Main Area and Open Discussion > General Software Discussion
Why 24-bit/192kHz music files make no sense - and may be bad for you!
40hz:
^If they did, they were using an old studio trick from back in the early 60s. The story (as I heard it) was that a mix wasn't finalized until it sounded good played both through a car radio and a small hand-held transistor radio. (Not much was stereo back then.)
Since it was estimated that 90% of the USA's music listening was done either in a car, or with a small radio, it made good sense to optimize the mix for the lowest common denominator - with the sure knowledge that if it sounded good there, it would sound even better on a home stereo or the family Victorola.
Edvard:
Rumor had it that was a trick used extensively by Motown Records and picked up on by others. Part of the rumor is they had their own low-power radio station that would broadcast to the cars in the parking lot where musicians and producers could listen to the final mix over the car radio.
Renegade:
I think you're really just making an appeal to ignorance and elevating the value of the theoretical here, which could be the beginning of science perhaps (if it inspires investigation), but is really just speculation.
-JavaJones (September 24, 2013, 01:03 PM)
--- End quote ---
No. I am not speculating. I am stating facts and that the implications of those have not been applied to this particular topic.
Perhaps I should have been stronger in my statements and put out more evidence. I was trying to be relatively brief. I'll get to more evidence below.
It is in fact fairly easy to test the limits of what our sound reproduction equipment can produce, and that is ultimately all that actually matters in this consideration because in the end all the recording, mixing, and mastering has to get squeezed through those limited speakers/headphones on the listener's end.
-JavaJones (September 24, 2013, 01:03 PM)
--- End quote ---
No argument there. Each step in the process creates a bottle neck, with the final one being the playback equipment.
But even if you somehow believe the measuring capabilities we have now can't account for every possible effect, as I said above there is really a simple way to find out if any of those "woo-woo" audio stuff is *practically detectable by humans* (whether directly or otherwise!), and yet so far such tests have failed to show a difference even between existing high quality (but lossy) audio formats and their lossless sources, much less a difference between two ultra high quality lossless sources.
-JavaJones (September 24, 2013, 01:03 PM)
--- End quote ---
It should be no surprise that the test don't show a difference because they're not actually testing for the right differences!
Again, they have unrealistically limited the scope of the question to the range of human hearing, which exactly what I am disputing. They are NOT testing the right things.
As I mentioned above, we still need more research done on what those "right things" are. We KNOW for a fact that they exist. This is not indispute, except in the audio industry, ironically. (Evidence below.)
That being said I will say that to my knowledge no one has done such a blind test with 16/44.1 vs. 24/192 audio, so if indeed these inaudible frequencies are somehow reproduced by audio equipment, even though they're well outside their rated range, and if somehow humans are able to detect them, then there may be value in Pono and other ultra high quality audio storage approaches.
-JavaJones (September 24, 2013, 01:03 PM)
--- End quote ---
While I wouldn't phrase things quite like that, you've kind of summarized what I've been saying above.
But I think the problem I have with your argument is that it essentially relies on the supposed limitations in our knowledge of audio science, when in fact, as I've pointed out, we don't need to know everything about audio to test *the effects* (to *understand* the effects we perhaps do, but not to *test whether they exist*).
-JavaJones (September 24, 2013, 01:03 PM)
--- End quote ---
But they aren't even testing for the effects! They run some tests for human hearing and do some analysis on the signals, but they do not test for what I've been describing.
I don't think we need to wait until some possible future breakthrough in audio science to determine whether Pono is worthwhile.
-JavaJones (September 24, 2013, 01:03 PM)
--- End quote ---
The question is whether the costs outweigh the potential benefits. Maybe, maybe not. My gut reaction is that storage is so cheap now that unless we're talking about orders of magnitude differences, then probably the costs are insignificant. Will that pan out? Dunno. I guess it's just a gamble, and a gamble that I'd take.
This is like someone saying "Homeopathic medicine works but our existing science has no way to measure it", to which I say do some controlled studies and we'll soon see. We can measure effects even if we cannot directly measure methods of action.
-JavaJones (September 24, 2013, 01:03 PM)
--- End quote ---
Yes - as long as we actually do some better testing, and explore more about what we should be looking for. We know some of what we should be looking for, but it's a BIG ocean out there to discover.
So who wants to run a blind test with Pono? I can guarantee you Neil Young won't be doing any fair comparisons (i.e. blind, same audio source, multiple subjects) any time soon. :D
-JavaJones (September 24, 2013, 01:03 PM)
--- End quote ---
Funny that you should say blind... hehehe! Let's get into the evidence, starting with...
...wait for it...
...wait for it...
GHOSTS!
Yes. Ghosts. Those spooky things that haunt houses, and sometimes... even laboratories. ;)
http://skepdic.com/infrasound.html
Several years earlier, Tandy was working late in the "haunted" Warwick laboratory when he saw a gray thing coming for him. "I felt the hairs rise on the back of my neck," he said. "It seemed to be between me and the door, so the only thing I could do was turn and face it."* But the thing disappeared. However, it reappeared in a different form the next day when Tandy was doing some work on his fencing foil. "The handle was clamped in a vice on a workbench, yet the blade started vibrating like mad," he said. He wondered why the blade vibrated in one part of room but not in another. The explanation, he discovered, was that infrasound was coming from an extractor fan. "When we finally switched it off, it was as if a huge weight was lifted," he said. "It makes me think that one of the applications of this ongoing research could be a link between infrasound and sick-building syndrome." When he measured the infrasound in the laboratory, the showing was 18.98 hertz--the exact frequency at which a human eyeball starts resonating. The sound waves made his eyeballs resonate and produced an optical illusion: He saw a figure that didn't exist.*
--- End quote ---
Perhaps the tests shouldn't be "blind". ;) ;D
More on the story here:
http://ghosts.monstrous.com/infrasound.htm
Have any of these compression or audio tests tested for ghosts? Because sound can produce ghosts!
