The No.1 Website for Pro Audio
Digital Audio and Sampling Rates
Old 19th October 2011
  #1
Gear Guru
 
UnderTow's Avatar
 
🎧 10 years
Lightbulb Digital Audio and Sampling Rates

Hi all,

There is a lot of talk on these forum on how sampling rates affect the sound of a recording. Many claims are made about how increasing the sampling rate is always better. Terms like "higher resolution" or "more accurate" are thrown around. Some of it is true. Some of it isn't. This article attempts to bring some clarity on some aspects of the subject.

Some points I will try to address: (Time permitting)

- Are higher sample rates more accurate?
- Inter Sample Peaks
- Oversampling Converters and Anti-alias filters
- Processing at higher rates and why (and why not)


I might try to add more as I have more time. This post is subject to being edited at any time to add things, clarify things, correct errors etc. It started off in another thread (now deleted) so bear with me as I find time to adjust it to make it more generally applicable.


============ Are higher sample rates more accurate? ============

Let's start with some waveform pictures to get an idea where some of the misunderstandings about sampling come from.

The first picture is a 5Khz sine wave sampled at 44.1Khz as shown in Sound Forge:


The sine wave doesn't look much like a sine wave at all. It looks very jagged and spiky.

Now the same 5Khz sine wave but resampled at 88.2Khz also shown in Sound Forge:



Notice that as we have added sample points by resampling, in this application, the sine wave looks more rounded and less jagged.

And now the same 5 Khz saine wave re-sampled to 176.4 Khz.



Now this looks much more like as sine wave.

Based on these pictures alone and other similar waveform representations in other editors and DAWs, it would be easy to conclude that as we increase the sampling rate, we have increased the accuracy of the sine wave.

But there is an obvious snag in that theory: We started with a 5 Khz at 44.1Khz. How does Sound Forge know how to make the sine wave more rounded as we re-sample and add sample points? If the wave was not a proper sine at 44.1 Khz, Sound Forge would not know where to add the sample points to form the waves you see in the latter pictures. So what is going on here?

Here is a picture of that same 5Khz sine wave at 44.1 Khz but this time as shown in Audition:



This looks again like a proper sine wave. How is this possible?

The big difference between these two applications is that Sound Forge (at least this particular version) creates it's visual representations by simply drawing a straight line between the sample points.(The little blue squares). Audition on the other hand shows a picture of how the reconstructed waveform will look. The reconstructed waveform is the waveform that is actually produced at the output of your converters. The analogue signal. So how does Audition do this?

How do we go from seemingly jagged lines to nice curves in Digital Audio? Here is some background that might help to understand.


The clue to this whole story is the way that complex waveforms can be seen as a series of summed up sine waves. (Thanks to Mr Fourier for figuring that out).

To illustrate this, the following animation starts off with a single sine wave and adds an increasing number of odd harmonics to form what approaches a square wave:





The more odd harmonic one adds, the closer one gets to a square wave. With an infinite number of odd harmonics, we could create a perfect square wave. (And as we can never have an infinite number of anything, there are no perfect square waves in nature).

Here is a similar animation but this time we create a triangle wave:




The reason it becomes a triangle wave is because the level of the harmonics rolls off faster as they get higher compared to the harmonics of a square wave.

And here we have an animation of a sawtooth wave being created: (In this case we add even and odd harmonics)




Now, if you go in the opposite direction and start with a waveform that is a theoretical square, sawtooth or triangle wave and start removing the harmonics, as you progressively remove them, the rounder and more curve like the waveform becomes. The exact same thing happens if we start with a more complex but jagged and pointy waveform like the first picture in this post. Remove the harmonics (all those jagged and pointy angles in the waveform) of the signal and you end up with a smooth rounded waveform just like you see in picture two.

What are harmonics and how do we remove them? Harmonics are higher frequency sine waves that have a mathematical relation to the base frequency. And how do we remove higher frequencies? We filter them out with a low-pass filter and a low-pass filter is exactly what you will find in the output of any quality DAC.

If one were to start with a perfect square wave and remove all the harmonics and just keep the fundamental base frequency we would get a perfect sine wave. Not only that, we can mathematically predict every single aspect of that sine wave before we even start removing the harmonics. The same rules apply to sawtooth waves or triangle waves.

