Strange (for me) frequency analysis

Hi,
I have recently acquired a very nice Cohen show, for which unfortunately it came with no lineage whatsoever . Of course this means I cannot post it here, but I was wondering if someone could help me understand the strange frequency analysis results I get and let me know if there is lossy ancestry. There is definitely a drop, but then it looks fine at higher frequencies .

Here is the frequency analysis and the spectral view from EAC (let me know if you need anything else). To me it looks like somebody hacked away a range of frequencies, but would anyone know how that could happen?

Any help would be much appreciated and I thank you in advance.

[wazoo2u]

I would agree with you.

It's a totally FUBAR'ed audio signal. Not a product of a Codec. Why would someone do this ?? Only through ignorance or incompetence, not knowledge or skill.

Looks like a botched de-hiss/NR job. But RainDawg is the guy to ask about this.

[RainDawg]

Uh-huh, that was my first impression upon looking at it too. Someone noticed some hiss or noise (likely someone who aslo doesn't know the difference between the two) and tried to remove it. To this person, I am sure this was a fantastic result...

This appears to be a prime example of why layman should not tinker with NR, and if they do, why they shouldn't seed their results.

[paddington]

it looks like a friggin notch filter to me. I wonder of there was some terrible 11khz squeal on the tape or something. Or something sweeping between 10k and 12k, where he notched it.

I take it you have no idea how it was recorded, right?

[Five]

I've seen this twice, one was a cdr transferred MiniDisc a client gave me recently, the other was a transfer from a VHS tape I made myself. I'm stumped as to how this "death star groove" happens, but on the bright side this doesn't look like a lossy source to me.

Thanks KingBongo for all the rare Cohen you continually unearth.

[ssamadhi97]

Err? What's with all this Noise Reduction talk here? I see NO evidence whatsoever for NR here.

Actually it's blatant aliasing introduced by bad resampling from 22.05kHz to 44.1kHz (check how the content above 11.025kHz is more or less just a flipped image of what's below). The "dropout" is there because the signal was lowpassed at 10kHz before this resampling process for some reason. Maybe a lossy audio encoder operating at a rather low bitrate. Who knows.

"If it sounds like ass, it's probably sourced from a 56kbps 22050Hz mp3" >_>

Originally Posted by ssamadhi97

... (check how the content above 11.025kHz is more or less just a flipped image of what's below) ...

I had noticed this but couldn't account for it. Well I'll be damned.

[wazoo2u]

Originally Posted by ssamadhi97

Actually it's blatant aliasing introduced by bad resampling from 22.05kHz to 44.1kHz (check how the content above 11.025kHz is more or less just a flipped image of what's below). The "dropout" is there because the signal was lowpassed at 10kHz before this resampling process for some reason. Maybe a lossy audio encoder operating at a rather low bitrate. Who knows.

"If it sounds like ass, it's probably sourced from a 56kbps 22050Hz mp3" >_>

Nice lesson ssamadhi. Thanks. I didn't notice the duplication in the freq graph, but to be honest, I was just transfixed by the huge notch

Good to always be learning the fine points.... Thanks for sharing an excellent observation.

Thanks for all the info... it went a little bit beyond what I can understand (ssamadhi97's answer seems to find consensus, but I >think< I managed to spot it as it flew really high above my head ).

The original source is supposed to be SBD, but from what you say someone did a conversion somewhere and/or resampled the signal and totally botched it... sad, really, it's a magnificent show. I guess I can consider it as NOT MP3-sourced but lossy in other ways, right?

[paddington]

it could be labeled

"FUBARed by junior's mastering studio"

[ssamadhi97]

Originally Posted by KingBongo

I guess I can consider it as NOT MP3-sourced but lossy in other ways, right?

Oh I definitely would not rule out mp3 compression before the bad resampling step.. just can't seem to come up with a n00b-friendly way of figuring out wtf is going on right now.

[Five]

Originally Posted by ssamadhi97

Oh I definitely would not rule out mp3 compression before the bad resampling step.. just can't seem to come up with a n00b-friendly way of figuring out wtf is going on right now.

How can I reproduce this groove in a controlled environment?

thanks for the lessons.