Bad science in fiction Hall of Shame

[Green Baron]

Ok :-)

[Gargamel]

5 minutes ago, Green Baron said:

Ok :-)

And I think that post qualifies you as an audiophile snob

Edited August 31, 2018 by Gargamel

[Green Baron]

*walks away whistling*

[Superfluous J]

When I was a kid, I could tell if a TV was on in the room I was in even with no sound and no screen. Nobody believed me until we did blind tests of it. I was like 9 or so.

[Kerbart]

2 hours ago, Green Baron said:

(...) but in the case of cds number goes over quality. The cd was a dictate from the industry. Maybe a reason why vinyl records have a revival. (...)

The revival is more likely to be hipster BS in general. While I'm happy to believe that a high end vinyl setup will yield a better sound to a qualified listener than a high end CD setup, the realtiy is that 99% of the market are made of semi-deaf untrained listeners (semi-deaf because of visiting clubs and concerts) using equipment that can best be described as "the cheapest the supermarket around the corner has to offer." Add a smaller form factor, hi fidelity sound because the element hasn't worn out (surely one could replace that, but again... 99% of the market...) and impervious to mishandling (a CD can take a lot more abuse than vinyl before the damage on the medium becomes audible) and the fact is that, again for the majority of the consumers, CD offered by far better quality than vinyl.

2 hours ago, 5thHorseman said:

When I was a kid, I could tell if a TV was on in the room I was in even with no sound and no screen. Nobody believed me until we did blind tests of it. I was like 9 or so.

That high-pitched noise! I remember that.

[ARS]

Real life land mine does not goes "click" and explodes once you lift your leg off of the switch, they'll instantly blow up whether or not your leg keep pushing the plate or releasing it, and the only benefit of pressure-release trigger is delaying infantry advance by forcing them to slow down helping their poor friend and advance cautiously, which is far less useful than if the mine just blow the guy's leg off, forcing their teammates to evacuate the area, but for the sake of drama, land mines in fiction almost always explodes once the pressure is released (even if the real life counterpart does not use pressure release-trigger). Pressure release triggers are known to be used in some mine designs mainly as an anti-handling device (booby-trap) to protect a bigger mine from being disarmed. The action of lifting this sort of mine up after it's set will trigger it - which doesn't help you if you've already stepped on it, though...

[wumpus]

14 hours ago, Green Baron said:

I have quite a few vinyl disks and cds for direct comparison. Back in the day the audiophile people were complaining about quality losses of the cd in comparison to vinyl records. This is true for many recordings, but not generally. An extremely well made cd recording can be as good as an extremely well made vinyl recording, but in the case of cds number goes over quality. The cd was a dictate from the industry. Maybe a reason why vinyl records have a revival.

The best reproduction technique is imo high bandwidth digital recording direct to a master disk(*) and then a high quality cast on vinyl (direct cut or direct to disk). On a good hifi rack in a room with reasonable acoustics you can hear the difference clearly.

So, the answer is, it depends on the recording, the listener, the room and the reproduction equipment, but in general people are amazed how good vinyl sounds when they can hear the direct comparison on a reasonable rack (let's say 1000,- component price).

(*) "disk" is really a block of metal where the analogue transformed signal is carved in during recording. It's authentic :-)

The reason I specified the 1980s wasn't the quality of the recordings (although most of them were taken from analog masters, digital mastering was a new thing and you were still likely better off having an experienced analog engineer using the toolchain he was familiar with), but the output devices.  It all came down to sampling theory and the electronics of the day:

CDs are sampled at 44.1kHz.  According to Shannon and the Nyquist criterion that means that we can't reproduce any signal under 22.05 kHz.  More importantly, all frequencies >22kHz are copies of the spectrum from 0-22kHz and since roughly half the output power of music is <1kHz, that means you have a godawful screetch at 22-23kHz.  To get around this, "oversampling" is done where zeros are inserted between the samples on the disc giving a new sampling rate of at least 176kHz (at first they used "four way supersampling".  It might be higher now, but 2 is probably enough and I have no idea if your dog will thank you about using more.  Even your kids won't hear it).  Note that this doesn't do anything about the screetch at 23kHz, but at least it allows the engineers to fix the digital signal instead of the analog signal.  Modern cd players almost certainly use 1024 point FFTs to completely eliminate the screetch (that algorithm* is required to play MP3s), but that type of thing was completely out of the question in the 1980s and early 90s.  More likely the engineers of the time used notch filters (to just scrub the screetch) and IIRs (which have nasty phase issues).  I'd expect any 1980s cd player to have lousy reproduction of high frequency ranges, especially messing up the phase.  No idea if the recording engineers tried to compensate for this type of thing (which would mess up modern players).