Thanks to the work of geniuses like Mr Fourier, Mr Shannon and Mr Nyquist, we also know that the same thing applies to complex periodic waveforms and even random waves. We can predict mathematically exactly what will happen when we filter out higher frequencies. That is why an application like Adobe Audition or iZotope RX can show us exactly what the reconstructed waveform (the one that has been filter by your DAC) will look like before it gets anywhere near your DAC!


This also brings us to the extremely important point in all these discussions about increasing sampling rates: It does not give us any more resolution! The increased sampling rate just allows us to sample higher and higher (inaudible) harmonics. There is not any more precision in the output within the frequency range we want to sample. The stuff we can hear. Again, this is important!

That waveform that looks like a jagged mess in some audio applications, just like in the first picture in this post, will look like the nice smooth wave in the second picture by the time it comes out of your DAC.

To drive the point home it is important to understand that our ears also function as low-pass filters. Even if you increase the sampling rate of your system, your ear is filtering out all those upper harmonics anyway!


I hope this brings some clarity to the topic.

Later I'll write about aliasing in (plugin) processing and how higher sampling rates can sometimes benefit this process (and sometimes not)...

Alistair

Last edited by UnderTow; 21st April 2016 at 02:12 PM..
Old 19th October 2011
  #2
Gear Guru
 
UnderTow's Avatar
 
🎧 10 years
=========================== Inter Sample Peaks ==============================================


Which brings us to the topic of Inter Sample Peaks: Certain signals, or rather I should say certain combination of sample points (it isn't exactly an audio signal until it has gone through the reconstruction filter of your DAC as that is part of system), when filtered at the DAC will cause wave forms that "overshoot". This example I created a few years ago illustrates the point well:




Although all the sample points are within 0 dB FS, if you look at the highest peak this signal causes after filtering by the reconstruction filter, it is at above +6 dB FS! If your DAC has enough headroom in the analogue components, it will happily recreate that +6 dB FS peak (or rather a peak 6 dB higher in voltage than whatever level 0 dB FS represents depending on what analogue level your DAC is calibrated to). If there is not enough headroom in the design, this signal will cause analogue clipping in your DAC.
Old 19th October 2011 | Show parent
  #3
Gear Guru
 
UnderTow's Avatar
 
🎧 10 years
=========================== Oversampling Converters and Anti-alias filters ==============================================

The Nyquist-Shanon sampling Theorem tells us: "A band-limited analogue signal that has been sampled can be perfectly reconstructed from an infinite sequence of samples if the sampling rate exceeds 2B samples per second, where B is the highest frequency of the original signal". To fulfil these requirements, all Analogue to Digital Converters need anti-aliasing filters to remove all frequencies at or above half the sampling rate. (Also known as the Nyquist frequency).

At a 44.1Khz sampling rate this means that all frequencies at or above 22050 Hz need to be removed (half the sampling rate) but we want to preserve everything within the audible range which is below about 20Khz. This gives us a bandwidth of 2050 Hz to achieve the filtering (also known as the filter's transition band). With a sampling rate of 88.2 Khz, the Nyquist frequency moves up to 44100 Hz but our lower limit remains at about 20 Khz. We now have a possible transition band of 24100 Hz. At a sampling rate of 176.4 Khz the allowed transition band of the filter becomes 62500 Hz.

In the past converters used analogue filters to remove frequencies above Nyquist. It is impossible to design an analogue filter that let's everything through transparently at 20Khz but filters everything out at 22050 Hz so compromises had to be made. Either you allow a significant amount of frequencies through that are above Nyquist. This causes audible aliasing. Or you let the filter encroach on the audible range below 20Khz. In this kind of design, increasing the sampling rate and relaxing the anti-aliasing filter requirements will make a significant difference to the sound. 88.2 Khz (and 96 Khz etc) will sound better than 44.1 or 48 Khz sampling rates.

The solution to this problem is to sample at (much) higher rates and then to use digital filters to convert the signal down to usable sample rates. That is exactly what all modern converters do.

The flagship converters of the top four ADC chip manufacturers (Cirrus Logic CS5381, Wolfson WM8786, Asahi Kasei ak5394a, TI Burr Brown PCM4222) all run their modulators (the part that actually does the sampling by measuring the analogue voltages), at 5.6448 Mhz for sampling rates of 44.1, 88.2 or 176.4 Khz and 6.1444 Mhz for sampling rates of 48, 96 or 192 Khz. This means that the analogue anti-aliasing filter now only needs to remove frequencies at or above 2.8224 Mhz (or 3,072 Mhz for 48, 96 or 192 Khz sampling rates). Such an analogue filter can be made absolutely transparent at 20Khz (and even higher).