Just how damn cool is that? :Thmbsup:
While it might not be great for children's music, can you imagine some band like Slayer creating a song about ghosts/monsters/whatever that made you hallucinate? You could take drug music to entirely new highs! ;D
There's lots more evidence out there to illustrate what I've been saying above.
Another fun topic is sonic weaponry. I'll leave that out for now as it should suffice to say that sonic weaponry can cause serious effects in people.
The practical side of the question is whether consumer level gear will ever reach the point that it can reproduce those kinds of sounds. That's where the mass market is. (I'm assuming that we'd have professional/military level gear capable of that before consumer level gear.)
JavaJones:
Hmm, how do you know the tests aren't testing for the right things? They are *not* as specific as you are suggesting. Here's how a blind audio test works: a person listens to playback of 2 (or 3) audio segments with identical *content*, but that differ in compression/bitrate/storage media/etc. They are not told which is which, but they know they are listening for a difference (sometimes they listen to 2, then a 3rd, and are supposed to identify which of the first 2 the 3rd corresponds to). If they can reliably detect a difference (or match the 3rd sample), they could correlate that difference with e.g. lossless formats vs. lossy compression. Multiple tests confirm that people are unable to make such distinctions with high enough bitrates in lossy compression (vs. CD audio 16/44.1 as a comparative). It doesn't matter one teeny tiny bit if the way they were able to detect a difference was because of "subsonic" or subconscious frequencies; they are not measuring the specific method of differentiation, only *whether there is any reliable differentiation*. There is not. Therefore the idea that they're not "measuring the right thing" is incorrect. They are measuring the *effect*, not a specific and limited set of criteria.
As I said I'm not aware of any such tests being yet performed on 24/192 audio, but since people are almost universally unable to detect a difference even between lossy and lossless 44.1 audio, I'm doubtful that the results would be any different. The only possible way they would is if you're right about the subconscious frequencies, which is highly speculative since speakers aren't built to reproduce such frequencies, and are broadly incapable of producing them even when intentionally induced to do so. The crux of my argument is focused on the limitations of audio reproduction equipment, not on whether such effects actually exist in the real world, with live sound (they obviously do). Still, I'd be curious to see the results of such a test, if only to answer your doubts.
- Oshyan
Renegade:
It's been about 10 years or so since I was doing a lot of this, so my memory may be rusty in some places.
IIRC, testing is done on reference speakers or monitors. Those usually have a frequency response of 20 Hz to 40 Hz up to 50 KHz, with most maxing out below 30 KHz.
None of those speakers is capable of reproducing the kinds of sounds that I've mentioned above.
At 50 KHz, sampling beyond 100 KHz is irrelevant because of the Nyquist frequency/sampling theorem there. So, what was mentioned above about "air" at the extreme high end is filtered out.
However, at the low end below 40 or 20 Hz, that's cut. The examples I gave were in the cut range there, so it's safe to say that no, they have not tested for that. But we would also need to include those kinds of effects in the sound being tested in order to test for it. That's not going to happen in your average, every day music.
So practically, it's pretty safe to say that 24/192 is useless, but ONLY because we are not using what is available to its full potential.
That's almost like complaining about being hungry after finishing a nice steak meal because you only ate the gravy/sauce and didn't touch the steak or anything else.
You can imagine a musician creating a song and working in sound that makes you nauseous at certain points. We don't have that now, but do we want to preclude the possibility of that? That's the question that we need to address when talking about ranges here. By excluding those possibilities, we limit artistic expression right from the get-go.
Will I go out and buy a set of speakers with a frequency response range of 0.3 Hz to 96 KHz? Hell no. Right now that's some serious engineering and custom work that I can't afford. But, that doesn't mean that in the future we won't see that. We already have consumer level speakers and headphones rated for up to and beyond 30 KHz. Some go down to 12 hz or so.
Traditionally, hardware leads the technology race. However, this is one case where software is clearly leading, which is a bit of a bizarre flip-around.
Does it make sense to have a software standard ready for hardware manufacturers to catch up to? I don't see why not.
Nixing 24/192 may be premature if the attitude is, "Oh well, I can't buy speakers for it today, so, forget it." That's not really a good reason. If we stuck to that kind of mentality, we'd still be saying, "Oh, forget that stupid wheel invention because there are no Ferraris to put them on."
BACK TO THE XIPH ARTICLE
What I've outlined is an illustration that the conclusions in the Xiph article are premature and based on incomplete science.
For the IMMEDIATE PRACTICALITY TODAY IN THIS MOMENT... a 96 KHz sampling rate is about the highest ever needed, and more realistically for existing equipment, 48 KHz is more than enough. Bit depth decides what happens within that, and I've not tried to address that question as I'm still rusty on some of this and would need to go back and read up for a refresher before I could comment as I have above.
Another way to phrase this is "what is practical today" vs. "what will be practical tomorrow". Or something like that.
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version