Need help understanding the frequency spectrum in some audio files

[Juan Uribe]

The following is the frequency spectrum of the same song by Savatage- St Patrick's . the first is the Japanese press CD ,the second is the remastered "Narrated version" CD.

It seems that the blue color (and may be the the red-orange as well) look denser in the "Narrated Edition" CD . does that mean that the "Narrated version" CD, has more sonic resolution?

 

[Juan Uribe]

another question:


I use a program called Fakin' The Funk to detect the quality (true bit rate, etc) of some of my audio files.

from time to time It shows cuts in the Frequency Spectrum in the file under check

like that:

viny rip of Iron Maiden killers

there is a cut on 20.000.. another on 40.000


I found some other cuts, on other audio files that were ripped from CD's .

any explanation?

 

[AnalogSteph]

It seems that the blue color (and may be the the red-orange as well) look denser in the "Narrated Edition" CD . does that mean that the "Narrated version" CD, has more sonic resolution?

It appears the noise floor is louder, so maybe it's just generally louder / more compressed, as generally the case for remasters. It's definitely more than just "get the CD audio and make it louder" though, as only the original exhibits some interference in the 15.5-16 kHz range (horizontal from video monitors?) as well as apparently worse ADC anti-alias filters. Back in the 1980s mastering CDs for local distribution from 1st generation tape copies wasn't entirely unheard of. In addition, recording equipment with several different sample rates was in use at the time (from 44.056 kHz to 48 and even 50 kHz), and if you needed to do a sample rate conversion, you pretty much had to go D/A and A/D, potentially making the signal vulnerable to interference. High-quality SRC ASICs (let alone software SRC) still were years off by this point, and we can assume that a lot of equipment would have been plagued with Pin 1 Problems still.

Your "cuts" are steady tones of some kind that were recorded along the desired content. It is odd to be seeing them at pretty much exactly 20 and 40 kHz. Could be some sort of power supply interference - the 40 kHz one in particular looks rather fuzzy. It'll be hard to tell at which point exactly they were picked up.

 

[kongwee]

It just show the dB in all frequency in a single timeline. Before 0:30, the red color is lesser because usually it is the intro. More "quiet". After that the more sound kick in, the red extend to more upper frequency.

Why the same song the spectrum frequency are different? This is mastering. I just check the two album. The Narrated version has addition tracks, so the mastering is different from the first.

The Narrated version has greater red and blue mean the tracks is being pushed up by dB. Could be more limiter at work. Or conventionally more recognised as "compressed".

In the end, it is not increase in resolution. You just can heard more in that frequency range.

You can check the DR or LUFS, you will get a different value.

 

[Juan Uribe]

Thanx for the comments


the dynamic range of both :



The Japanese press:

--------------------------------------------------------------------------------
Statistics for: 06-St. Patrick's
Number of samples: 11382504
Duration: 4:18
--------------------------------------------------------------------------------

Left Right

Peak Value: -1.87 dB --- -1.96 dB
Avg RMS: -14.64 dB --- -15.38 dB
DR channel: 8.41 dB --- 8.78 dB
--------------------------------------------------------------------------------

Official DR Value: DR9

Samplerate: 44100 Hz
Channels: 2
Bits per sample: 16
Bitrate: 714 kbps
Codec: FLAC
================================================================================


The Narrated version:


--------------------------------------------------------------------------------
Statistics for: 15/31-St. Patrick's
Number of samples: 11281368
Duration: 4:16
--------------------------------------------------------------------------------

Left Right

Peak Value: -0.05 dB --- 0.00 dB
Avg RMS: -15.27 dB --- -16.35 dB
DR channel: 10.72 dB --- 11.45 dB
--------------------------------------------------------------------------------

Official DR Value: DR11

Samplerate: 44100 Hz
Channels: 2
Bits per sample: 16
Bitrate: 798 kbps
Codec: FLAC
================================================================================


Astonishingly The Narrated version has more dynamics ! ?

 

[DVDdoug]

Astonishingly The Narrated version has more dynamics ! ?

The way it was measured, yes. There is more than one way to measure dynamics and none is perfect. The narration probably has a lower RMS level than the music so including it could result in "more dynamics".

Sometimes the "dynamic range" is calculated as the crest factor (peak-to-RMS ratio). That doesn't seem to be what you're using, but one short-duration peak that's to short to sound "loud" will make a high crest factor. (Making a vinyl record or an MP3 will also often increase the crest factor without affecting the sound of the dynamics.)

As a rule, you can't "look at a graph" to determine sound quality. You just have to listen. i.e. You can make a "pretty spectrum" with audible distortion and noise. Harmonic distortion (and often noise) adds to the high frequencies. And frequency content that extends beyond the hearing range doesn't matter. When you are trying to compare quality, it's best to match the loudness (RMS or LUFS). In your case, the narrated version seems to be only slightly louder so it may not be that important.

But in this case, the most important thing is if you want to hear the narration or not... You're not going to simply choose the "better sounding" version.

 

[ebslo]

It seems that the blue color (and may be the the red-orange as well) look denser in the "Narrated Edition" CD . does that mean that the "Narrated version" CD, has more sonic resolution?

The dark blue and black (leading to appearance of lower density blue) is < -105dB from 16-bit source material. My interpretation is the original was not dithered correctly. Properly dithered source material should have somewhat higher resolution WRT improperly dithered, but IMO the difference would only be audible under very specific and rare (maybe only contrived) circumstances. In terms of what you see, I think the smooth dense blue areas of the "narrated" version is dither noise, you are not directly seeing extra musical content.