Go Back   Cockos Incorporated Forums > REAPER Forums > REAPER General Discussion Forum

Reply
 
Thread Tools Display Modes
Old 04-27-2018, 03:49 PM   #41
Bri1
Banned
 
Join Date: Dec 2016
Location: England
Posts: 2,432
Default

Quote:
Maybe it's in my head.


Have you guys tried recording some tracks in 96k? Then rendering to 48k and see if you notice anything?
^Heh-m8 it's not in your head at all--it's just data-- you can easily render and both see + hear the differences at ~normalised~ listening levels---
your basically capturing twice the info-- how can anybody argue that's a waste?? lol...
your compressors,eq's and plugins love the fact they do not have to work so hard trying to do their jobs--- it's basically a computing bottleneck (not being optimized for super high sampling rates) that halts people in their paths..

there is a point though where human perception goes into #non physical reality# though-- it's the same as gamers reporting to actually 'see the differences' with high frames per sec over 120fps.. users swear they can-- the data will support that.simplez.
we are talking waves of differring rates and amplitudes-- radio frequencies and such have a great effect on people's perceptions-a lot of digital equiptments are channelling radio frequencies.... this is also part of the 'distortions' some may report. =)
Bri1 is offline   Reply With Quote
Old 04-27-2018, 04:00 PM   #42
karbomusic
Human being with feelings
 
karbomusic's Avatar
 
Join Date: May 2009
Posts: 29,260
Default

Quote:
Originally Posted by Judders View Post
I think we need to be careful about conflating delivery format with DSP.

Mr. Nyquist was not writing about aliasing from compression or saturation plugins (with the caveat that aliasing is generally much less of a problem than people who stare at frequency plots of 1 kHz sine tones all day instead of making music would have you believe).

I've recorded and mixed at 96 kHz for a few years now, but when it comes to bounce time, I'm going 44.1 kHz, either 16 bit WAV or 320 kbps MP3.
ABX it and that will settle it, seriously, 99% of the people who discuss this don't actually do that (even when they say they do) and... I still stand behind those guys (Nyquist etc.) until someone proves differently with something I can chew on - going on 20 years of not being able to prove anything *substantial* outside of corner cases. We can come up with scenarios all we want but if someone (not you) wants to impress, they'll start posting audio files to compare instead of words - they almost never do because it's easier to say than to prove.

Again though, sonic differences in the result, not mixing workflow or other stuff. People (again not you) just don't get it until they really start comparing in a way that removes all bias - I get it, people are scared to or think it's obvious when it isn't - been there! I also tested quite a bit on unsuspecting subjects not long ago, people with great ears, it was jaw dropping how I could make changes that were 100 times larger than we are talking here that went 100% unnoticed (will explain that on another day).
__________________
Music is what feelings sound like.

Last edited by karbomusic; 04-27-2018 at 04:07 PM.
karbomusic is offline   Reply With Quote
Old 04-27-2018, 04:26 PM   #43
Judders
Human being with feelings
 
Join Date: Aug 2014
Posts: 11,044
Default

Quote:
Originally Posted by karbomusic View Post
ABX it and that will settle it, seriously, 99% of the people who discuss this don't actually do that (even when they say they do) and... I still stand behind those guys (Nyquist etc.) until someone proves differently with something I can chew on - going on 20 years of not being able to prove anything *substantial* outside of corner cases. We can come up with scenarios all we want but if someone (not you) wants to impress, they'll start posting audio files to compare instead of words - they almost never do because it's easier to say than to prove.

Again though, sonic differences in the result, not mixing workflow or other stuff. People (again not you) just don't get it until they really start comparing in a way that removes all bias - I get it, people are scared to or think it's obvious when it isn't - been there! I also tested quite a bit on unsuspecting subjects not long ago, people with great ears, it was jaw dropping how I could make changes that were 100 times larger than we are talking here that went 100% unnoticed (will explain that on another day).
Sure, I didn't have to ABX to realise that I couldn't hear any difference between a finished song at 44.1 kHz or 96 kHz, or even PCM and MP3 in most cases.

