Author Topic: Can you truly make 16bit audio sound better by upping it to 32bit?  (Read 16132 times)

0 Members and 1 Guest are viewing this topic.

Offline peteb2Topic starter

  • Regular Contributor
  • *
  • Posts: 242
  • Country: nz
...and call it HiFi/HiQuality Audio.... Sorry. I am missing something here. This latest Kickstarter looks like a nifty wee cased thing to plug your headphones into but really?

I followed a bit of the excitement for a while over high bitrate compressed audio files (.flac etc) thing. We mere individuals can now supposedly access what is as close if not a direct copy of the original studio mix by the recording artist. Some are released with the assurance you hear what the original mix which will always sound better being the studio recording, not some now-considered degraded version 16bit 44.1k music CD sold by the music industry that they adopted when vinyl records and cassette tapes went away... All that's happened i feel is it's a way to pay way more for your fave music you might like to own... but i digress.

So add to all this that (apparently) the audiophools who'll will point out that any .mp3 audio files are an abortion because much of the music's 'soul' will be missing and even often distorted and that devices like your average iPhone do not contain a DAC chip that can operate at a bitrate anywhere near being able to decode 92bit/192k HiRes files let alone a 'HiFi" quality file format such as .flac  They will shun using Bluetooth to remote feed the file pointing out a high number of bits will be being lost in the process, so they therefore own an expensive dedicated played from the likes of Sony, FiiO or Astell & Kern for their portable needs.... To some extent they have a point but then it's all about what the individual cares for i guess

Enter this device: It appears to convert the low quality bitrate of Bluetooth up to 32bit.... so that makes it SOUND
better.... Hmmmmm ...

I'd love to know anyone's thoughts on the subject but in this Kickstarter unit i'm seeing an audio up-scaler that's not HiFi let alone being true HiRes even though they are using a device capable for the doing so... bottom line you can't replace actual lost bits (only approximate them).... SO WHY BOTHER IN THIS APPLICATION?

So cut to gratuitous Rigol scope and dodgy low speaker playing some sound video:

https://ksr-video.imgix.net/assets/017/835/487/a4cca3311c291557106869067e1963e6_h264_high.mp4

The Kickstarter Page:...

AQUA+

http://tinyurl.com/y8wnnen4


 

Offline DrGeoff

  • Frequent Contributor
  • **
  • Posts: 793
  • Country: au
    • AXT Systems
Re: Can you truly make 16bit audio sound better by upping it to 32bit?
« Reply #1 on: August 10, 2017, 10:39:07 pm »

Enter this device: It appears to convert the low quality bitrate of Bluetooth up to 32bit.... so that makes it SOUND
better.... Hmmmmm ...

I'd love to know anyone's thoughts on the subject but in this Kickstarter unit i'm seeing an audio up-scaler that's not HiFi let alone being true HiRes even though they are using a device capable for the doing so... bottom line you can't replace actual lost bits (only approximate them).... SO WHY BOTHER IN THIS APPLICATION?


Bit rate and bit depth are two different things.
I'd recommend watching Monty Montgomery's excellent video on this to further understand how the two affect signals. In short, 44.1K 16 bit is an excellent choice, and there is nothing wrong with it.





Was it really supposed to do that?
 
The following users thanked this post: alexanderbrevig, peteb2

Offline peteb2Topic starter

  • Regular Contributor
  • *
  • Posts: 242
  • Country: nz
Re: Can you truly make 16bit audio sound better by upping it to 32bit?
« Reply #2 on: August 11, 2017, 12:50:32 am »





Enter this device: It appears to convert the low quality bitrate of Bluetooth up to 32bit.... so that makes it SOUND
better.... Hmmmmm ...

I'd love to know anyone's thoughts on the subject but in this Kickstarter unit i'm seeing an audio up-scaler that's not HiFi let alone being true HiRes even though they are using a device capable for the doing so... bottom line you can't replace actual lost bits (only approximate them).... SO WHY BOTHER IN THIS APPLICATION?


Bit rate and bit depth are two different things.
I'd recommend watching Monty Montgomery's excellent video on this to further understand how the two affect signals. In short, 44.1K 16 bit is an excellent choice, and there is nothing wrong with it.

