[TOS #8,#12] From: ...no point to distributing music in 24 bit/192 kHz, From Topic ID: 93853 Options

[forart.eu]

Just a note: it has been scientifically proven that sounds are not perceived only through ears...

Anyway, in our case, bit depth is *mutch* more important than frequency IMHO.

This post has been edited by forart.eu: Mar 15 2012, 08:19

[bandpass]

Just a note: it has been scientifically proven that sounds are not perceived only through ears...

I don't think it matters—when you're sitting in front of the speakers, doing an ABX test, you can use any organ of your body you like to help make the determination (you might want to lock the door first though).

[forart.eu]

I don't think it matters—when you're sitting in front of the speakers, doing an ABX test, you can use any organ of your body you like to help make the determination (you might want to lock the door first though).

Well, in this perspective lossless is useless if you can't ABX lossy...

No, ABX isn't the unique method to evaluate sound quality, IMHO.

BTW just think about low frequency effect on your floor (then to your feets): even if your ears hear certain frequencies, the listening experience wouldn't be the same if the floor would not vibrate.

[greynol]

Ok, I'll bite. Where is your scientific "proof" that 16/44.1 is inadequate as a delivery format?

This post has been edited by greynol: Mar 15 2012, 23:15

[forart.eu]

16/44 *is* sufficient, but 32/44 (or 48) would be *mutch* better, IMHO.

[Martel]

Certain percentage of customers clearly aren't satisfied with it (statistics). Listening to a "mere 44.1/16" source (they have to be aware of it) causes subjective degradation of the enjoyment for them.

I'm not advocating anything, I just applied science (statistics) to "prove" that 44.1/16 is not enough.
(Don't take it too seriously, it's a joke)

This post has been edited by Martel: Mar 15 2012, 10:09

[2Bdecided]

No, ABX isn't the unique method to evaluate sound quality, IMHO.

Do you remember the 10th of December 2009? It was the day you joined the wrong forum.


Seriously, it's pretty trivial to ABX the feeling of the floor vibrating beneath your feet as you play a deep bass line.

The other commonly known "non ear" perception of sounds is via bone conduction (though I think it's your hears that do the final transduction). You can "hear" 50kHz if the transducer is coupled to you. Plenty of research into this. All rigorous stuff using double-blind testing.

Ultra High Frequency bone conduction is irrelevant for normal listening (unless you have your speakers a lot closer to your head than most people! ).

Cheers,
David.

[forart.eu]

Do you remember the 10th of December 2009? It was the day you joined the wrong forum.

For your knowledge, i joined this forum before you on 2nd October 2001 mate.

Seriously, it's pretty trivial to ABX the feeling of the floor vibrating beneath your feet as you play a deep bass line.

It was just an example.

So you're substantially claiming that non-ABXable lossy-encoded sound is equal to a lossless one ?

If true, than what's the meaning of lossless encoders ?


BTW - once again - my position is that we don't need more Hz, but we need more bits !



This post has been edited by forart.eu: Mar 15 2012, 14:37

[bug80]

BTW - once again - my position is that we don't need more Hz, but we need more bits !

[there's that *** picture again ]

Please read the article in the first post, especially the part titled "Sampling fallacies and misconceptions"

So you're substantially claiming that non-ABXable lossy-encoded sound is equal to a lossless one ?

If true, than what's the meaning of lossless encoders ?

You cannot proof that two samples sound equal. You can proof that they sound different to you with a high enough probability, though.

This post has been edited by bug80: Mar 15 2012, 14:43

[matt the cat]

http://www.joelstrait.com/blog/2009/10/12/..._per_sample.png

Wasnt it explained by xiphmont why thinking like this is wrong?

Edit: Got ninja'd

This post has been edited by db1989: Mar 15 2012, 20:33

Reason for edit: demoting image to link

[probedb]

Do you remember the 10th of December 2009? It was the day you joined the wrong forum.

For your knowledge, i joined this forum before you on 2nd October 2001 mate.

