[redrol]

I know how it works. I don't see how this example would do anything, this file is not loud enough to need 24bit. Im not interesting in listening to this sample, im more interested to know if its real or not.

 

[71 dB]

...this file is not loud enough to need 24bit.

16 bit can do over 90 dB of technical dynamic range and up to 110-120 dB of perceptual dynamic range (using shaped dither). What music needs more than that? Symphony for space rockets and artillery?

 

[gregorio]

Excuse me if I sense a bit reluctance to understand what I mean.

No, I did understand what you meant and I agree with it. I responded to your questions/points literally simply to highlight how they can (and probably will be) misunderstood. My apologies that I didn’t make that clear.

However, even in those cases we might have a 200 seconds long track for example of which say 10 seconds are pauses/silences/etc. meaning 95 % of the time masking happens.

I know what you mean, I’m just questioning how you’re presenting it. Your response above could be interpreted as meaning that 5% of the time dither is not masked and therefore might/could be audible. In practice, masking is irrelevant because even if it’s not masked, it’s still inaudible at reasonable listening levels. And incidentally, I don’t think white noise can be masked can it?

Well, I certainly did NOT mean you working on a track in studio!

There are others besides studio engineers who listen/test this way. Some audiophiles do, some reviewers do. These are also “real life listening scenarios”. Again, I’m not questioning what you meant, just how you presented it. Generally I find it better to say “reasonable listening levels”.

G

 

[bigshot]

I know how it works. I don't see how this example would do anything, this file is not loud enough to need 24bit.

That's kind of the point. This is a test of something that just looking at it in terms of how digital audio works, you know that 24 bit isn't needed. It's a test where the result was a foregone conclusion before they even did the test.

I'm sure the test files are legit.

 

[gregorio]

What music needs more than that? Symphony for space rockets and artillery?

Not even then! The 1812 overture has artillery in it (canons) but still 16bit is more than enough!

G

 

[danadam]

I see what you mean. Im not sure how to find out if this is real or not.

I'm not entirely sure what you mean by "real". As for the bit-depth difference between the files, they have fade-in and fade-out (unfortunately, in the context of a test), so you can load them in Audacity and switch the view to dB scale:

 

[gregorio]

I know how it works. I don't see how this example would do anything, this file is not loud enough to need 24bit.

Don’t you mean “this file is not quiet enough to need 24bit”?

G

 

[castleofargh]

Spoiler: Illegal troll

Not even then! The 1812 overture has artillery in it (canons) but still 16bit is more than enough!

G

This isn’t music, a canon is not an intrument!

I fail with the order of the tags, oops.

 

[bigshot]

The cannon has orchestral backing. I actually might like 4'33" if it was a piano concerto and the piano was silent but with orchestral accompaniment. The cadenzas would be a tough slog though!

 

[SLAFL]

There may be no difference, but it doesn't make sense to test on random piece of music random people. If I familiarize myself with a piece of music (being song etc) then I check if I can hear the difference. I am not saying there is audible difference, but I mean we probably couldn't make a difference between fake Picasso and real one during 5 minutes either. Sorry for digging the old thread.

 

[gregorio]

I mean we probably couldn't make a difference between fake Picasso and real one during 5 minutes either.

If the differences between a fake Picasso and real one were all below the threshold of visibility then it wouldn’t matter if you looked at it for 5 minutes or 50 years, you could not see the difference. After 50 years (or just a few minutes) of staring though, it’s most likely your brain will make-up some imaginary differences (because it knows one is fake), although those differences will likely not be the actual differences and the judgement of which is the fake will be the exact same random chance.

G

 

[Davesrose]

If the differences between a fake Picasso and real one were all below the threshold of visibility then it wouldn’t matter if you looked at it for 5 minutes or 50 years, you could not see the difference. After 50 years (or just a few minutes) of staring though, it’s most likely your brain will make-up some imaginary differences (because it knows one is fake), although those differences will likely not be the actual differences and the judgement of which is the fake will be the exact same random chance.

G

The provenance of art has several factors. Picasso had many periods, but the uninitiated think his paintings were about abstract multi-faceted depictions of people. Such is also the same for Jackson Pollack: who I think could be easier to dispute. If you're a true historian of him, you'll see many transition periods. But his expensive artwork is the height of his drip paintings. Doesn't matter that he had some strong periods before that were drawings, variation of color with paintings, etc. There's a documentary I recommend if you can get to it: Who the f*ck is Jackson Pollack. Entertaining when it comes to a woman who got a painting from a flea market that could be an earlier Pollack. She shot herself in the foot when it came to provenance, but then also there was a forensic scientist who found pigments in the painting were the same in Pollack's studio. The film itself was kind of entertaining about how opinions varied with the art critics and friends of Pollack.

 

[bigshot]

Is a book better if the typeface used is 14 point instead of 12 point?

 

[Davesrose]

Is a book better if the typeface used is 14 point instead of 12 point?

Sorry, as a web developer, I'm wondering what base rem unit there is for that font face at browser screen

 

[castleofargh]

There may be no difference, but it doesn't make sense to test on random piece of music random people. If I familiarize myself with a piece of music (being song etc) then I check if I can hear the difference. I am not saying there is audible difference, but I mean we probably couldn't make a difference between fake Picasso and real one during 5 minutes either. Sorry for digging the old thread.

The thing is, if you have the ability to switch very fast(near instantaneously), then a majority of audible differences just "pop up" and we notice them while going back and forth. I always refer to the similar phenomenon with pictures on a computer. If you can just go back and forth between 2 almost identical pics on the same screen with a click or a scroll on your mouse, doing it a few times will be enough to just see the small differences in the picture(they’re the parts that "move". And knowing the picture or audio sample is unnecessary for that effect.

We discussed possible audio differences that could escape this trick and still have some impact we could sense, like tiny stuff over a long period of time maybe having some cumulative effect like being more tiring, or high energy ultrasounds hurting us while not strictly heard. But I ’ve seen no study of such a thing for your everyday music at reasonable listening levels.
The only legitimate concern for audio is that we absolutely need both samples to be at the same loudness(so we don’t detect loudness difference), and as it works best with short audio samples, we must first know what sample to pick that will contain the most likely to be audibly different passage.