I'd be very, very impressed if someone managed to get vinyl up to cd standards, especially if it didn't wear down below cd standards after the first listen or so.  One large error in the cd standard was that they used linear recording instead of a logarithmic encoding.  Expanding the A-law to a 16 bit linear input would have been more efficient, although I don't know how you would "spend" the efficiency: obvious choices are even greater range (although 16 bits was already >100dB S/N, pretty much beyond human hearing), expanding the sampling range (which as noted above would mostly only temporarily help, there's a reason modern systems rarely go beyond 44kHz sampling), and simply extend the amount of music on the disc (a cd can fit Beethoven's Ninth, and most labels seemed content with leaving the amount of music sold on a disc to ~40 minutes, at least during the 80s)

* ok, technically it is a DCT.  I'm pretty sure you can replace the FFT with a DCT with barely any change in your filter design (except of course the obvious change in complex coefficients).

 

PS:  Apparently not only do modern computers (at least LCD screens) make noise, you can even attempt to read the screen that way: https://arstechnica.com/information-technology/2018/08/researchers-find-way-to-spy-on-remote-screens-through-the-webcam-mic/

Edited August 31, 2018 by wumpus
added ars link.

[Green Baron]

That has been discussed over and again in the 80s and 90s. CDs lack a whole lot of harmonies and overtones. The difference can be heard in a good recording, e.g. of a symphonic orchestra. Sure, pop music mostly lacks the instruments as well as the art and rarely strains a music reproduction system to its limits as it is mostly generated by digital equipment itself and not always the best.

I can only recommend to go to a studio with decent equipment (not necessarily high end) and experience the difference if one wants to.

 

Be it as it may, i can hear my graphics card (970 clone) whistling when i switch off vsync and let it run full power. Could i write an opengl app that makes music ? :-)

Edited August 31, 2018 by Green Baron

[DAL59]

The collision in the ski speeders in the last jedi would have killed both Finn and Rose.  I'm beginning to think Rose is a spy...

[StrandedonEarth]

2 hours ago, DAL59 said:

The collision in the ski speeders in the last jedi would have killed both Finn and Rose.  I'm beginning to think Rose is a spy...

*waves hand, speaks like Obi-Wan* I have invented acceleration compensators and installed them in these rickety old jalopy ski speeders....

[Gargamel]

8 hours ago, ARS said:

Real life land mine does not goes "click" and explodes once you lift your leg off of the switch, they'll instantly blow up whether or not your leg keep pushing the plate or releasing it, and the only benefit of pressure-release trigger is delaying infantry advance by forcing them to slow down helping their poor friend and advance cautiously, which is far less useful than if the mine just blow the guy's leg off, forcing their teammates to evacuate the area, but for the sake of drama, land mines in fiction almost always explodes once the pressure is released (even if the real life counterpart does not use pressure release-trigger). Pressure release triggers are known to be used in some mine designs mainly as an anti-handling device (booby-trap) to protect a bigger mine from being disarmed. The action of lifting this sort of mine up after it's set will trigger it - which doesn't help you if you've already stepped on it, though...

Well this would completely depend upon the design.     First off, they might very well go click when depressed, but since there is usually other loud noises happening at the same time (the explosion), it might get drowned out.  If I'm a goverment contractor making mines for the military, and I can save a few bucks using cheaper switches that happen to click when depressed, then so be it. 

And yes, some land mines do wait to explode.  There are designs of anti-personel mines that have a 3-5 second fuse before detonating.  Unless the person who initially steps on the mine is running at a full sprint, they will easily still be within lethal range, and it allows trailing members of a squad to get closer.  In the case of bouncing mines, the triggering person needs to be clear of the mine so it has room to actually pop up.   The Bouncing Betty might be the most famous of this design, it had a 4 second fuse. 