However, extreme cases of aliasing foldback can be heard, and subtle cases can be seen, so it's not like that is some kind of placebo thing. My main line of thinking is just a pragmatic belt and braces approach: why risk these problems becoming audible when there is an easy fix to make the potential for problems disappear? Add to that the advantage of avoiding so much oversampling and latency benefits then I don't see a good case to change what I'm doing. I guess part of that is a "if it ain't broke" mentality, but I'm fine with that because (as much as I talk about this stuff on here) I don't want to have to worry about any of this when I'm recording or mixing.

I'm certainly not evangelical or think that I've got it right and everybody else has got it wrong though. Like I said, I can see pros and cons to both sides, so in the end it just comes down to personal preference. Who cares, as long as the music comes out the other end sounding good!
Judders is offline   Reply With Quote
Old 04-27-2018, 05:04 PM   #44
karbomusic
Human being with feelings
 
karbomusic's Avatar
 
Join Date: May 2009
Posts: 29,260
Default

Quote:
Originally Posted by Judders View Post

I'm certainly not evangelical or think that I've got it right and everybody else has got it wrong though. Like I said, I can see pros and cons to both sides, so in the end it just comes down to personal preference. Who cares, as long as the music comes out the other end sounding good!
I trust ya. As I eluded to Jimmy James, I'm totally pro mojo whether I actually personally heard it or not - I think forgetting about what's real and what's not is very important.

Conversely there was a defining moment a few months ago where I mixed for four hours and destroyed an already working mix badly, it was just terrible and it depressed me quite a bit at the time. But.... I rendered it out as-is, then came back a couple hours later and compared that with what it was before I started making changes...

I could not tell a difference hardly at all even with directly A/Bing in real time... it was all in my head, but while my brain was playing tricks, I was certain nothing could be more obvious, well I was wrong and my memory of that four hours when I thought it sounded dorked up still remains, even though it actually wasn't (that was god candy btw).
__________________
Music is what feelings sound like.
karbomusic is offline   Reply With Quote
Old 04-27-2018, 05:23 PM   #45
Judders
Human being with feelings
 
Join Date: Aug 2014
Posts: 11,044
Default

Quote:
Originally Posted by karbomusic View Post
I trust ya. As I eluded to Jimmy James, I'm totally pro mojo whether I actually personally heard it or not - I think forgetting about what's real and what's not is very important.

Conversely there was a defining moment a few months ago where I mixed for four hours and destroyed an already working mix badly, it was just terrible and it depressed me quite a bit at the time. But.... I rendered it out as-is, then came back a couple hours later and compared that with what it was before I started making changes...

I could not tell a difference hardly at all even with directly A/Bing in real time... it was all in my head, but while my brain was playing tricks, I was certain nothing could be more obvious, well I was wrong and my memory of that four hours when I thought it sounded dorked up still remains, even though it actually wasn't (that was god candy btw).
Hehe... I know that feeling!
Judders is offline   Reply With Quote
Old 04-27-2018, 08:41 PM   #46
pipelineaudio
Mortal
 
pipelineaudio's Avatar
 
Join Date: Jan 2006
Location: Wickenburg, Arizona
Posts: 14,047
Default

Not to wreck anybody's mojo, but even extreme amounts of oversampling, or gigaherz sampling rates, won't help with the aliasing in nonlinear processes all that much

https://www.youtube.com/watch?v=QCANuKa9ULU

Dan Lavry said as much years ago as well
pipelineaudio is offline   Reply With Quote
Old 04-27-2018, 10:22 PM   #47
mschnell
Human being with feelings
 
mschnell's Avatar
 
Join Date: Jun 2013
Location: Krefeld, Germany
Posts: 14,685
Default

Great Analyze !

Any idea why ReaComp seems to be so much worse with aliasing than the other compressor ?