For most generic pop type music i guess but along comes an audiophool who once handed me their Neil Young 'HiRes Pono' player that has a special comparison file set and i must say i actually noticed what i describe as an obvious clarity difference for the higher bit recording of the same song. Whether it actually mattered i guess not and i must agree the standard for music CD is good enough but there are those who seek perfection and will go for the 'extra bits' i guess...


 
The following users thanked this post: jiro

Offline WastelandTek

  • Frequent Contributor
  • **
  • Posts: 609
  • Country: 00
Re: Can you truly make 16bit audio sound better by upping it to 32bit?
« Reply #3 on: August 11, 2017, 01:10:18 am »
that was...awesome

that gentleman is a really good educator
I'm new here, but I tend to be pretty gregarious, so if I'm out of my lane please call me out.
 

Offline Richard Crowley

  • Super Contributor
  • ***
  • Posts: 4317
  • Country: us
  • KJ7YLK
Re: Can you truly make 16bit audio sound better by upping it to 32bit?
« Reply #4 on: August 11, 2017, 01:47:15 am »
Can you truly make 16bit audio sound better by upping it to 32bit and call it HiFi/HiQuality Audio?
Can you make the MonaLisa look better by putting it into a larger frame?  No.

Quote
Sorry. I am missing something here. This latest Kickstarter looks like a nifty wee cased thing to plug your headphones into but really?
No, you aren't missing anything. Kickstarter and Indiegogo are a rich source of phoolishness like this.  To give them the benefit of the doubt, perhaps the majority of these schemes are perpetrated by people who actually believe the hype themselves. It doesn't change the fact that it is phony.
 
The following users thanked this post: Jacon, Electro Detective

Offline b_force

  • Super Contributor
  • ***
  • Posts: 1381
  • Country: 00
    • One World Concepts
Re: Can you truly make 16bit audio sound better by upping it to 32bit?
« Reply #5 on: August 11, 2017, 07:05:00 am »
I can think of some applications where 16 bit isn't adequate enough. But those are all high professional situations, where noise is a serious issue (especially with highly sensitive horn drivers).
Keep in mind that most DACs are being sold as 16 or 24 bit, but practically they only can deliver 70-80% of that (speaking from a noise/dynamic range kind of perspective)

Offline Rerouter

  • Super Contributor
  • ***
  • Posts: 4694
  • Country: au
  • Question Everything... Except This Statement
Re: Can you truly make 16bit audio sound better by upping it to 32bit?
« Reply #6 on: August 11, 2017, 12:04:23 pm »
That was a very informative video, thanks :)

b_force, i would say as far as 24 bit dacs go, its really easy to end up with them having effective bits in the 14-18 ballpark, once you add up all the error terms, to actually get 24 bits effective takes not only some more $, but layout techniques that go against most audiophile mantras (e.g. ground planes, return routing, and serious crosstalk prevention) I'm approaching this purely from the perspective of having laid out 16 bit ADC/DAC for precision measurements, and that alone wasn't easy to guarantee less than 1 lsb of noise. (50uV on 3.3V)
 

Offline coppice

  • Super Contributor
  • ***
  • Posts: 8574
  • Country: gb
Re: Can you truly make 16bit audio sound better by upping it to 32bit?
« Reply #7 on: August 11, 2017, 12:18:39 pm »
that was...awesome

that gentleman is a really good educator
He has several videos related to common misunderstandings about audio. They are all worth watching, even if you are an expert, if only to see his slant on presenting the material.
 

Online brucehoult

  • Super Contributor
  • ***
  • Posts: 3972
  • Country: nz
Re: Can you truly make 16bit audio sound better by upping it to 32bit?
« Reply #8 on: August 11, 2017, 12:56:04 pm »
"Sound better" is a very subjective thing. Some people think pumping up the bass and/or treble makes music sound better. Or adding a bit of reverb or phasing.

You certainly can't add back any missing information by converting a 16 bit signal to 32 bit and/or increasing the sampling rate.

What you *can* do is make life easier for the analogue output components, to get the best from what you already have.

Going to 16 bit 44.1 kHz with no copy protection on CDs was perhaps the worst blunder the music industry ever made. It's TOO GOOD. Or at least, it's as good as anyone ever needs for playback. In a studio mixing down different tracks and doing various processing it's useful to have a bit better, so that the final output can be at the maximum quality possible for 16 bit 44.1 kHz. Which is very very good.

But not for end user delivery and playback.
 

