix].mp3
--------------------------------

Bitrates:
----------------------------------------------------
320 |||||||||||||||||||||||||||||||||||||||| 100.0%
----------------------------------------------------

Type : mpeg 1 layer III
Bitrate : 320
Mode : stereo
Frequency : 44100 Hz
Frames : 13223
Length : 00:05:45
Max Reservoir : 0
Emphasis : none
Scalefac : not used
Complete : yes
Encoder : Blade

--[ EncSpot 1.0 ]--[ http://www.guerillasoft.com ]--



Here's that Pilot I dl'd off you on OpenNap:

Pilot - Magic(256k).mp3
-----------------------

Bitrates:
----------------------------------------------------
256 |||||||||||||||||||||||||||||||||||||||| 100.0%
----------------------------------------------------

Type : mpeg 1 layer III
Bitrate : 256
Mode : stereo
Frequency : 44100 Hz
Frames : 7147
Length : 00:03:06
Max Reservoir : 125
Emphasis : none
Scalefac : 12%
Complete : yes
Encoder : Lame

--[ EncSpot 1.0 ]--[ http://www.guerillasoft.com ]--



Here's a Luther I ripped (CBR that is):

Luther Vandross - Give Me The Reason.mp3
----------------------------------------

Bitrates:
----------------------------------------------------
192 |||||||||||||||||||||||||||||||||||||||| 100.0%
----------------------------------------------------

Type : mpeg 1 layer III
Bitrate : 192
Mode : stereo
Frequency : 44100 Hz
Frames : 10925
Length : 00:04:45
Max Reservoir : 334
Emphasis : none
Scalefac : 1%
Complete : yes
Encoder : FhG (fastenc or mp3enc)

--[ EncSpot 1.0 ]--[ http://www.guerillasoft.com ]--


I KNOW you use Lame, Pod uses Blade and I use MMJB (7.0)




Cripes, almost forgot to mentioned that I had EncSpot all along......all I had to do was dust it off a bit <snicker>

[Snarkridden]

Well What ARE you like?

Cripes, almost forgot to mentioned that I had EncSpot all along......all I had to do was dust it off a bit <snicker>


Hey but you ahaed of me there, never bothered to print anything out before, well done


Watch your post box...

Snark..

[VWguy]

Originally posted by Snarkridden
Hi,

So many of the downloaded MP3's are of variable quality, due to many factors, one of which is the encoder used, a utility I have used for some time now is ENCSPOTwhich produces an Explorer like listing showing many features of each MP3 in any specific folder.

Interested then try...

http://www.guerillasoft.com/Encspot/

Sorry I'd post it as a link if I knew how...

Snark.

Very nice! I ran the proggy on about 2000 songs(had to stop it, it was taking too long to go through all of them) and all of the songs I encoded with the Lame encoder (388)were in the green. The songs in the red (bad) were all downloads, and almost all were encoded with xing or Blade.

Fruhauf (fhg)was ok to good.
bitrate was also a factor.

[TankGirl]

Thanks for the link, Snarkridden, a very nice utility!

Originally posted by VWguy
Very nice! I ran the proggy on about 2000 songs(had to stop it, it was taking too long to go through all of them) and all of the songs I encoded with the Lame encoder (388)were in the green. The songs in the red (bad) were all downloads, and almost all were encoded with xing or Blade.

Fruhauf (fhg)was ok to good.
bitrate was also a factor.

You have to be careful not to trust the results too much though as it is possible to fool this proggy by artificially 'upgrading' the mp3 bitrates - a practice that has unfortunately become more common as quality has become a trading factor. To separate a good genuine rip from a doctored one is a difficult technical task beyond the scope of Encspot.

- tg

[Snarkridden]

Hi Tg,

I thought it was only Cats that got doctored?

I presume you mean since quality is NOW a trading factor, that some unscrupulous individuals will be decoding & re-encoding at a false higher bit rate?

That should be easy to spot, by doing a FFR analysis of the spectrum, a genuine original will have the full spread, whereas the re-encoded one will be limited to the spectrum signiture of the original encoder.

To the point about the "reds" in a previous post, Yes I noticed that too as typical of the old Xing encodes, but the new one is supposed to be a lot better, though the Blade was always better than the Xing, it was always terribly slow.

I still reckon the fastest encoder is the FhG in AudioActive Pro.

There was in the other thread on MP3 quality, a throwaway comment about Sound Forge and its handling of MP3, I cannot find that thread entry again, so if the original poster would identify themselves here, I'd be pleased to have some review information on this reported problem...

Snark..

[TankGirl]

Hi Snark!

Originally posted by Snarkridden
I presume you mean since quality is NOW a trading factor, that some unscrupulous individuals will be decoding & re-encoding at a false higher bit rate?