-Michael
mschnell is online now   Reply With Quote
Old 04-27-2018, 11:18 PM   #48
drumphil
Human being with feelings
 
drumphil's Avatar
 
Join Date: Jun 2006
Location: Australia
Posts: 3,738
Default

Quote:
Originally Posted by Judders View Post
You could equally argue that oversampling is a workaround for lower sample rates.

I've read lots of plugin devs explaining to users that oversampling introduces its own problems and is not the magic bullet many presume it to be. If it were, some devs wouldn't spend so much effort on decramping methods that don't involve oversampling.

Another bonus of higher sample rates (well, 96k at least, as higher than that you start dealing with other potential problems) is reduced latency... if your system can handle it.
Last time I checked the answer is always filtering correctly. Then sample rate doesn't matter beyond what frequency range is necessary for the listener.

Feel free to correct me if I've missed something.


I guess this is what I'm getting at.

There is no reason that a plugin can't give the same output within the range of human hearing no matter the sample rate, so long as the sample rate covers the range of human hearing.

The only time this don't hold is if you want the plugin to react to inaudible frequencies, or output inaudible frequencies. Unless either of these two things is true, then absolutely any plugin can be written in a manner that gives the exact same audible output regardless of the project sample rate (so long as that rate is at least high enough to cover what our ears can hear).

Last edited by drumphil; 04-27-2018 at 11:50 PM.
drumphil is offline   Reply With Quote
Old 04-28-2018, 12:20 AM   #49
bezusheist
Human being with feelings
 
bezusheist's Avatar
 
Join Date: Nov 2010
Location: Mullet
Posts: 829
Default

Quote:
Originally Posted by mschnell View Post
Great Analyze !

Any idea why ReaComp seems to be so much worse with aliasing than the other compressor ?

-Michael
ReaComp uses very low quality resampling (LPF)
It’s kind of a joke. (does not work)
__________________
I like turtles
bezusheist is offline   Reply With Quote
Old 04-28-2018, 01:04 AM   #50
somebodyelseuk
Human being with feelings
 
somebodyelseuk's Avatar
 
Join Date: Jun 2011
Location: Birmingham, UK
Posts: 1,125
Default

Depends on the quality of yer AD converters, n'est ce pas?
__________________
"As long as I stay between the sun & my shadow, I guess I'm doing well."
somebodyelseuk is offline   Reply With Quote
Old 04-28-2018, 01:09 AM   #51
mschnell
Human being with feelings
 
mschnell's Avatar
 
Join Date: Jun 2013
Location: Krefeld, Germany
Posts: 14,685
Default

Quote:
Originally Posted by bezusheist View Post
ReaComp uses very low quality resampling (LPF)
In fact I don't see why a compressor should introduce any aliasing with a steady signal. Here it's supposed just to impose a constant volume control which is not prone to any aliasing or distortion in a floating point system.

-Michael
mschnell is online now   Reply With Quote
Old 04-28-2018, 01:55 AM   #52
bezusheist
Human being with feelings
 
bezusheist's Avatar
 
Join Date: Nov 2010
Location: Mullet
Posts: 829
Default

Quote:
Originally Posted by mschnell View Post
In fact I don't see why a compressor should introduce any aliasing with a steady signal. Here it's supposed just to impose a constant volume control which is not prone to any aliasing or distortion in a floating point system.

-Michael
Compression is not a simple gain change where each sample is multiplied by the same number. You would need a “hold” parameter set to a value >= one cycle of input Fc.
__________________
I like turtles
bezusheist is offline   Reply With Quote
Old 04-28-2018, 02:11 AM   #53
Bri1
Banned
 
Join Date: Dec 2016
Location: England
Posts: 2,432
Default

Question= "how many points do we need to make a circle?"