Offline b_force

  • Super Contributor
  • ***
  • Posts: 1381
  • Country: 00
    • One World Concepts
Re: Can you truly make 16bit audio sound better by upping it to 32bit?
« Reply #9 on: August 11, 2017, 04:22:19 pm »
"Sound better" is just pure and only marketing b*llsh*t.
That virus is even so bad that even crap is coming out of some very smart people unfortunately.

Noise levels etc, don't have anything to do with 'sounding better'.
From experience I can tell that in some professional situation noise can be very annoying and unwanted.
Like quiet moment on stage with a speaker monitor for the musician for example.

I have been developing quite some audio related projects over the years, and I can confirm that it's all about grounding and using the right components, like Rerouter already said.
Unfortunately with DSPs (digital filtering) noise levels are not getting any better, especially when you also need to have an universal input stage for different input standards.

Offline Richard Crowley

  • Super Contributor
  • ***
  • Posts: 4317
  • Country: us
  • KJ7YLK
Re: Can you truly make 16bit audio sound better by upping it to 32bit?
« Reply #10 on: August 11, 2017, 04:34:56 pm »
I have been recording audio professionally from several decades before the Digital Age. The goal while "tracking" (recording the original sounds) is to maintain the optimal SNR (Signal to Noise Ratio) by keeping the average audio levels as high as possible without hitting the upper limit (clipping).  Clipping back in the analog era was much "softer" and more forgiving/flexible compared to digital recording where when you hit 0dBFS (Full Scale) you are up against a HARD limit, beyond which there is nothing but severe distortion.

When you do the math, you will see that a dynamic range of 16 bits is the equivalent of 96.33dB.  If you take the noise-floor of the Real World at around 20dB you will see that 96dB is quite more than enough to produce sounds over the complete range that our ears can handle.  Unless you want to ruin your ears with head-banging heavy metal at 120dB, I'll leave you to decide that for yourself.

The overwhelming majority of music consumed these days is typically both dynamic range compressed (see the "Loudness Wars") as well as data-compressed (i.e. MP3) and it is rare to find people listening to music that even amounts to 14-15 bits of dynamic range.

OTOH, when I am recording live performances (especially of amateur musicians) I prefer to use 24 bit recording sample depth. Not because it offers any significant improvement of SNR, but because it offers a significant increase in headroom so that I can record at a comfortable level while maintaining a substantial safety margin to handle unexpected amplitude excursions.  Of course, it is also true that the electronics (mostly the microphone preamp) has nowhere near the 144dB of dynamic range implied by a 24-bit sample depth. 

32 bits of sample depth implies 196dB of dynamic range which isn't even possible to achieve under normal atmospheric pressure on our planet.  Not to mention that you would need cryogenically-cooled electronics to even get half that dynamic range.  But 32-bit (or even floating-point) arithmetic is commonly used for intermediate calculations in most modern audio software to handle mixing and other effects without worrying about arithmetic overflow ("clipping").  Of course the final result is truncated back down to 16 bit for consumption in the Real World.

Ref:
https://en.wikipedia.org/wiki/Audio_bit_depth
https://en.wikipedia.org/wiki/Loudness_war
« Last Edit: August 11, 2017, 04:47:38 pm by Richard Crowley »
 
The following users thanked this post: janoc, thm_w, OldNeurons, alm, juan_ee

Offline ejeffrey

  • Super Contributor
  • ***
  • Posts: 3669
  • Country: us
Re: Can you truly make 16bit audio sound better by upping it to 32bit?
« Reply #11 on: August 11, 2017, 08:25:30 pm »


When you do the math, you will see that a dynamic range of 16 bits is the equivalent of 96.33dB.  If you take the noise-floor of the Real World at around 20dB you will see that 96dB is quite more than enough to produce sounds over the complete range that our ears can handle.  Unless you want to ruin your ears with head-banging heavy metal at 120dB, I'll leave you to decide that for yourself.

Also, if you _do_ decide that, 20 dBA won't be your personal noise floor for long, so 16 bit will still suffice :)

But ++ to all of this.  24 bit is useful in the sense that it is affordable overkill that takes one piece of the signal chain (the ADC) off the table as something you have to get "just right" when recording.  For playback, a signal with more than 16 bits of dynamic range is unlistenable.

The same thing goes for sample rate.  44.1 kHz or 48 kHz is enough to capture everything you can hear.  But going to higher sample rate makes digital filters easier to implement.  It would also help with analog anti-aliasing filtering except that now basically all audio CODECs are sigma-delta based which upconvert internally to a higher effective sample rate so that simple anti-aliasing filters can be used.
 