Once the analogue source signal is sampled at this high oversampling rate (128x for 44.1 or 48Khz) and turned into a digital signal, it is then converted down to the base target rate using a digital "decimation" filter. In the digital domain it is very possible to create a brick-wall filter that is transparent in the audible band yet removes enough at 22050 Hz to avoid any audible aliasing. That is why quality converters can be transparent at 44.1hz.



Here are the respective specification sheets of these converters for anyone that is interested:

http://www.asahi-kasei.co.jp/akm/en/...5394a_f03e.pdf
http://www.wolfsonmicro.com/document.../en/WM8786.pdf
http://www.ti.com/lit/ds/symlink/pcm4222.pdf
http://www.cirrus.com/en/pubs/proDat.../CS5381_F2.pdf

Of course this is all about quality converters. Not all converters use these chips (even though they are relatively cheap) and the chips are not the only part of the full design. The analogue sections, the power supply, any external clock used etc all can have an impact on the performance of any particular converter or interface.
Old 19th October 2011 | Show parent
  #4
Gear Guru
 
UnderTow's Avatar
 
🎧 10 years
Here is a little blurb about dither, bit depth and resolution. I'm copying it here as it seems relevant to the topic of Digital Audio in general. (It isn't entirely proof read and comes from another thread so excuse any out of context comments or references).

Below is an example of a sine wave that is low enough to only modulate the lowest bit (LSB) of wave file. A properly dithered signal that modulates only the lowest bit won't look like a sine wave because of all the dither noise. Here is an example with a 23 Khz sine modulating the lowest bit of a 48Khz file:



The sine wave is still there (See further on). Actually, it goes even further than that. You can take a 23 Khz sine wave that peaks at let's say -60 dB in a 48 Khz 24 bit file, convert to 8 bit and the sine wave will still be there if you used proper dither.

That is right. You can encode a -60 dB signal in a 8 bit file. It will be in the noise floor but it will still be there. You can easily see that sine wave on a frequency analyser:



(I used TPDF dither in the above example so there is a lot of noise in the signal).



Now to make things more interesting, here is a 1Khz sine wave at -60 dB in a 48 Khz 32 bit float file:



This is the same file converted to 8 bit with heavy noise shaping dither:



(Note the heavy noise shaped dither signal peaking at the top of the image. It is actually louder (brighter) than the 1 Khz sine).

This is what it looks like on a spectrum analyser:



Note the level at which the 1 Khz sine is peaking.

This is what it sounds like: http://puretone.nl/1Khz-60dB-8bit.wav The sine is perfectly clear and undistorted.

As you can see in the spectrum analyser, I could have made that sine another 35 dB softer and it would still be easily detectable and probably easily audible. That would be a -105 dB signal still completely whole and easily detectable in a 8 bit file.

So to resume in short: bit depth determines the dynamic range and noise floor of a digital audio signal. Sample rate determines the bandwidth. Not resolution.

Alistair
Old 19th October 2011 | Show parent
  #5
Gear Guru
 
UnderTow's Avatar
 
🎧 10 years
<place holder 4>
Old 19th October 2011 | Show parent
  #6
Gear Guru
 
theblue1's Avatar
 
🎧 15 years
One point that may seem largely semantic, yet I think is fundamental to the processes you describe...

I would suggest to you that, while the data points (sample values) do not -- by themselves -- constitute signal, those data points, in combination with the context of the interpretive rules of a given digital format do form a synergism that represents signal in the digital realm.