It would be great if we could have 96,000 snapshots that have a 25hz > 22khz boundry. none go under-none go over.
sampling and frame rates can be comparable-- but for certain videos just 1 frame is suffice-- for other videos we need more-- gamers report 'lag' choppiness in between frames (which is visually noticable) - so more frames equal much much smoother games..
human brain can only process so much info per sec,so the rates do have a kind of finate target.

going back to question^ : we can make a 'circle' with just 1 dot or point visually- but the resolution of dots must increase as the size of a circle increases-- it's a sizing thang... (lod,draw distance)
Bri1 is offline   Reply With Quote
Old 04-28-2018, 02:16 AM   #54
drumphil
Human being with feelings
 
drumphil's Avatar
 
Join Date: Jun 2006
Location: Australia
Posts: 3,738
Default

Those sorts of analogies don't work well for digital audio.

Quote:
It would be great if we could have 96,000 snapshots that have a 25hz > 22khz boundry
You don't need 96,000 snapshots if you don't need frequencies above 22Khz.

Watch this:



Should be mandatory watching before anyone who isn't suitable qualified already says anything about digital audio and sampling theory. Stuff that happens between sample points IS accurately represented so long as there are no frequencies above nyquist.

Now, the practical realities of filtering slightly complicates things, but the magnitude of this problem is rather oversold, especially with modern digital filters and oversampling ADC/DAC systems.

Last edited by drumphil; 04-28-2018 at 02:26 AM.
drumphil is offline   Reply With Quote
Old 04-28-2018, 02:31 AM   #55
Bri1
Banned
 
Join Date: Dec 2016
Location: England
Posts: 2,432
Default

Quote:
Those sorts of analogies don't work well for digital audio.



You don't need 96,000 snapshots if you don't need frequencies above 22Khz.

Watch this:


Should be mandatory watching before anyone who isn't suitable qualified already says anything about digital audio and sampling theory.
Oh please.... your missing the point m8.
it's a bit like midi- outdated practices just passing for a modern environment.
Bri1 is offline   Reply With Quote
Old 04-28-2018, 02:37 AM   #56
drumphil
Human being with feelings
 
drumphil's Avatar
 
Join Date: Jun 2006
Location: Australia
Posts: 3,738
Default

??

A better analogy than the number of points it takes to draw a circle is the amount of information you need to draw a circle. You plot a circle of any size perfectly with only one piece of information. Be it the circumference, radius, or diameter.

In digital audio, you can reproduce a high frequency sine wave with just a few sample points per cycle just as perfectly as a low frequency sine wave with many more sample points per cycle. It's a unique mathematical solution.

I strongly encourage you to watch the video.

You don't get a more accurate plot of audible frequencies by going to double the necessary sample rate. If you want ultrasonic information for some reason, that's a different story.

Last edited by drumphil; 04-28-2018 at 05:07 AM.
drumphil is offline   Reply With Quote
Old 04-28-2018, 02:50 AM   #57
Bri1
Banned
 
Join Date: Dec 2016
Location: England
Posts: 2,432
Default

Quote:
I strongly encourage you to watch the video..
I strongly encourage you to acknowledge a straight line can not be a curved line..
1 is trying to reapresent the other-different.

closer the buildings-the less gaps there are to fall into-- only thing is- computers put huge gusts of wind to fight as well as 1 tries to hop from skyscraper to skyscraper!

just optimize audio for optimal amount of samples for recording + playback-- and lets all move on. i think 44100 is barely scraping the barrel for *professional users.*
Bri1 is offline   Reply With Quote
Old 04-28-2018, 02:58 AM   #58
drumphil
Human being with feelings
 
drumphil's Avatar
 
Join Date: Jun 2006
Location: Australia
Posts: 3,738
Default

"stair steps don't exist in digital audio" is the other name for the video I posted. Watch it and then tell me what you think. There is a unique solution for every set of dots, that passes smoothly through all of them. The size of the gap between the dots tells you the maximum frequency you can encode. Adding more dots doesn't give you more information about the signal unless you want frequencies above what can be stored with the number of dots you started with.