Offline b_force

  • Super Contributor
  • ***
  • Posts: 1381
  • Country: 00
    • One World Concepts
Re: Can you truly make 16bit audio sound better by upping it to 32bit?
« Reply #12 on: August 12, 2017, 01:36:19 pm »
Like said before, a 24 bit ADC (or DAC) is still no guarantee at all.
In fact, there are enough ADC/DAC/CODECs that can only do around 95-100dB SNR/Dynamic range *

Can tell you that it's not enough, or very difficult to handle with in professional situations.

@Richard Crowley
I can imagine that these extra bits are just enough to not feeling limited.
Not that you'll always use these, but it's just handy to have the extra space.
For that reason a lot of video editors use 4K.
A normal human being isn't able to see the difference either (with 1080p or 720p).
(same marketing BS)

@ejeffrey
Lower sample rates have more other benefits, like (practically) lower latencies, smaller in storage etc etc
Ones again, there is no 'better', just use what is needed for your purposes.


(* Dynamic range and SNR aren't the same thing. Bit unfortunately in every datasheet these two numbers are just being copied)

Offline WastelandTek

  • Frequent Contributor
  • **
  • Posts: 609
  • Country: 00
Re: Can you truly make 16bit audio sound better by upping it to 32bit?
« Reply #13 on: August 12, 2017, 04:23:21 pm »

For that reason a lot of video editors use 4K.
A normal human being isn't able to see the difference either (with 1080p or 720p).
(same marketing BS)


well, for watching TV perhaps

Gamers are using the extra real estate to achieve a wider FOV, I am already seeing attempts at 3 x 4K surround setups among the serious (and wealthy) hardware enthusiasts.

There will never be a limit to the gamer appetite for pixels, framerate, input lag or anti-aliasing.
I'm new here, but I tend to be pretty gregarious, so if I'm out of my lane please call me out.
 

Offline b_force

  • Super Contributor
  • ***
  • Posts: 1381
  • Country: 00
    • One World Concepts
Re: Can you truly make 16bit audio sound better by upping it to 32bit?
« Reply #14 on: August 12, 2017, 07:00:33 pm »

For that reason a lot of video editors use 4K.
A normal human being isn't able to see the difference either (with 1080p or 720p).
(same marketing BS)


well, for watching TV perhaps

Gamers are using the extra real estate to achieve a wider FOV, I am already seeing attempts at 3 x 4K surround setups among the serious (and wealthy) hardware enthusiasts.

There will never be a limit to the gamer appetite for pixels, framerate, input lag or anti-aliasing.
Gamophiles  ;) ;D

Offline WastelandTek

  • Frequent Contributor
  • **
  • Posts: 609
  • Country: 00
Re: Can you truly make 16bit audio sound better by upping it to 32bit?
« Reply #15 on: August 12, 2017, 07:11:57 pm »

Gamophiles  ;) ;D

Well sure, lol, but at least we are reaping actual benefits, unlike phools.   I have been running 5760x1080 here for like 6 years already, I would never go back to a tighter FOV.
I'm new here, but I tend to be pretty gregarious, so if I'm out of my lane please call me out.
 

Offline MT

  • Super Contributor
  • ***
  • Posts: 1616
  • Country: aq
Re: Can you truly make 16bit audio sound better by upping it to 32bit?
« Reply #16 on: August 12, 2017, 08:26:37 pm »
Can you truly make 16bit audio sound better by upping it to 32bit and call it HiFi/HiQuality Audio?
Can you make the MonaLisa look better by putting it into a larger frame?  No.
But inside a golden baroque frame makes her appear more impressive!

 

Offline nidlaX

  • Frequent Contributor
  • **
  • Posts: 663
  • Country: us
Re: Can you truly make 16bit audio sound better by upping it to 32bit?
« Reply #17 on: August 13, 2017, 03:08:42 am »
BTW, for small but close to eye applications, such as cellphones, maybe 2k and 4k don't make much a difference, and to some people, maybe 1080p and 2k don't make a difference, but 720p and 1080p are definitely different, very different.
I can still make out sub-pixels in Pentile configuration 1440p OLEDs at 5.5", so there is still some room for (not very practical) improvement. Not to mention, VR applications will require greater than 4K densities in small screens.
 