Yep, that is unfortunately the case. It is easy enough to do an arbitrary MP3-WAV->MP3 conversion with MusicMatch, and then there are special converter programs that make the job even easier. Some 'mp3 doctors' are genuinely ignorant of the technical aspects - they just think that this new wonder conversion program performs some digital magic to their 128 kbs files so that they end up being fine 256 kbs files - and they may honestly not be able to tell the difference with their crappy PC speakers. Then there are the greedy traders who know what they are doing but do it anyway to 'manufacture' more acceptable trading material in hope of getting some genuine quality in return.

Originally posted by Snarkridden
That should be easy to spot, by doing a FFR analysis of the spectrum, a genuine original will have the full spread, whereas the re-encoded one will be limited to the spectrum signiture of the original encoder.

It surely is possible to spot a faked high-bitrate mp3 if you have the original CD or a quality rip of it as a reference, high-quality audio gear and good ears to do comparision tests, spectrum analysis tools and the expertise to use them and so on... but what's the point of trading if you already have the original or a good rip of it? Forged bitrates are like forged money... specialists may be able to spot the difference easily but normal users watch just the simple bitrate numbers and act accordingly - knowingly or unknowingly polluting the shared mp3 pool with faked quality. On anonymous filetrading networks like Gnutella and AG the quality problem is even worse as you lose the advantage of knowing reliable sources for quality rips.

- tg

[Snarkridden]

Hi Tg,

I like the little wave...

You know you have really spoiled things for me now, by saying that about trading factors, I supoose in my innocence I didn't think people would bother with such things, I shall now have to look very carefully at my Acoustic Alchemy and Ambrosia collections, built up over the year mostly at 320k sample rate.

Strangely just recently a new CD appeared in the set, never found it before, then there it is complete all tracks at 320k, none at lower rates as you would expect from normal multi-sourcing, so looks like you could be right.

On the other hand, what better way to discredit a source of good quality material, in the battle against sharing, could this be another example of the weapons of war?

Will we ever find out... Back to the spectral analysis job on all
the suspects...

Snark..

[TankGirl]

Hi Snark! (<---- this just to delight you!)

Originally posted by Snarkridden
Strangely just recently a new CD appeared in the set, never found it before, then there it is complete all tracks at 320k, none at lower rates as you would expect from normal multi-sourcing, so looks like you could be right.

A recent case I know for a fact: a long-time collector/trader suddenly offered all his material (tens of gigabytes collected over long periods of time in mixed bitrates from various sources) to a new trading partner in 256 kbs. Suspicious eh? Talk about wholesale quality!

Then there are also similarly sad cases of lost quality libraries - also brough about these handy bitrate conversion utilities. People have been happily reporting how they managed to make more room to their full HD by converting all their 'stuff' to 128 kbs... and afterwards they wonder why their old trading partners suddenly lose all interest in their library....

Originally posted by Snarkridden
On the other hand, what better way to discredit a source of good quality material, in the battle against sharing, could this be another example of the weapons of war?

Polluting the mp3 pool with faked quality files would indeed be a potential weapon in the hands of the RIAA against p2p sharing. If bitrates become an unreliable measure for quality at least some people might be motivated to buy the CD instead of grabbing it from p2p. The greedy traders are already doing a good job for the RIAA by pushing their forged merchandise into circulation.

The good news is that certain technical countermeasures can go a long way in ensuring the quality of p2p-distributed material. The basic idea (as I mentioned in this thread on P2P) is to have trusted sources publish unique signatures for their quality rips and to have p2p software to support the verification of these signatures. Similar measures would help to keep p2p-distributed software free from virii and other unpleasant additions.

- tg

[MikeHunt]

http://privatewww.essex.ac.uk/~djmrob/mp3decoders/


go to FaQ's and encoder ratings section. Then you will know!

[Snarkridden]

Thanks Mike,

Interesting site, and quite local to me too...

Nice to know I have been using one of the best coders for a while now.. LAME.

Snark..

P.S vers 3.91 is out now

[MikeHunt]

Snarkey...thanks for encspot...nice referral..great tool!
And you just can't beat Brit education when it comes right down to it.

[thinker]

Just dl'ed this and going to use it soon. Thanks for the heads-up.

[thinker]

Hmmm...out of 75 files I have left (all were downloads), 31 were rated bad, and all bad were encoded with Xing or Blade. So...

bad overall - 31 of 75 - 41.333%
bad Xing - 24 of 26 - 92.308%
bad Blade - 7 of 10 - 70.000%
bad FhG - 0 of 39 - 0.000%

[Snarkridden]

I dare NOT tell you how my tests fared, but I know one thing, i'm sharing a lot less now than I was?

Standby for blasting over the re-encoding issue, when I can get my act together with all the FFR tests (pictures) I will post the results of some serious tests done with Acoustic Alchemy tracks downloaded from different sources, different users, compared to my own direct CD rips.

Tg, you opened up a can of worms with your re-encoding post.

Snark..