Question about using Trader's Little Helper to Spot Lossy

[Powerfan]

So I'm FINALLY organizing and categorizing my 300+ CD-Rs of shows, and I'm checking them out for lossy before archiving them.

I'm using EAC for making WAVs and the using Trader's Little Helper (version 2.4.1, build 160) to test them and so far I've tested 4 CDs for lossy and I've gotten the following results:

1, source of track cannot be qualified
2, track looks like CDDA with probability 100%
3, track looks like MPEG with probability 95%

What do these results mean?
I got the idea that "track looks like MPEG with probability 95%" means the disc is lossy and that "track looks like CDDA with probability 100%" means it's lossless, but what does "source of track cannot be qualified" mean? Is it lossless or lossy?

Thanks for any help!

[darkside02]

I'd like to add the following questions:

how accurate is Trader's Little Help recognizing lossy sources?
what if I got a track looks like CDDA with probability 99%? does that makes it lossy?

thanks

ps- I didn't even know TLH had that option :lol:

[Powerfan]

I'd like to add the following questions:

how accurate is Trader's Little Help recognizing lossy sources?
what if I got a track looks like CDDA with probability 99%? does that makes it lossy?

thanks

ps- I didn't even know TLH had that option :lol:

I got a track with CDDA 56%. So again, does that mean it's lossy?

[darkside02]

I got a track with CDDA 56%. So again, does that mean it's lossy?

Again, I didnt even know you could do that with TLH. I just test a cd and got a few tracks with 100% and some with 94% and it's just one source apparently so let's wait for someone more tech savvy to help us out :wave:

[Powerfan]

I was actually directing my last question to others but yea. I got one that is definitely MP3 sourced, but it reports it as 61% CDDA.

[Five]

the mp3 detection in TLH is a joke, its not much better than a ouija board. if its very very obviously lossy it will pick up on it but beware of false negatives and positives.

check this thread (http://www.thetradersden.org/forums/showthread.php?t=4288) for a bit of info, also take a look in a couple of the threads in this forum to get an idea what to look for.

[darkside02]

thanx Five

[Powerfan]

Thanks five. What would you recommend for easy lossy detection?

[Powerfan]

What would you recommend for easy lossy detection?

Never mind my last post, I saw the link you posted which I previously missed.

Thanks again!

[darkside02]

So, my question would be then, is there any way to check a file WITHOUT the spectrum analysis?

[rspencer]

No. There is no easy automated way. they're more useful for the negatives than positives. If it says it's clean, it's pretty much clean. But almost all software has a crappy percentage when it comes to false positives for lossiness.

[darkside02]

I guess I'll have to learn how to read spectrums, I've already tried a few times but I just don't get it.

thanks :)

[michael123]

Hi rspencer,
What is meant by "all software has a crappy percentage when it comes to false positives for lossiness. " ?
is there software which make them clean?

[tonebloke]

I am not tech savvy and am resistant to change but one thing I have learned very slowly is that you need to "learn" how to use various bits of software; like TLH, Foobar, CEP, EAC etc, to work with your boots.
Awhile back on TC I took umbrage with a member there for going on about Specs and Freqs to prove lossy/lossless. Well the worm has turned and I have to give him an apology (I will once I find him). The best way to find out if the show you have is OK or not is to look at a spectral graphic and a frequency analysis. Sure it takes time to learn and figure out what you're looking at but it's worth it.
There are good guides here (thanks Drgiggles1) and there are members who will help out enormously when asked.

As "Molly" used to say, "do yourself a favour" and learn and use software that's out there. I am.

[AAR.oner]

Hi rspencer,
What is meant by "all software has a crappy percentage when it comes to false positives for lossiness. " ?
is there software which make them clean?

he means any program that analyzes and detects lossiness automatically...i know TLH has the feature [which is a GREAT windows program other than that]


to detect lossiness, you need to be able to visually analyze the file(s) yerself using Frequency Analysis (FA) & Spectral Analysis (SA)...there are various programs which can do one or both, some better than others...but any "built-in automatic lossy-detection" feature is not gonna be accurate

[rspencer]

Additionally, I meant that any software tends to do better proving lossless than lossy. If it says it's lossless, usually it's right. But far too many factors can produce a "lossy" designation when the recording in question is not in fact lossy.

[TheWarlock64]

http://s28.postimg.org/utyvy7acc/littlefeat_aragon1975_spectral.jpg ok i ran the first track thru TLH with a verbose accuracy of 0.40 where "0" is the most acc. and just listened to enough of the first track to tell you all the possibility that this little feat aragon ballroom 1975 source has/had a virgin lossy lineage is about 99.5%. take a look at the spectral & it hurt my ears right off the bat & for some1 that has had 4 instances of bilateral mirongotomy(tube implantation in eardrums) throughout his 5 decades, i can assure every1 this info is valid and albix should not be held liable for any coding errors that are to be....

[jpeace]

The problem with this whole process is that it can all be fooled. Taking a lossy source and using harmonic synthesis to restore the high end will certainly fool TLH (I've tried it), and will also make spectral look lossless. I tried this approach and sent the spectral to one of my MQR label mates who is skilled at reading spectral views, and he couldn't tell it was an mp3 to begin with.