There must be continuity of signal in order for it to pass along a signal chain. Signal can be analog. Signal can be digital. But it is still conveying the same information along the path.
Old 19th October 2011 | Show parent
  #7
Lives for gear
 
🎧 15 years
Thanks for taking the time to explain, I too get fed up with the misinformation being spread in those frequent threads as it seems like few search before posting. Unfortunately you have marketers on this site that push "the higher the number, the better it is" mentality to sell product. Additionally there is product out there not using the better converter chips with cheap or bad analog circuits where 96k can sound better than 44.1k or 48k with those SPECIFIC converter units, so users of those units posting 96k sounds better are just sharing their personal experience (it would help with the discussion if they also posted what converters they were using too). To further cloud the waters some converter units do not aim for transparency and instead add pleasing color (I want my converters to be transparent). You might want to touch on the analog side of converters while you are at it. Bottom line is with a decent converter, 44.1k or 48k will get the job done. I thought Lavry's position was that 60k sampling would have been a great standard but that ship has sailed.
Old 19th October 2011
  #8
Gear Guru
 
UnderTow's Avatar
 
🎧 10 years
Thanks for the feedback. In Place holder 4 I will try and wrap everything up with comments about real world situations as Bassmankr mentions in the post above.

Also I'd like to make a list of suggested reading. If anyone has anything they feel should be linked for further reading on the topic, please mention it.

theblue1, I'm thinking of how best to phrase the relationship between the digital and analogue signals. Thanks for the comment.

Alistair
Old 19th October 2011 | Show parent
  #9
MediaMix
Guest
Quote:
Based on these pictures alone and other similar waveform representations in other editors and DAWs, it would be easy to conclude that as we increase the sampling rate, we have increased the accuracy of the sine wave.

But there is an obvious snag in that theory: We started with a 5 Khz at 44.1Khz. How does Sound Forge know how to make the sine wave more rounded as we re-sample and add sample points? If the wave was not a proper sine at 44.1 Khz, Sound Forge would not know where to add the sample points to form the waves you see in the latter pictures.
Question. Does this apply to recording a string section at 44.1 vs 96khz/192khz or just a resampled 5k sine wave?
Old 19th October 2011 | Show parent
  #10
Gear Guru
 
UnderTow's Avatar
 
🎧 10 years
Quote:
Originally Posted by MediaMix View Post
Question. Does this apply to recording a string section at 44.1 vs 96khz/192khz or just a resampled 5k sine wave?
It applies to any signal. And to go even further, a recorded string section is a summing of sine waves. Any audio signal can be de-constructed into the component sine waves.

Alistair
Old 19th October 2011 | Show parent
  #11
MediaMix
Guest
Quote:
Originally Posted by UnderTow View Post
It applies to any signal. And to go even further, a recorded string section is a summing of sine waves. Any audio signal can be de-constructed into the component sine waves.

Alistair
So recording a rock band thru lavry converters at 192khz will sound no different then recording them at 44.1khz?
Old 19th October 2011 | Show parent
  #12
Gear Guru
 
theblue1's Avatar
 
🎧 15 years
Quote:
Originally Posted by UnderTow View Post
Thanks for the feedback. In Place holder 4 I will try and wrap everything up with comments about real world situations as Bassmankr mentions in the post above.

Also I'd like to make a list of suggested reading. If anyone has anything they feel should be linked for further reading on the topic, please mention it.


theblue1, I'm thinking of how best to phrase the relationship between the digital and analogue signals. Thanks for the comment.

Alistair
Yeah... I knew what you meant. If one simply refers to digital signal or analog signal, I think that eliminates some potential confusion.

Certainly, in the sense I'm using it, signal exists in both realms. In the case of an analog audio signal, the 'interpretation' is considerably more straightforward and the context more 'forgiving.' Even if we don't know reference levels and such, we can probably extract meaningful audio program material out of it.

Signal in the digital realm can be viewed, I think, as a synergism between the data points of the individual sample values collected and knowledge of the digital format under which they were sampled, with which we can extract the information encoded in that digital signal.


I'll leave the hard, tweaky, mathematical stuff to you guys with the proper education... I feel more comfortable with the big picture stuff because, well, you can usually just approach it with some basic knowledge and some basic common sense and sort it out. heh
Old 19th October 2011 | Show parent
  #13
Gear Guru
 
theblue1's Avatar
 
🎧 15 years
Quote:
Originally Posted by MediaMix View Post
So recording a rock band thru lavry converters at 192khz will sound no different then recording them at 44.1khz?
Trick question!

heh

Lavry Engineering


I think Bassmankr's point about theoretical potential versus real world implementations is an important one.