Really?

Group: Members
Posts: 74
Joined: 10-December 09
From: italy
Member No.: 75798

[Martel]

Now do these pictures for 16 vs. 24 bit.

[dhromed]

Photoshop won't let me create an image with 2^24 pixels to a side.

This post has been edited by dhromed: Mar 15 2012, 15:01

[Porcus]

my position is that we don't need more Hz, but we need more bits !

... than two? I'm buying that even without your ABXing it.

[skamp]

For your knowledge, i joined this forum before you on 2nd October 2001 mate.

Explain.

If true, than what's the meaning of lossless encoders ?

Lossless encoders have other applications besides playback, like transcoding.

BTW - once again - my position is that we don't need more Hz, but we need more bits !

I'm no engineer, but as I understand it: because of the way digital audio works, 1 bit adds 6dB to the scale, giving you twice the loudness. That's no matter how many bits you have, 16, 24 or 32, the value will always be the same. Increasing the number of bits will not make for a finer scale (more detail), but a bigger scale (more values, leaving original values intact). 1 bit will still be +6dB. Without dithering, 16 bits gives you 96dB to play with, 24 gives you 144, and 32 gives you a whopping 192dB. Suffice to say, that's a stupidly large scale, that would "allow" you to record sounds from the absolute threshold of hearing (0dB) to way above the threshold of pain (120-140dB if I'm not mistaken?).

Somebody correct me if I'm wrong (and if I am, please explain like you would to a 10 year old).

[SebastianG]

BTW - once again - my position is that we don't need more Hz, but we need more bits !
http://www.joelstrait.com/blog/2009/10/12/..._per_sample.png

I don't see how these pictures support your position. As Monty said already in his article: Pictures can be deceiving. You seem to lack an understanding of how and why dithering works. Go back to Monty's article and read it again.

This post has been edited by db1989: Mar 15 2012, 20:32

Reason for edit: demoting image to link

[Canar]

http://www.joelstrait.com/blog/2009/10/12/..._per_sample.png

The fact that you're posting this image proves that you did not read the article. Linear interpolation is not how the signal is reconstructed.

This post has been edited by db1989: Mar 15 2012, 20:32

Reason for edit: as above

[Porcus]

Increasing the number of bits will not make for a finer scale (more detail), but a bigger scale (more values, leaving original values intact).

Well, that is not right. You can always use the volume knob to give you a 'bigger scale' in the sense of sound pressure difference between 'full' volume and 'silence'. 1 bit can tell you 'sound on/sound off'. Two bits can give you 4 levels including silence. n bits can give you 2^n levels including silence.

But assume that while you vary the number of bits, you adjust the volume knob so that the largest signal -- all bits to '1' -- gives you the loudest useful output. Any other signal will then be some dB's less than this. Then the 'bigger scale' interpretation is the difference between the smallest *NONZERO* signal and the full volume. This smallest signal is the least-significant bit.

But that also gives you a 'fineness' (resolution): say you have an 8 bit signal, then for any of the first seven you can have *******0 or *******1. The difference between the two, is this smallest-nonzero-signal -- the 'fineness' of the scale. You can add this least-significant bit to every signal except the full volume (and by the 'loudest useful output' assumption, you do not wish to do so for full volume).

If you have so many bits that the 000[etc]0001 signal is at the hearing threshold and 111[etc]1 is the loudest useful, then what? Not only can you feed the listeners the quietest signal (s)he could detect in a dead room, you can also add this signal to anything below full volume. If we assume that there is never a 'negative masking effect' -- a hearing threshold sound is no easier to detect if another louder signal is played simultaneously, than if played in a dead chamber -- then this gives you a scale that is sufficiently fine when you have turned the volume knob up so much that it is sufficiently big. Any volume difference so small that the ADC rounds it off, is also so small that it is inaudible.



(All of this is slightly complicated by positive/negative amplitude. I've written as if everything is positive.)

This post has been edited by Porcus: Mar 15 2012, 15:50