Offline IanMacdonald

  • Frequent Contributor
  • **
  • Posts: 943
  • Country: gb
    • IWR Consultancy
Re: Can you truly make 16bit audio sound better by upping it to 32bit?
« Reply #18 on: August 13, 2017, 11:05:45 am »
The main quality issue is not the bit depth but the filtering used to eliminate RF. If the digital clock is 44.1kHz then that is only just over twice the audio range. That means your low-pass filter needs to be able to provide a substantial amount of high-cut in just one octave, to eliminate remnants of the clock frequency. Typically you need a compound active filter to achieve this, and the filter design itself can make or break the sound quality. In particular a poor filter design will introduce phase changes in the passband.

The solution is oversampling, which is basically a case of multiplying the digital clock so the filtering requirement is shifted further away from the audio range. That means a simple low-pass filter can be used, which is far less likely to introduce artifacts into the audio range. Even some of the expensive early CD players lacked oversampling, and that is why the sound was 'fuzzy' and unclear.  Some manufacturers then started making it a point of 'specmanship' to use silly amounts of oversampling. In fact this gains no advantage. As long as the carrier is high enough in frequency to avoid the need for a filter with extremely sharp cutoff, that's all that's needed.

Incidentally, a 44.1kHz sampling rate means that a 10kHz signal consists of four points per cycle*. A 1kHz signal, 40. 100Hz bass, 400. Since 16-bit sample depth gives 65536 levels (or 32768 in each signal polarity)  that is far more resolution than you can actually use, when you only have 400 points per cycle.  Hence 32-bit encoding is pointless UNLESS a much higher clock rate is also used. 

* The human ear cannot tell the difference between sine, triangle and square waves at above a few kHz. So, this is actually quite acceptable. Believe it or not.  :-//
« Last Edit: August 13, 2017, 11:09:43 am by IanMacdonald »
 

Online brucehoult

  • Super Contributor
  • ***
  • Posts: 3972
  • Country: nz
Re: Can you truly make 16bit audio sound better by upping it to 32bit?
« Reply #19 on: August 13, 2017, 12:08:05 pm »
* The human ear cannot tell the difference between sine, triangle and square waves at above a few kHz. So, this is actually quite acceptable. Believe it or not.  :-//

Given that both triangle and square waveforms consist of the fundamental tone plus odd harmonics (in different ratios), the first overtone in both cases is three times the frequency of the fundamental tone. So it stands to reason that even a fundamental of 5 or 6 kHz puts the first overtone near or past (depending on age) the limit of hearing.
 

Offline b_force

  • Super Contributor
  • ***
  • Posts: 1381
  • Country: 00
    • One World Concepts
Re: Can you truly make 16bit audio sound better by upping it to 32bit?
« Reply #20 on: August 13, 2017, 12:15:26 pm »
That's simply because our ears work as a low-pass filter as well.
Most (older) adults don't hear anything above 10kHz relativly speaking with other frequencies.
(with a single tone and enough dB's you will eventually able to hear it).


Offline kalel

  • Frequent Contributor
  • **
  • Posts: 880
  • Country: 00
Re: Can you truly make 16bit audio sound better by upping it to 32bit?
« Reply #21 on: August 13, 2017, 12:25:28 pm »
* The human ear cannot tell the difference between sine, triangle and square waves at above a few kHz. So, this is actually quite acceptable. Believe it or not.  :-//

Given that both triangle and square waveforms consist of the fundamental tone plus odd harmonics (in different ratios), the first overtone in both cases is three times the frequency of the fundamental tone. So it stands to reason that even a fundamental of 5 or 6 kHz puts the first overtone near or past (depending on age) the limit of hearing.

At 6 kHz in Audacity (tone generator), I can hear the difference between Sine, Square and Sawtooth. I believe there's surely some other reason for that though. I know that audio is complicated as so many factors influence it. Who knows what frequencies I'm actually hearing.

At 10k, there's a huge difference too. Square makes you hear "more simultaneous tones" than Sine. 10k Sine at decent loudness hurts my ears, so be careful with trying any of these.

It's interesting how the "Sine" looks when zoomed in:



Square (also 10000 Hz)



It would be even more interesting to compare it with what comes out of the headphones, but that's not easy.
« Last Edit: August 13, 2017, 12:32:50 pm by kalel »
 

Offline b_force

  • Super Contributor
  • ***
  • Posts: 1381
  • Country: 00
    • One World Concepts
Re: Can you truly make 16bit audio sound better by upping it to 32bit?
« Reply #22 on: August 13, 2017, 02:27:22 pm »
there is nothing more 'complicated' about audio.
Just a lot of fairy tales.

