These are discussions made over and over again with some audiophile friends, and the end result has always been the same: they will not make headway, at least for now, for a long series of reasons.
The first reason is that quality costs, and it costs a lot!!
Between the lines, in the previous article, reference is made to very expensive hardware, both from the point of view of audio reproduction (a 400 euro player), and from the point of view of listening (listening ranges from 200 euros up to 2500!).
In fact, it is perfectly useless to have a high-resolution audio file if you listen to it on iPhone earphones. To demonstrate this, I challenge anyone to recognize the qualitative difference between 192 kbps MP3 files and a high-resolution lossless file.
Beyond the consumer scope, even in the professional field, it becomes very difficult to distinguish between the two types of files.
It has been shown that, in some contexts—it seems incredible to say—MP3 has received greater consensus than lossless files, demonstrating a smoother listening experience for the ear. (I did not conduct these tests myself, but I had the chance to read them; they are purely statistical tests carried out on a sample of 1000 people).
Ultimately, defining the MP3 compression algorithm as something that reduces the size and audio quality of the final file is, in my opinion, absolutely reductive. Objectively, there is a loss of data, as we move from uncompressed formats with a certain weight to much lighter compressed files. The MP3 algorithm was born out of necessity, at a time when available bandwidth could not guarantee a usable data stream for the end user. It was therefore thought to invent a compression method that would significantly lighten the size of the file to be transferred without, however, perceptibly compromising the final quality of the audio file. It is precisely this latter idea that makes the definition usually given to the MP3 algorithm reductive. The algorithm is based mainly on psychoacoustic reduction criteria, meaning it removes those frequencies within the file—or more generally, those signal components—that our ear is unable to perceive. We are talking about masking phenomena. An example could be related to very fast attack transients. If we record the hit on a snare drum of a drum kit, it will have a very fast attack and a very high volume. Probably if there were other frequencies at play, for example those of a flute heard at a much lower volume, the frequencies recorded in a certain vicinity of those belonging to the drum sound would be hidden from our ear. Not only those following the drum hit, but also some frequencies that precede, in millisecond order, the violent attack of the drum. Since these frequencies cannot be heard, the MP3 algorithm, after performing a complete analysis of the audio file, excludes them, dedicating a smaller number of bits to these hidden frequencies for conversion. This is just one of many modifications, in a psychoacoustic sense, that the MP3 algorithm performs on the file.
At certain resolutions, by dedicating a good number of bits to conversion, for example 320 kbps, the MP3 algorithm produces psychoacoustically almost impeccable copies of the original uncompressed file.
To listen to the cancellation product of the MP3 file compared to the uncompressed file, we would have to put them out of phase, so as to be able to hear what the MP3 algorithm has gone and deleted.
I downloaded a 74 MB high-resolution FLAC audio file 96 kHz/24 Bit (Vivaldi: Recitative and Aria from Cantata RV 679, "Che giova il sospirar, povero core" Tone Wik & Barokkanerne (period instruments)) from the site www.2l.no/hires, imported it into a new project, and compressed it into 320 kbps MP3 format.
I propose to you a part of the audio compressed in mp3 to understand what we are listening to, and immediately after, the listening of the MP3 cancellation product between the compressed and uncompressed file out of phase. Anyone who wants to listen to the high-resolution file can download it from the aforementioned site (There are also many other formats).
Reference file @320 Kbit/s
http://www.pianoconcerto.it/condivisione/hiresaudio/mp3_320.mp3
Cancellation product between hires and mp3 files @ 320 Kbit/s
http://www.pianoconcerto.it/condivisione/hiresaudio/cancellazione.mp3
Let's do an experiment that is much easier than the previous ones; it could also be interesting to understand a bit whether this mp3 works as well as it promises. Unfortunately, on the forum we cannot make listening differences for high-resolution audio, streaming is not allowed, and one would be able to understand very quickly—though I won't say which is the hires file—so let's do it with MP3s (VBR encoding): recognizing the bitrate
of an mp3 file among 5 available listens: 48 - 64 - 96 - 128 - 192 Kbit/s.
For each question you have 5 files to listen to. You must assign the bitrate to each listening file.
Ex. Listen 1 - A) 192 kbit - B) 96 kbit - C) 320 kbit - D) 128 kbit
I am telling you right away that if you do not have good listening equipment, you probably will not be able to hear the difference starting from 96 Kbit and up; as a qualitative reference, I will provide an mp3 file with VBR encoding at 320 Kbit/s. For those who wish to download high-resolution references, they can do so on the website mentioned above; if it asks for a username and password, they are listed on the site www.2l.no/hires -> Username: 2L and Password: 2L
For the first listen, the file is: http://www.lindberg.no/hires/test/7041888513520_01_13_96kHz.flac
For the second listen, the file is: http://www.lindberg.no/hires/test/HD7041888514725_01_01.flac
For the third listen, the file is: http://www.lindberg.no/hires/test/2L-064_stereo96kHz_01_08.flac
For the fourth and final listen, the file is: http://www.lindberg.no/hires/test/2L50SACD_tr01_stereo_96.flac
Listen 1:
Reference @320 kbit: http://www.pianoconcerto.it/condivisione/hiresaudio/rif1.mp3
A ) http://www.pianoconcerto.it/condivisione/hiresaudio/1a.mp3
B ) http://www.pianoconcerto.it/condivisione/hiresaudio/1b.mp3
C ) http://www.pianoconcerto.it/condivisione/hiresaudio/1c.mp3
D ) http://www.pianoconcerto.it/condivisione/hiresaudio/1d.mp3
E ) http://www.pianoconcerto.it/condivisione/hiresaudio/1e.mp3
Listen 2:
Reference @320 kbit: http://www.pianoconcerto.it/condivisione/hiresaudio/rif2.mp3
A ) http://www.pianoconcerto.it/condivisione/hiresaudio/2a.mp3
B ) http://www.pianoconcerto.it/condivisione/hiresaudio/2b.mp3
C ) http://www.pianoconcerto.it/condivisione/hiresaudio/2c.mp3
D ) http://www.pianoconcerto.it/condivisione/hiresaudio/2d.mp3
E ) http://www.pianoconcerto.it/condivisione/hiresaudio/2e.mp3
Listen 3:
Reference @320 kbit: http://www.pianoconcerto.it/condivisione/hiresaudio/rif3.mp3
A ) http://www.pianoconcerto.it/condivisione/hiresaudio/3a.mp3
B ) http://www.pianoconcerto.it/condivisione/hiresaudio/3b.mp3
C ) http://www.pianoconcerto.it/condivisione/hiresaudio/3c.mp3
D ) http://www.pianoconcerto.it/condivisione/hiresaudio/3d.mp3
E ) http://www.pianoconcerto.it/condivisione/hiresaudio/3e.mp3
Listen 4:
Reference @320 kbit: http://www.pianoconcerto.it/condivisione/hiresaudio/rif4.mp3
A ) http://www.pianoconcerto.it/condivisione/hiresaudio/4a.mp3
B ) http://www.pianoconcerto.it/condivisione/hiresaudio/4b.mp3
C ) http://www.pianoconcerto.it/condivisione/hiresaudio/4c.mp3
D ) http://www.pianoconcerto.it/condivisione/hiresaudio/4d.mp3
E ) http://www.pianoconcerto.it/condivisione/hiresaudio/4e.mp3
Results in a few days, try to participate in large numbers, it is not a challenge, it is just a test.