Lets say you want frequencies up to 30Khz. Using a 96K or 192K sample rate will not give you a more detailed recording of those frequencies than 88K will.
drumphil is offline   Reply With Quote
Old 04-28-2018, 03:03 AM   #59
Bri1
Banned
 
Join Date: Dec 2016
Location: England
Posts: 2,432
Default

Quote:
Lets say you want frequencies up to 30Khz. Using a 96K or 192K sample rate will not give you a more detailed recording of those frequencies than 88K will.
M8- your totally missing the point! a compressor does not give a shit what your hearing--it works on info per sec- if a compressor has only 1 dot per sec to work with--- well--get the point m8.? =hardwork.
Organically a sound can appear at any moment- if a recorder is not capturing at that precise 'point' that info is lost=simplez.
Bri1 is offline   Reply With Quote
Old 04-28-2018, 03:12 AM   #60
mschnell
Human being with feelings
 
mschnell's Avatar
 
Join Date: Jun 2013
Location: Krefeld, Germany
Posts: 14,685
Default

If you know that the result is perfectly round, two points are enough to define a circle.

If you take care by hardware that the result does not feature frequencies above a certain limit, half of that limit as a sample frequency is enough to exactly define the curve. (There is no straight line at all !!! )

And this is exactly what the D/A converter hardware does with very good exactness (when a decent product is used).

And the necessary points (samples) are created by the A/D converter in exactly the correct way (when a decent product is used).

You can better see the "samples" as parameters for constructing a curve, instead of amplitude values.

Of course if some processing is done in between the A/D and the D/A converter, this is prone to any kind of distortion.

-Michael
mschnell is online now   Reply With Quote
Old 04-28-2018, 03:22 AM   #61
Bri1
Banned
 
Join Date: Dec 2016
Location: England
Posts: 2,432
Default

70khz seems a sensible setting as a new standard.
ditch the rest for hq audio.
the rest is for other types of communications... both above and below the nyquist rates...
most people not need to know what goes on both above and below at them rates other than military or medicals.
we work in a fixed range for 'audioists' =so lets optimize it! please. =)

*think* optimizing the lower octaves is essential- the ranges of 38hz>1khz>5khz upto around 18khz being most sensitive to human ears.
a lot of eq's and plugs are not optimized for lower end of octaves eh..we mainly work below 7khz eh...?

Last edited by Bri1; 04-28-2018 at 03:30 AM. Reason: 1Xtra.
Bri1 is offline   Reply With Quote
Old 04-28-2018, 03:26 AM   #62
Judders
Human being with feelings
 
Join Date: Aug 2014
Posts: 11,044
Default

Quote:
Originally Posted by drumphil View Post
Last time I checked the answer is always filtering correctly. Then sample rate doesn't matter beyond what frequency range is necessary for the listener.

Feel free to correct me if I've missed something.
As I said, that filtering is always a compromise. Using a linear phase filter will always introduce latency and pre-ringing (though not always audible).
Judders is offline   Reply With Quote
Old 04-28-2018, 05:03 AM   #63
Bri1
Banned
 
Join Date: Dec 2016
Location: England
Posts: 2,432
Default

Quote:
If you know that the result is perfectly round, two points are enough to define a circle.
^Heh...o really?
i do not know of a circle that is not 'perfectly round' otherwise your distorting a circle--right?
so,by your thinking-- i can graphically reapresent a curve with just 2 points over time? kool,show me. please.
Bri1 is offline   Reply With Quote
Old 04-28-2018, 05:14 AM   #64
mschnell
Human being with feelings
 
mschnell's Avatar
 
Join Date: Jun 2013
Location: Krefeld, Germany
Posts: 14,685
Default

It's just a matter of band limiting. It the band is limited to exactly one frequency, two points define the complete sine curve (i.e. amplitude and phase).