In terms of practical reality, I tend to lean toward the opinion expressed by many that even though a 44.1 kHz sample rate with a well-designed, properly performing device can deliver accurate signal across the nominal audible band,* it's probably easier to deliver accuracy across a suitable frequency bandwidth by using somewhat higher sample rates, which allows the use of more gradual anti-alias roll off and so forth. (I'm going to leave modern oversampling out of this because, frankly, thinking about it makes my head hurt. heh )

And that would seem to be supported by anecdotal and other evidence that suggests that some high end converters deliver performance at 44.1 kHz indistinguishable from, say, 96 kHz, across the audible band even as some lesser quality converters are perceived to deliver better performance at 96 than at 44.1.



* Of course, we typically peg the nominal audible band at 20-20 kHz, despite the fact that there are outliers, particularly among the very young, who can perceive audio somewhat above that range, possibly as high as 22-24 kHz and there are some who perceive very low frequency as sound while others do not.
Old 19th October 2011 | Show parent
  #14
MediaMix
Guest
Let me rephrase:

So recording a rock band thru PRISM/APOGEE/AVID converters at 192khz will sound no different then recording them at 44.1khz?
Old 19th October 2011 | Show parent
  #15
Lives for gear
 
🎧 15 years
Quote:
Originally Posted by MediaMix View Post
Let me rephrase:

So recording a rock band thru PRISM/APOGEE/AVID converters at 192khz will sound no different then recording them at 44.1khz?
Correct. If you can hear a difference then one of the following maybe true:

1. Your test is flawed and you have some other sound modifying processes engaged.
2. Your DAC is broken

The fundamental thing to understand is that there will be no sonic material out of audio band, > 22kHz, that is significant. Your mics might get to 15kHz, your instruments might get to 10kHz, there is no source material anywhere near the 22kHz to worry about.
Old 19th October 2011 | Show parent
  #16
Moderator
 
narcoman's Avatar
 
🎧 10 years
vote for sticky.

Well done UT.
Old 19th October 2011 | Show parent
  #17
MediaMix
Guest
Quote:
Originally Posted by David Spearritt View Post
Correct. If you can hear a difference then one of the following maybe true:

1. Your test is flawed and you have some other sound modifying processes engaged.
2. Your DAC is broken
We work at 24/44 or 24/48. But I never argue with someone who says they work at 88.2/96 because it sounds better. Whatever floats their boat.

Another question. So working in the daw at 88.2/96 is a waste of resources? Many have stated that plug-ins sound better at the higher sample rates and the overall sound is better, including many of the developers here like Cytomic. Is this a myth as well?
Old 19th October 2011 | Show parent
  #18
Gear Guru
 
UnderTow's Avatar
 
🎧 10 years
Quote:
Originally Posted by MediaMix View Post
Another question. So working in the daw at 88.2/96 is a waste of resources? Many have stated that plug-ins sound better at the higher sample rates and the overall sound is better, including many of the developers here like Cytomic. Is this a myth as well?
It is neither truth nor myth in all cases as it depends on the approach used by the plugin (or digital hardware) developers. It also depends on the actual process (EQ, dynamics, distortion etc). I hope to address some of these points in my upcoming post about processing. That is why there are still place holders. :-)

Alistair
Old 19th October 2011 | Show parent
  #19
Lives for gear
 
🎧 10 years
Quote:
Originally Posted by UnderTow View Post
It is neither truth nor myth in all cases as it depends on the approach used by the plugin (or digital hardware) developers. It also depends on the actual process (EQ, dynamics, distortion etc). I hope to address some of these points in my upcoming post about processing. That is why there are still place holders. :-)

Alistair
In all seriousness. Dude, you should design plugins or something that will put your views on native processing or converters into practical use. Unless you do that already.
Old 19th October 2011 | Show parent
  #20
Gear Head
 
🎧 10 years
The Pacific Northwest section of AES did an extensive discussion on this matter.

February 2004 From Hear to Eternity - Sampling, Conversion and the Limits of Hearing. An all-day seminar with James Johnston, Dr. Melissa Harrison, Dr. Richard Cabot, Steven Green, and Bob Moses.

I was luck enough to attend this meeting, but I am not an AES member so could not access the archives. For those truly interested in the science of sample conversion it would be worth trying to get a transcript.

