Oh spectrograms? They're pretty easy to read. Standard middle-school science basics: always label your x- and y-axes in a plot. In the spectrogram you have time in seconds on the x-axis (so on the left is when the music track starts, on the right is when the track ends), frequency in kilohertz on the y-axis (so the bottom is 0 kHz and the top is __ kHz; 22.05, 48, or 96 kHz in these cases). Intensity is coloured based on the decibel scale and is labeled according to the legend on the right (the more purple/black it is, the quieter it sounds). Stereo audio has two channels, so the top graph is the left channel and the bottom graph is the right channel.
This is the same exact thing as the spectrogram visualiser in Foobar, but with the whole song in view as opposed to 30 seconds or however long it is in Foobar
It also happens to be a visualiser in Rockboxed devices too
Everyone here knows that CD-quality audio by
RedBook standards is 16-bit, 44.1 kHz, which produces a
maximum frequency of 22.05 kHz (half the sampling rate) from the Nyquist Theory.
Have 96 kHz or 192 kHz sampling rate? The maximum frequencies that can be reproduced are 48 kHz and 96 kHz respectively. From that, you can see
the plots in my previous post that the 24/96 and 24/192 plots only show significant information in the ~22 kHz range, which is
exactly where the cutoff frequency for RedBook standards is. The purple gradient haze and random horizontal line in the spectrograms are not relevant to the music.
Well okay, how do I know that? Because
if you have a proper 24/96 recording, you would get peaks above the ~22 kHz cutoff that you wouldn't get if you were just sampling at 44.1 kHz.
^ that's a legitimate 24/96 master, with true frequency content above ~22 kHz; this track happens to be a percussion track, which tend to have a lot of
natural ultrasonic frequency content
Too long; didn't read, the underlined portions are what's important and relevant for the current discussion. XD
[rule]
tgdinamo
If you really think you can hear a difference between HD sampling rates and RedBook standard sampling rates, I've prepared tests you can try for yourself
https://dl.dropboxusercontent.com/u/2816447/96vs441.zip (46 MB)
This is the same file as the spectrogram above, with legitimate frequency content over 22 kHz. The only differences between the two are the sampling rates.
Or how about 16- vs 24-bit?
https://dl.dropboxusercontent.com/u/2816447/16vs24.zip (21 MB)
Same song. The only differences between the two are the bit-depth (both are at their native 96 kHz sampling rate).
Chances are you heard a completely different master of the same song despite it being the same album. Same album ≠ same master. Lucky for you I prepared a different test for that as well.
https://dl.dropboxusercontent.com/u/2816447/Different_Masters.zip (21 MB)
Seal's "Crazy" song. Same song, different masters. I even downsampled the original HD version from 24/88.2 to 16/44.1 to make the test fair in terms of sampling rate and bit depth. The test should be easy compared the previous two and be sure to volume-match them (I already calculated the ReplayGain values) since they are different in volume.
And again, "It's the Masters, Damit!!!"
http://www.innerfidelity.com/content/its-masters-damit
Because Neil Young claims MP3 sounds so bad, here's another test (using the 24/96 file above)
https://dl.dropboxusercontent.com/u/2816447/24_96vs256CBRLAMEMP3.zip (20 MB)
Well that's MP3, what about iTunes quality (256 kbps AAC)? Same track
https://dl.dropboxusercontent.com/u/2816447/24_96vs256CBRAAC.zip (30 MB)