Similarly band limited signals can be defined by samples with a frequency of twice the upper limit frequency. (Nyquist theorem)

I suggest you try to understand the video mentioned above...
-Michael

Last edited by mschnell; 04-28-2018 at 10:43 AM.
mschnell is online now   Reply With Quote
Old 04-28-2018, 06:06 AM   #65
Judders
Human being with feelings
 
Join Date: Aug 2014
Posts: 11,044
Default

Quote:
Originally Posted by Bri1 View Post
^Heh...o really?
i do not know of a circle that is not 'perfectly round' otherwise your distorting a circle--right?
so,by your thinking-- i can graphically reapresent a curve with just 2 points over time? kool,show me. please.
Here you go:

Judders is offline   Reply With Quote
Old 04-28-2018, 06:31 AM   #66
cyrano
Human being with feelings
 
cyrano's Avatar
 
Join Date: Jun 2011
Location: Belgium
Posts: 5,246
Default

Quote:
Originally Posted by karbomusic View Post
I could not tell a difference hardly at all even with directly A/Bing in real time... it was all in my head, but while my brain was playing tricks, I was certain nothing could be more obvious, well I was wrong and my memory of that four hours when I thought it sounded dorked up still remains, even though it actually wasn't (that was god candy btw).
As they say, time heals all...

I never listen to what I record immediately after recording, if I can avoid it. Seems my ears or brain never like what's recorded when they are set to the live sound.
__________________
In a time of deceit telling the truth is a revolutionary act.
George Orwell
cyrano is offline   Reply With Quote
Old 04-28-2018, 07:29 AM   #67
karbomusic
Human being with feelings
 
karbomusic's Avatar
 
Join Date: May 2009
Posts: 29,260
Default

Quote:
Originally Posted by cyrano View Post
As they say, time heals all...

I never listen to what I record immediately after recording,
Totally agree but this wasn't what that was, it was just one of many mixing sessions for my latest project which released as of yesterday. The recording had happened months earlier.
__________________
Music is what feelings sound like.
karbomusic is offline   Reply With Quote
Old 04-28-2018, 07:31 AM   #68
karbomusic
Human being with feelings
 
karbomusic's Avatar
 
Join Date: May 2009
Posts: 29,260
Default

Quote:
Originally Posted by Judders View Post
Here you go:

Absolutely and to add to what phil said, that's the beauty of it, those few points can EXACTLY recreate the original waveform, in fact, it MUST be that way, there is no other choice if within the Nyquist range. But we only even need such few points at the very edge of the frequency range.

Someone didn't watch the video yet.
__________________
Music is what feelings sound like.
karbomusic is offline   Reply With Quote
Old 04-28-2018, 08:30 AM   #69
Judders
Human being with feelings
 
Join Date: Aug 2014
Posts: 11,044
Default

Quote:
Originally Posted by karbomusic View Post
Absolutely and to add to what phil said, that's the beauty of it, those few points can EXACTLY recreate the original waveform, in fact, it MUST be that way, there is no other choice if within the Nyquist range. But we only even need such few points at the very edge of the frequency range.

Someone didn't watch the video yet.
Though 44.1 kHz will mess up your groove if you're playing at 3,000,000,000 bpm
Judders is offline   Reply With Quote
Old 04-28-2018, 08:52 AM   #70
karbomusic
Human being with feelings
 
karbomusic's Avatar
 
Join Date: May 2009
Posts: 29,260
Default

Quote:
Originally Posted by Judders View Post
Though 44.1 kHz will mess up your groove if you're playing at 3,000,000,000 bpm
HAHA!
__________________
Music is what feelings sound like.
karbomusic is offline   Reply With Quote
Old 04-28-2018, 11:26 AM   #71
Bri1
Banned
 
Join Date: Dec 2016
Location: England
Posts: 2,432
Default

Quote:
Absolutely and to add to what phil said, that's the beauty of it, those few points can EXACTLY recreate the original waveform, in fact, it MUST be that way, there is no other choice if within the Nyquist range. But we only even need such few points at the very edge of the frequency range.