The conclusion was that there is improvement in audio that can be perceived by the human ear after exceeding a sample rate of 60hz, however there is not an audible benefit once exceeding that sample rate.
Old 20th October 2011
  #21
Lives for gear
 
🎧 5 years
Excellent read. Another favourite added to my list of sources that dismisses this outrageous fallacy about high sample rates. Smh

Thanks!
Old 20th October 2011 | Show parent
  #22
Audio Alchemist
 
Lagerfeldt's Avatar
Good post, looking forward to the currently missing graphics. Will link to this thread from the FAQ in my own forum.
Old 20th October 2011 | Show parent
  #23
restpause
Guest
Quote:
Originally Posted by UnderTow View Post
=========================== Oversampling Converters and Anti-alias filters ==============================================

The Nyquist-Shanon sampling Theorem tells us: "A band-limited analogue signal that has been sampled can be perfectly reconstructed from an infinite sequence of samples if the sampling rate exceeds 2B samples per second, where B is the highest frequency of the original signal". To fulfil these requirements, all Analogue to Digital Converters need anti-aliasing filters to remove all frequencies at or above half the sampling rate. (Also known as the Nyquist frequency).

At a 44.1Khz sampling rate this means that all frequencies at or above 22050 Hz need to be removed (half the sampling rate) but we want to preserve everything within the audible range which is below about 20Khz. This gives us a bandwidth of 2050 Hz to achieve the filtering (also known as the filter's transition band). With a sampling rate of 88.2 Khz, the Nyquist frequency moves up to 44100 Hz but our lower limit remains at about 20 Khz. We now have a possible transition band of 24100 Hz. At a sampling rate of 176.4 Khz the allowed transition band of the filter becomes 62500 Hz.

In the past converters used analogue filters to remove frequencies above Nyquist. It is impossible to design an analogue filter that let's everything through transparently at 20Khz but filters everything out at 22050 Hz so compromises had to be made. Either you allow a significant amount of frequencies through that are above Nyquist. This causes audible aliasing. Or you let the filter encroach on the audible range below 20Khz. In this kind of design, increasing the sampling rate and relaxing the anti-aliasing filter requirements will make a significant difference to the sound. 88.2 Khz (and 96 Khz etc) will sound better than 44.1 or 48 Khz sampling rates.

The solution to this problem is to sample at (much) higher rates and then to use digital filters to convert the signal down to usable sample rates. That is exactly what all modern converters do.

The flagship converters of the top four ADC chip manufacturers (Cirrus Logic CS5381, Wolfson WM8786, Asahi Kasei ak5394a, TI Burr Brown PCM4222) all run their modulators (the part that actually does the sampling by measuring the analogue voltages), at 5.6448 Mhz for sampling rates of 44.1, 88.2 or 176.4 Khz and 6.1444 Mhz for sampling rates of 48, 96 or 192 Khz. This means that the analogue anti-aliasing filter now only needs to remove frequencies at or above 2.8224 Mhz (or 3,072 Mhz for 48, 96 or 192 Khz sampling rates). Such an analogue filter can be made absolutely transparent at 20Khz (and even higher).

Once the analogue source signal is sampled at this high oversampling rate (128x for 44.1 or 48Khz) and turned into a digital signal, it is then converted down to the base target rate using a digital "decimation" filter. In the digital domain it is very possible to create a brick-wall filter that is transparent in the audible band yet removes enough at 22050 Hz to avoid any audible aliasing. That is why quality converters are transparent at 44.1hz.



Here are the respective specification sheets of these converters for anyone that is interested:

http://www.asahi-kasei.co.jp/akm/en/...5394a_f03e.pdf
http://www.wolfsonmicro.com/document.../en/WM8786.pdf
http://www.ti.com/lit/ds/symlink/pcm4222.pdf
http://www.cirrus.com/en/pubs/proDat.../CS5381_F2.pdf

Of course this is all about quality converters. Not all converters use these chips (even though they are relatively cheap) and the chips are not the only part of the full design. The analogue sections, the power supply, any external clock used etc all can have an impact on the performance of any particular converter or interface.
Thanks for this more modern info.
Old 20th October 2011
  #24
Lives for gear
 
oudplayer's Avatar
Since many converters that are widely used do not use these particular chips, it might be good to overview other implementations of filtering? That said, I really appreciate the work you've done already on this, it's a great contribution.

