Need help analysing files!

[duckandcover]

Hi all,

I got four shows from an unreliable source, and after analysis, I'm not sure about the material, the EAC outputs seems weird to me.
It may be obvious for some, but as I'm new into this stuff, I'd like to have your confirmation: I didn't want to spread lossy material...

(All recorded with a Sony DAT TCD-D3 & self-made micros)

Thanks! :thumbsup

Show 1:

Frequency Analysis (http://img4.imageshack.us/img4/2868/fa1.png)
Spectral View (http://img4.imageshack.us/img4/2593/sv1q.png)

Show 2:

Frequency Analysis (http://img39.imageshack.us/img39/981/fa2.png)
Spectral View (http://img39.imageshack.us/img39/8882/sv2f.png)

Show 3:

Frequency Analysis (http://img4.imageshack.us/img4/6426/fa3.png)
Spectral View (http://img268.imageshack.us/img268/3150/sv3.png)

Show 4:

Frequency Analysis (http://img4.imageshack.us/img4/6696/fa4l.png)
Spectral View (http://img268.imageshack.us/img268/107/sv4.png)

[Drgiggles1]

1st one to me appears lossless the other 3 lossy

[duckandcover]

That's a bad news, but thanks for the confirmation! :thumbsup

[Drgiggles1]

I would definately wait till a mod from here confirms. I'm new at this also. You can try zooming in on a 2 second frame in CEP and go into settings and change the colors. High energy to white, low energy to black and changing the gamma to 0.9. That will literally give you a clear black and white pic and be easier to decipher. If it is solid black above 16000 than you know its lossy.

[Drgiggles1]

I would definately wait till a mod from here confirms. I'm new at this also. You can try zooming in on a 2 second frame in CEP and go into settings and change the colors. High energy to white, low energy to black and changing the gamma to 0.9. That will literally give you a clear black and white pic and be easier to decipher. If it is solid black above 16000 than you know its lossy.
Sorry, I screwed up, High energy = black, low energy = white, and gamma = 0.4.

[duckandcover]

If it is solid black above 16000 than you know its lossy.

Solid white you mean? ;)

Thanks for the info man, I didn't know this CEP tool! (I guess I have to read the sticky threads...)
Ok, this is another confirmation: show 1 looks fine (see cap (http://img200.imageshack.us/img200/8789/56028510.png)) but others three are badly haircuted (2 (http://img18.imageshack.us/img18/1056/46748831.png),3 (http://img709.imageshack.us/img709/2767/97613038.png),4 (http://img18.imageshack.us/img18/7401/11930392.png))! It's not "plain" white, but I guess it's enough to make those lossy!

Thanks again! :wave:

[Drgiggles1]

Solid white you mean? ;)

Thanks for the info man, I didn't know this CEP tool! (I guess I have to read the sticky threads...)
Ok, this is another confirmation: show 1 looks fine (see cap (http://img200.imageshack.us/img200/8789/56028510.png)) but others three are badly haircuted (2 (http://img18.imageshack.us/img18/1056/46748831.png),3 (http://img709.imageshack.us/img709/2767/97613038.png),4 (http://img18.imageshack.us/img18/7401/11930392.png))! It's not "plain" white, but I guess it's enough to make those lossy!

Thanks again! :wave:
When it's zoomed in on a 2 second section it probably won't be solid. However after reading the sticky's for the last hour or so if at 16000 when it cuts off if you see blocks than you can be assured its lossy. If it seems to be a clean cut than its DAT. Thats what I take from the sticky's. For comparison I also have a topic in this section on going and I am more convinced that is DAT than MP3, just waiting on confirmation. As far as Cool Edit Pro goes yea there is a link in the sticky for download and they say it is best for spectral analysis. Read the sticky's, its tedious but very good info there. Also once you start playing around with a lossy file and a dat file and a lossless file you actually see the differences on CEP especially in Black and White. What is throwing me off on your samples are the FA's. They don't seem to have the big drop off at 16000. As far as FA's go also read the sticky's.

[Drgiggles1]

I just looked at the 3 new samples. If you scroll with the mouse on your files you can zoom in on a 2 second section compared to having a SA of the whole file, by just keep on scrolling and watching the time stamps in the bottom right hand corner window. Once your zoomed in to 2 seconds the spectral analysis at the cutoff will be obvious. If you see a bunch of squares, than its lossy.

[Drgiggles1]

This is what I mean by squares on the cutoff at 16000. This is a file I converted to MP3 to see. Here it is.
http://i733.photobucket.com/albums/ww335/DrGiggles1/MP3.jpg

[Five]

EAC is a crude tool for spectral analysis, that having been said I agree with the doctor 2-4 look bad. post a sample of the first show to megaupload or something if you can't figure audacity/cep correct settings so we can get a hold of it and have a look/listen.

ps check this link
http://www.thetradersden.org/forums/showthread.php?t=4288

[AndyLGR]

Many thanks for this thread, it saved me posting my own.

I have a wav show that I now know is converted from mp3. Everything above 16000 comes up white on CEP.

As a comparison I ran a proper wav file and this was black from 0 and gradually fading all the way up to a dark grey at 20000

[tonebloke]

Many thanks for this thread, it saved me posting my own.

I have a wav show that I now know is converted from mp3. Everything above 16000 comes up white on CEP.

As a comparison I ran a proper wav file and this was black from 0 and gradually fading all the way up to a dark grey at 20000


The cut-off for a lossy source made from eMPty3 will stand out like dogs......
As you saw from the wav you checked lossless will also stand out. The problem in the future will be to tell what everything inbetween is, either sourced from or, (which is almost impossible) what possible device did the recording.

[AndyLGR]

The cut-off for a lossy source made from eMPty3 will stand out like dogs......
As you saw from the wav you checked lossless will also stand out. The problem in the future will be to tell what everything inbetween is, either sourced from or, (which is almost impossible) what possible device did the recording.

I tried some other shows and here is what I came up with. I assume that image 4 (which I dont know the source of) is a lossy show:

1. Taken from a wav file

http://i366.photobucket.com/albums/oo103/AndyLGR/misc/wav-1.jpg

2. Taken from an mp3

http://i366.photobucket.com/albums/oo103/AndyLGR/misc/mp3source-2.jpg

3. Taken from a BBC podcast recorded as a wav file in cooledit

http://i366.photobucket.com/albums/oo103/AndyLGR/misc/Wavrecordeddirectfrompodcast-2.jpg

4. Unknown source

http://i366.photobucket.com/albums/oo103/AndyLGR/misc/unknownsource-2.jpg

[tonebloke]

1 is fine. 2 is crap; delete it. 3 is fine due to the source. 4 is probably sourced from MD with some work done to it - even with the cut-off at 14000. Some would call "this" particular spectral lossy, others lossless. I'd keep it.



"All views noted above are personal. Others may/will differ".