Someone didn't watch the video yet.
Heh-maybe that guy was u? because if u actually watched it shows an reaconstruction approximation via quantization which is actually prone to rounding errors . haha? i guess if it's close enough--it's good enough right? yep.nope.
so,remove any type of filtration/filter and the posted image has no 'approximated curves' and\or sinc interpolation-just points of data-holding to the next point of data=simplez.
if the current sampling system is really wonderfull,why is it using a filtration process to "faithfully reacreate" a signal? that's a distortion from this angle.. yep.

here's a question for the exberts here->> "how many samples does it take to create a 1hz signal cycle?" because @ even 44100hz--that seems a waste of samples__right,no?

a variable sampling system would be different.modern. but if people do not feel a need to change or fix what seems to work,then all is good eh..
sampling could be seen as a fairytale where the porridge is either a little too hot,too cold--or just right. who likes porridge anywayzz?- well-i do. ooo.
Bri1 is offline   Reply With Quote
Old 04-28-2018, 11:34 AM   #72
pipelineaudio
Mortal
 
pipelineaudio's Avatar
 
Join Date: Jan 2006
Location: Wickenburg, Arizona
Posts: 14,047
Default

Quote:
Originally Posted by Bri1 View Post
Question= "how many points do we need to make a circle?"
Two. You need to know two points to know everything about an actual circle. More points will not help, less points will not work.
pipelineaudio is offline   Reply With Quote
Old 04-28-2018, 11:58 AM   #73
mlprod
Human being with feelings
 
Join Date: Jul 2015
Location: Stockholm, Sweden
Posts: 1,343
Default

Like Dan Lavry said, around 60kHz would probably be about the maximum you'd need, well the next step up from that is 88,2, which is what I use nowdays. My Skylake X machine can handle that with a breeze even in big mix projects and HD space is not an issue for me.
Can you make great records in 16/44,1? Of course you can, I simply choose to not paint myself into a corner now in 2018.
__________________
Magnus Lindberg Productions - VRTKL Audio - Redmount Studios
magnuslindberg.com
mlprod is offline   Reply With Quote
Old 04-28-2018, 12:00 PM   #74
Bri1
Banned
 
Join Date: Dec 2016
Location: England
Posts: 2,432
Default

Quote:
Two. You need to know two points to know everything about an actual circle. More points will not help, less points will not work.
^hey-ok! please draw that with only straight lines..

hmm-i dunno mayngz-- i feel the current system can really be improved!
<<food thought for your theorums >>
Bri1 is offline   Reply With Quote
Old 04-28-2018, 12:22 PM   #75
Lokasenna
Human being with feelings
 
Lokasenna's Avatar
 
Join Date: Sep 2008
Location: Calgary, AB, Canada
Posts: 6,551
Default



There you go. Two points tell you everything you need to know about a circle.

Every sound behaves like a sine wave. Nyquist proved that any sine wave can be accurately captured by two points because you know that it's a sine wave.

Nobody has been able to prove otherwise (yet).
__________________
I'm no longer using Reaper or working on scripts for it. Sorry. :(
Default 5.0 Nitpicky Edition / GUI library for Lua scripts / Theory Helper / Radial Menu / Donate
Lokasenna is offline   Reply With Quote
Old 04-28-2018, 12:36 PM   #76
Bri1
Banned
 
Join Date: Dec 2016
Location: England
Posts: 2,432
Default

Quote:


There you go. Two points tell you everything you need to know about a circle.
Lol-Lokasenna-you know your cheating with the compass! ( the reaconstruction filter) -- try the same thing with a straight ruler-- define a circle like that m8
thinking u may find 1 may need an infinate amount of time and a central crossover point just to demonstrate !
begin.

try this thought for food>>



does it all matter? maybe.
cheerz.
Bri1 is offline   Reply With Quote
Old 04-28-2018, 12:48 PM   #77
Lokasenna
Human being with feelings
 
Lokasenna's Avatar
 
Join Date: Sep 2008
Location: Calgary, AB, Canada
Posts: 6,551
Default

It was a coffee cup, actually.

Why is a straight ruler required? If I know that those two points are a circle, I know exactly where every other point of the circle is.
__________________
I'm no longer using Reaper or working on scripts for it. Sorry. :(
Default 5.0 Nitpicky Edition / GUI library for Lua scripts / Theory Helper / Radial Menu / Donate
Lokasenna is offline   Reply With Quote
Old 04-28-2018, 02:55 PM   #78
mschnell
Human being with feelings
 
mschnell's Avatar
 
Join Date: Jun 2013
Location: Krefeld, Germany
Posts: 14,685
Default

Quote:
Originally Posted by Bri1 View Post
^hey-ok! please draw that with only straight lines..
Neither when mathematically defining a circle nor when doing D/A for sampling there is any straight line involved. Those exist only in your head.

