To all the phenomena who came here to say that an MP3 sounds immediately...

Piccinesco wrote:

CromaDiBrera, just a few days ago I saw a video from fonderie sonore in which the Grado Sr60 were recommended, specifically for their precision during listening.

This is a site where you can compare many models. I also had your Sony ones, but they are very dark and not very linear. http://www.headphone.com/pages/build-a-graph

Yes, in fact the Sony ones I mentioned emphasize the mid/lows a lot.

Thanks for the link, I'll take a look now.

2/6 with a Focal hi-fi system. I am a disaster.

No, simply at certain bit rates your ear probably favors the compressed format more than the uncompressed one. Maybe you hear it as softer and prefer it.

In any case, this test wanted to demonstrate something else...

I mean? What are you referring to?

Essentially, some people came onto the forum saying that MP3s can be recognized on the fly (the classic poseurs who bought 100-euro headphones and want to act like audio professionals). I had already conducted acoustic tests and physical comparisons between compressed and uncompressed formats, and I had reached the conclusion that compressed files, once a certain bitrate threshold was passed (which I identified starting from 192 kBit/sec VBR), were practically indistinguishable from uncompressed ones. This test demonstrates this. If it is difficult to recognize them by comparing the same track compressed in different compressed formats versus uncompressed, let alone recognizing them on the fly without any point of comparison! Beyond this, in some ways compression softens certain things, so it is very likely that due to a matter of subjective taste, the listener—even if they hear differences—is inclined to appreciate compressed formats more (it's a bit what happened to you, but also to me, as I recognized slightly more than half).

The MP3 compression system is a truly complex algorithm... For those interested, I can post something.

To summarize, it takes into account not only acoustic physics but also psychoacoustics: all masking phenomena regarding power and time, and frequency proximity. I'll just tell you that I have a book that explains how to implement the MP3 algorithm in Matlab... It is about 120 pages of extremely complex Mathematics, Acoustic Physics, Psychoacoustics, and Computer Science, and it speaks only of this (just to say that the 120 pages do not wander into the 'whys' but focus solely on the algorithm), but as I repeat to you, people come here and suddenly everyone is an expert. I wish I had found this test online sooner. I should mention that the tests I performed were done in recording studios with something like 30,000 euros worth of calibrated listening environments, and they yielded the same result.

I imagine it is quite heavy. Certainly not within my reach. Anyway, the point you made about taste is interesting. In fact, I selected the one that I felt was most pleasant.

What Stefanus replied is interesting, for a specific reason: you write that you selected based on listening pleasure, but (in quotes, though not even that much) the parameter that should have been evaluated was the cleanliness, the definition of the sound of each excerpt.

Perhaps by "pleasant" you meant exactly that version of the same piece that you perceived as more defined and consequently more pleasant, but it is the term itself that induces (at least me) to think that the approach was not the most appropriate.

I don't know if I am making myself clear.

I listened with walkman earphones (initially), and I sensed (not in every variant of the same piece) small "variations" in terms of perceived sound cleanliness (in general); I mean that in one of the three versions of a piece, it was possible to notice some passages being more "sharp," i.e., clean—for example, the mids/highs being more emphasized and therefore a sound that was overall less muffled, less loaded with mid/low frequencies, and less crystalline than the other versions.

No, not petty, perhaps I explained myself poorly. At least for me, the only way to try to find the uncompressed one was to try to identify possible imperfections. So I tried to evaluate based on what appeared sharpest to me. As Simone rightly says, however, sharpness is difficult to appreciate beyond a certain level and, for example, the sound of a drum cymbal might result in being sharper than the compressed one because perhaps in compression it loses details that made it less "constructed" and therefore comparable to the compressed file.

At least I think that what he/she wanted to say was this.

Ok, all clear.

Yes, the summary is exactly what Stefano expressed.

LucaCavaliere wrote:

io 3 out of 6 . . . with Sennheiser HD 590

i am a little proud to have caught the three where there is more acoustic "softness". Suzanne Vega first and foremost.

for the rest I am not happy: for some time now I have had suspicions about my ears

That Suzanne Vega track was the very first one to be encoded in mp3 format, precisely because it was "soft" (so it didn't stress the algorithm too much). So the legend goes.

--

alberto

Mmmm. I really think it's a legend. The MP3 compression algorithm takes into account the physical limits of the ear and the psychoacoustic limits of our brain, and does not pay attention to the type of audio signal it processes. Therefore, the argument should be viewed exactly the opposite way, starting from the sensory sphere and not from the characteristics of the file to be processed.

TheSimon wrote:

Mmmm. I really think it's a legend. The MP3 compression algorithm takes into account the physical limits of the ear and the psychoacoustic limits of our brain and does not pay attention to the type of audio signal it processes. Therefore, the discussion should be viewed exactly the opposite way, starting from the sensory sphere and not from the characteristics of the file to be processed.

I should preface that the song I have in mind is "Tom's Diner" and it's not necessarily the one included in your selection :-), by "soft" I meant exactly the song as perceived: for the singer's voice, the song itself, and the arrangement.

Anyway, I checked: it's not a legend but the truth, it's just that I understood exactly the opposite (oops). Suzanne Vega's song was chosen precisely because it was an excellent testbed for the algorithm.

https://en.wikipedia.org/wiki/Tom%27s_Diner#The_.22Mother_of_the_mp3.22

Ciao.

--

alberto

So it makes sense...

Well certainly. If the piece has very strong attacks, harmonic variety, or other psychoacoustic phenomena, it could be interesting to use it to test the algorithm. In this case, the discussion certainly makes sense.

I only guessed 1

How pathetic!

Ahahahahaha. It depends on what you listened with. Headphones? Speakers? What kind of headphones? What kind of speakers? If you don't have good listening equipment, it is impossible to pass a test like that.

Irina, I only guessed 2 of them. I am not much better than you. At least you are young and beautiful; I am old and senile, so I surely got those two right by chance.