Also,
Quote:
So recording a rock band thru PRISM/APOGEE/AVID converters at 192khz will sound no different then recording them at 44.1khz?
Cymbals will "sound" different, since they produce energy up to 30khz. I put "sound" in quotes since we can easily show a difference in spectral content, but not everyone might be able to hear the difference, nor would different people hear the same difference in a similar way. Whether or not the inclusion of sound between 20khz-30khz is beneficial or detrimental is a lingering question.
Old 20th October 2011 | Show parent
  #25
Lives for gear
 
frans's Avatar
 
10 Reviews written
🎧 15 years
Sticky!
Thank your for cutting through the noisefloor of halfcooked blah about A/D D/A conversion. I hope the audio-barbarians will be repelled by the ammount of words you used and not dilute this thread.
May I ask if you will include the topic of intersample peaks vs. consumer-level-D/A chips and the tarpits of the uneven distribution of dynamic resolution inherent in PCM?
Old 20th October 2011 | Show parent
  #26
Lives for gear
 
🎧 10 years
Quote:
Originally Posted by MediaMix View Post
Many have stated that plug-ins sound better at the higher sample rates and the overall sound is better, including many of the developers here like Cytomic. Is this a myth as well?
Short answer: This is not a myth. Many plugins, especially the ones that do nonlinear processing (saturation, compression) sound better at higher sample rates.

The reason is NOT that the higher sample rates are more accurate, but that nonlinear procession generates overtones/harmonics above nyquist (fs/2) which results in aliasing.

To avoid this there has to be some bandlimiting implemented into the code, which can be very difficult to code (oversampling is one of the common methods)

So in realitiy, plugins that sound better at higher rates are just "badly" coded (the developer didn't take care about avoiding aliasing)
Old 21st October 2011
  #27
Lives for gear
 
🎧 10 years
+1 For sticky!!

Quote:
Originally Posted by oudplayer View Post
Cymbals will "sound" different, since they produce energy up to 30khz. I put "sound" in quotes since we can easily show a difference in spectral content, but not everyone might be able to hear the difference, nor would different people hear the same difference in a similar way. Whether or not the inclusion of sound between 20khz-30khz is beneficial or detrimental is a lingering question.
I'm confused are you saying any human can hear this?
Old 21st October 2011 | Show parent
  #28
Audio Alchemist
 
Lagerfeldt's Avatar
Quote:
Originally Posted by frans View Post
May I ask if you will include the topic of intersample peaks vs. consumer-level-D/A chips [...]
Here's some stuff while you're waiting ;-)

Inter-sample peaks
https://www.gearslutz.com/board/tips-...ple-peaks.html

0 dBFS+ Levels in Digital Mastering
http://www.tcelectronic.com/media/ni...0_0dbfs_le.pdf

Overload in Signal Conversion
http://www.tcelectronic.com/media/ni...3_overload.pdf

Stop Counting Samples
http://www.tcelectronic.com/media/lu...les_aes121.pdf

Programmed for Distortion
Listen to the artifacts produced when hot CDs are sample rate converted or reproduced in a CD player.
http://www.tcelectronic.com/media/Pr...Distortion.zip

Why Very High Sample Rates Do Not Sound Better
https://www.gearslutz.com/board/1234224-post72.html
Old 21st October 2011 | Show parent
  #29
Lives for gear
 
🎧 10 years
Nice one Alistair - another vote for stickie
Old 21st October 2011 | Show parent
  #30
Lives for gear
 
jupiter8's Avatar
 
🎧 10 years
Quote:
Originally Posted by UnderTow View Post
This means that the analogue anti-aliasing filter now only needs to remove frequencies at or above 2.8224 Mhz (or 3,072 Mhz for 48, 96 or 192 Khz sampling rates). Such an analogue filter can be made absolutely transparent at 20Khz (and even higher).
Cool post thanks a lot,it is sorely needed i'm afraid.

Question:in the above scenario is it even necessary to have an analog filter, since there will be no energy present at those frequencies and even if it is, it will be filtered by the digital filter, since the aliasing will be way above hearing range. Am i missing something ? Just curious.
πŸ“ Reply

Similar Threads

Thread / Thread Starter Replies / Views Last Post
replies: 184 views: 74285
Avatar for audiomizer
audiomizer 22nd June 2015
replies: 609 views: 50017
Avatar for psycho_monkey
psycho_monkey 7th February 2012
replies: 9 views: 2841
Avatar for Robo
Robo 15th December 2020
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
πŸ–¨οΈ Show Printable Version
βœ‰οΈ Email this Page
πŸ” Search thread
πŸŽ™οΈ View mentioned gear
Forum Jump
Forum Jump