 

But yeah, having the damsel in distress step on a mine and wait for the hero to defuse it is just Dramatic license.   There might be a design of mine out there that does this, and it just happens to make a clicking noise, and the ambient noise is low enough for the hero to hear it click, and then yell out "freeze!", AND the damsel has good enough reflexes to actually freeze in place, AND the Hero can diffuse it.... so i can maybe possibly slightly deal with this in movies, as it makes for good drama, as long as it's not used just for coolness sake, but the delay in the story has an overriding plot aspect to it. 

Edited August 31, 2018 by Gargamel

[kerbiloid]

CD is just a bit storage, it can't lack any part of sound if it's written with enough high resolution.

Recorders, players, and compressed file formats do.
As most of people do not feel highest and lowest frequencies as a music, you can limit your player with 10 kHz and treat everything upper as a unused noise.
As your player doesn't need >10 kHz frequencies, you don't need to store them properly in your file format, and can compress the sound to keep the file size reasonable.
As you anyway don't properly store and reproduce >10 kHz, you can loose them in a recorder.
The same with time resolution.

So, use a high-quality recorder, and you will get a CD with the same quality as the rough spiralling scratch in vinyl.

(Though, maybe CD will be not enough capable, DVD is better)

P.S.
And why all of them use demolition charges with huge red numbers?
(I would understand self-made ones, but no.)

Edited August 31, 2018 by kerbiloid

[wumpus]

4 hours ago, Green Baron said:

That has been discussed over and again in the 80s and 90s. CDs lack a whole lot of harmonies and overtones. The difference can be heard in a good recording, e.g. of a symphonic orchestra. Sure, pop music mostly lacks the instruments as well as the art and rarely strains a music reproduction system to its limits as it is mostly generated by digital equipment itself and not always the best.

I can only recommend to go to a studio with decent equipment (not necessarily high end) and experience the difference if one wants to.

 

Be it as it may, i can hear my graphics card (970 clone) whistling when i switch off vsync and let it run full power. Could i write an opengl app that makes music ? :-)

Here's a link to Shannon's mathematical proof of the Nyquist criterion : https://en.wikipedia.org/wiki/Nyquist–Shannon_sampling_theorem  You can perfectly reconstruct a sound wave from 0-22.05kHz up to 90dB S/N ratio.  Good luck finding a recording to master with that has a better S/N ratio, but children capable of hearing >22kHz might notice the difference.  Aside from that, every "harmony and overtone" in the original recording is on the CD (if the recording engineer wanted to put it there).  If the CD was mastered for "loudness" (the recording engineer's equivalent of dumping salt and sugar all over the food) you won't hear the subtle cues over the "loudness" (it will be masked much the way the stuff that gets removed in MP3s are masked), but that isn't the fault of the media.  Anything you can record and hear will fit in a cd (with the possible exception once you play over 90dB, but I'm almost positive that will still mask your hearing, and some things only young children can hear) but don't really think the same is true on vinyl.

Note: 90dB hits  "regular sustained exposure may cause permanent damage", so be careful in listening at levels where it is even possible to hear the difference.  As mentioned above this is the weak point in cd recording, they should have used logrithmic levels or encoded some sort of "maximum volume of the next few miliseconds" and recorded a ratio of that.  But even then 90dB of range is next to impossible to beat with vinyl and equally difficult to design ADCs (analog to digital converters) to put on a CD.

[Green Baron]

43 minutes ago, wumpus said:

[snip]

This has all been discussed over decades until it ended in telling myths about either vinyl and cd.

It is all true what you say, but this is not about sampling rates and s/n rations, the latter being dismal for vinyl records. It is about a steady analogue signal, that even if it went through a digital (uncompressed) recording step, sound less sterile when reproduced from an analogue machine. Not always, if it is a bad recording or just flat synthesizer music then it won't matter what plays it, but in many cases when played directly side by side, the analogue reproduction sounds better when played from a reasonable equipment, let's say 4.000 for players, amp and speakers. The turntable needs a little love when positioning and adjusting, speakers standard stereo triangle, no obstructions or corner positions. And nope, not every harmony is on the cd, that's theoretically impossible because of the quantification. You will probably not hear much difference on a 1 or 2k hifi rack, it needs a little (not much) more.

I have listened to digital reproductions (not CD, a network player with a heavenly sample rate) as well as turntable side by side on a friends high end rack, the speakers alone 15k pounds (Linn Akubarik), there i had to admit that the digital reproduction actually did sound better. But the difference was so minimal and i needed many drinks. And i personally could never even think about spending only a fraction of that money *waves hand".

I can only suggest to lay aside the calculator, do the test, side by side, by switching from CD to record player when they play the same piece of music.

:-)

Edited August 31, 2018 by Green Baron

[Gargamel]

2 hours ago, Green Baron said:

I can only suggest to lay aside the calculator, do the test, side by side, by switching from CD to record player when they play the same piece of music.

The only true way to test it, would be through a blind study.   The listener would have no idea what the device was, but could only hear the result.  Only then would an opinion be truly impartial.  Nobody will ever convince me that there isn't a strong bias in these types of arguments.  You like what you like and that's that.    But to objectively state one is better than the other, objectively, you must remove the device bias.  

[Green Baron]

Oh, we did that more than once ;-)

The outcome is not always clear to one or the other side and it only works once with a given pair.

Edited August 31, 2018 by Green Baron

[wumpus]

2 hours ago, Gargamel said:

The only true way to test it, would be through a blind study.   The listener would have no idea what the device was, but could only hear the result.  Only then would an opinion be truly impartial.  Nobody will ever convince me that there isn't a strong bias in these types of arguments.  You like what you like and that's that.    But to objectively state one is better than the other, objectively, you must remove the device bias.  

You'd want more than that.  One of the tests would be to digitally sample the vinyl and see if it can exactly reproduce whatever Green Baron prefers.  The math simply says that you can store an arbitrary waveform on the CD and that there can exist (but certainly didn't in the 1980s but should now) a player that can reproduce that waveform exactly as recorded.  It doesn't mean you will prefer it to whatever distortion happens when you press to vinyl. 

I know I've seen warnings about "studio monitor" headphones and speakers claiming that while they may produce wildly better sound, consumer recordings are mastered with such flat levels in mind.

[kerbiloid]

All that fancy symphonic muzzick can't beat the real warm sounding proven by time.

Spoiler

 

[KG3]

12 hours ago, Gargamel said:

The only true way to test it, would be through a blind study.   The listener would have no idea what the device was, but could only hear the result.  Only then would an opinion be truly impartial.  Nobody will ever convince me that there isn't a strong bias in these types of arguments.  You like what you like and that's that.    But to objectively state one is better than the other, objectively, you must remove the device bias.  

I've heard that vinyl records are able to transmit certain frequencies better than others.  Just something about how the groove is cut and how the needle works I guess.  And to compensate for these deficiencies sound engineers would boost the signal for some frequencies and lower the signal for others when they made the original masters so that I would sound right on the vinyl after they were pressed.  When CDs were made from the original masters they just didn't sound like the familiar vinyl versions.  So they had to "Digitally Remaster" to get the music to sound right on the CDs.

It sounds plausible to me.  Does anyone known if I got this right? 

[Gargamel]

4 hours ago, KG3 said:

I've heard that

The plural of anecdote is not evidence.    While this may be true, only rigorous scientific studies can prove it.   Observer's Bias is strong in this area, especially when the outcome is very subjective. 

[kerbiloid]

Why do the coal-powered airships not have a cruise nozzle from the rear side?
They have 3 lifting ones and RCS. They could fly much faster.
 

Spoiler

 

 

[Bill Phil]

20 minutes ago, Gargamel said:

The plural of anecdote is not evidence.    While this may be true, only rigorous scientific studies can prove it.   Observer's Bias is strong in this area, especially when the outcome is very subjective. 

Well, we also need to consider that there are various different levels of quality for both systems. 

[Gargamel]

7 minutes ago, Bill Phil said:

Well, we also need to consider that there are various different levels of quality for both systems. 

Can you imagine the setups required to do this study right?  It would be both an audiophile's and technophile's wet dream. 

[Bill Phil]

10 minutes ago, Gargamel said:

Can you imagine the setups required to do this study right?  It would be both an audiophile's and technophile's wet dream. 

Certainly. 

[kerbiloid]

The scratch in vinyl is just a line with rough edge, coding the amplitude and timeline in micrometers instead of bits. The needle just roughly measures this.
CD is these values written on disks.