Re the video:
Of course he is right that reconstruction filters are not perfect hence imposing artifacts, and a higher sampling rate needs a less perfect reconstruction filter hence imposes less audible artifacts.

But the reconstruction filter your "client" will use is not in your studio, but in his audio equipment, So you can't do much about it, and recording at a higher sampling rate will not help much .

-Michael

Last edited by mschnell; 04-28-2018 at 11:36 PM.
mschnell is online now   Reply With Quote
Old 04-28-2018, 03:35 PM   #79
Bri1
Banned
 
Join Date: Dec 2016
Location: England
Posts: 2,432
Default

ooh goodtiming! - thing is..time as we count it is actually a distortion of a greater much unknown 'fact'- as much i can gather so far...
considering this^ we still as audio users are stuck with using a timed scale-and fixed rates--- analogus signalling is infinate in possible variety and is not fixed as such-- we use fixed sampling rates to define our scalings and frequency ranges,but currently i do not see that being used as being computationally efficiently as it may well be able todo so.. we can use a variable bit system currently,but not variable sample rates???

think about it-- if indeed a newer *standard recording sampling rate* is reaconised as being: for eg: fixed rate of 60khz --- this would help with aliasing-but!- a slightly modified variable sampling system could help to rea_arrange the actual sample point positions to suit frequency modulated responses...because even using the current system,a 20hz cycle @ a sampling rate of 60khz, in theory only really would need roughly xxxxx samples to produce that tone.... right,no?

not all music is created as either fixed 20hz or 1khz sine waves-- they are very complex waves over time-- but they are not optimized for modern computing imofwiw..
analog is truer fm-itb daw is frppm (fixedratepulsedphasemodulation) there will always be distortions like that< with or without filters- but who can tell riiight..

oh and computers like todo straight-linear-on-offs --< nature (fm) does not conform to those rules luckily!! it's rather organized but extremely chaotic at the same time!!
computers just chug along...merrily.. tic,toc,clock,reach for the glock?? na call spock,for a reality shock n reastock. =)

it would be nice if cockos took a longer look at all these type of 'whacky theorums'-- because outthebox thinking can do wonders for the creative processes... newer techs are moving along--daws need to keepup! if not innovating in their own ways and methods..
try> https://en.wikipedia.org/wiki/Direct_Stream_Digital for eg-- even this'tech' is gettting old,but is moving along i guess...
Bri1 is offline   Reply With Quote
Old 05-02-2018, 04:19 AM   #80
emarsk
Human being with feelings
 
Join Date: Dec 2016
Posts: 218
Default

Bri1, no offence, but it's very apparent that you don't understand the maths that's behind signal processing.

Did you even care to watch Monty's video?

Do you understand that any signal IS a sum of sinusoidal waves? It's not just a convenient representation.

Do you understand that if you have have an analog signal, properly band-limited below the Nyquist frequency, you can sample it and convert it back to analog again, and obtain the EXACT same signal? There's no approximation, no distortion, no "steps" or "straight lines". It's not "tricking our brain". You truly have the same signal again. The only caveat is that the filters can't be ideal filters, so you need some headroom and start filtering below the Nyquist frequency. That's why we don't use 40kHz but a bit more. (And, to be fair, a typical adult human can't hear anything above 16kHz already).

Do you understand that a higher sampling rate can only extend the frequency of the processed signal, and CAN'T improve anything at the lower frequencies? This means that the ONLY thing that you gain by going from 48kHz to 96kHz is encoding frequencies that only your cat or dog can hear. (And possibly introducing intermodulation distortion when you listen on cheap hardware.)
emarsk is offline   Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT -7. The time now is 01:11 AM.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.