With headphones you simply make the SNR MUCH better, so all effects or differences are going to be more clear.

With a sawtooth contains even and odd harmonics. So with 6kHz it is possible to just hear the 2nd harmonic and maybe the third (although difficult).

A square wave and triangle wave only have odd order, so with 6kHz the next harmonic is on 18kHz.
Probably there are some issues with sampling or other side effects.
All depends how good/bad programmers have done their home work.
In a lot of cases there are also resample issues in Windows, which can be very clearly heard with just sine waves and alike.
With my standard onboard soundcard 44kHz music sounds horrible when I don't select the proper output samplerate for example.

Offline CatalinaWOW

  • Super Contributor
  • ***
  • Posts: 5160
  • Country: us
Re: Can you truly make 16bit audio sound better by upping it to 32bit?
« Reply #23 on: August 13, 2017, 03:59:06 pm »
I have been recording audio professionally from several decades before the Digital Age. The goal while "tracking" (recording the original sounds) is to maintain the optimal SNR (Signal to Noise Ratio) by keeping the average audio levels as high as possible without hitting the upper limit (clipping).  Clipping back in the analog era was much "softer" and more forgiving/flexible compared to digital recording where when you hit 0dBFS (Full Scale) you are up against a HARD limit, beyond which there is nothing but severe distortion.

When you do the math, you will see that a dynamic range of 16 bits is the equivalent of 96.33dB.  If you take the noise-floor of the Real World at around 20dB you will see that 96dB is quite more than enough to produce sounds over the complete range that our ears can handle.  Unless you want to ruin your ears with head-banging heavy metal at 120dB, I'll leave you to decide that for yourself.

The overwhelming majority of music consumed these days is typically both dynamic range compressed (see the "Loudness Wars") as well as data-compressed (i.e. MP3) and it is rare to find people listening to music that even amounts to 14-15 bits of dynamic range.

OTOH, when I am recording live performances (especially of amateur musicians) I prefer to use 24 bit recording sample depth. Not because it offers any significant improvement of SNR, but because it offers a significant increase in headroom so that I can record at a comfortable level while maintaining a substantial safety margin to handle unexpected amplitude excursions.  Of course, it is also true that the electronics (mostly the microphone preamp) has nowhere near the 144dB of dynamic range implied by a 24-bit sample depth. 

32 bits of sample depth implies 196dB of dynamic range which isn't even possible to achieve under normal atmospheric pressure on our planet.  Not to mention that you would need cryogenically-cooled electronics to even get half that dynamic range.  But 32-bit (or even floating-point) arithmetic is commonly used for intermediate calculations in most modern audio software to handle mixing and other effects without worrying about arithmetic overflow ("clipping").  Of course the final result is truncated back down to 16 bit for consumption in the Real World.

Ref:
https://en.wikipedia.org/wiki/Audio_bit_depth
https://en.wikipedia.org/wiki/Loudness_war

While a 20 dB noise floor may be realistic in a recording booth there are darn few real world situations where the noise is that low.  Most modern urban people get uneasy in truly quiet environments.  Maybe in a quiet room with earphones.  All of this adds fuel to your argument that 16 bits is more than enough for reproduction.
 

Offline Lukas

  • Frequent Contributor
  • **
  • Posts: 412
  • Country: de
    • carrotIndustries.net
Re: Can you truly make 16bit audio sound better by upping it to 32bit?
« Reply #24 on: August 13, 2017, 06:33:40 pm »
It's interesting how the "Sine" looks when zoomed in:



Square (also 10000 Hz)



It would be even more interesting to compare it with what comes out of the headphones, but that's not easy.

Watch the video by the xiph.org guy. Playing 'connect the dots' as audacity does it is the wrong way of rendering sampled signals. The correct way is to use sinx/x interpolation since this is what you'll get after putting the signal through a brick-wall filter that cuts off at nyquist frequency. Audacity can do this for you using Tracks->Resample and choosing a high sample rate.

When generating a 10kHz square wave, select "Square, no alias" and you'll get something that doesn't look like a square but is the correct band-limited approximation of a band-limited 10kHz square.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf