Pono Player... is it worth $400...?

[Dafinga]

No, do not get this. This article makes a convincing argument that playing high sample rates and bit rates through the PONO will actually make music sound WORSE : 1) 44.1 already has perfect fidelity; anything higher is not only unnecessary, but also introduces artifacts; 2) anything higher than 16 bit adds nothing to the sound while taking up way more space. The only element of possibly higher quality in the Pono could be its DAC. Excellent article!

http://xiph.org/~xiphmont/demo/neil-young.html

[drumphil]

The only element of possibly higher quality in the Pono could be its DAC.

Or its headphone amp. But you can get stuff for less than half that price with a decent headphone amp these days. It's probably a decent unit, but it's massively overpriced, and is sold on the basis of marketing bullshit (Neil Youngs name, and a bunch of audiophile bullshit).

A very decent way to help suckers solve the problem of having too much money.

[serr]

Originally Posted by Dafinga

No, do not get this. This article makes a convincing argument that playing high sample rates and bit rates through the PONO will actually make music sound WORSE : 1) 44.1 already has perfect fidelity; anything higher is not only unnecessary, but also introduces artifacts; 2) anything higher than 16 bit adds nothing to the sound while taking up way more space. The only element of possibly higher quality in the Pono could be its DAC. Excellent article!

http://xiph.org/~xiphmont/demo/neil-young.html

Haha.
Yeah, companies the likes of Prism, Apogee, & RME have been ripping studios off for years now with this fake 24 bit crap and fake high sample rates. Playing us silly engineers that don't know what we're hearing for fools. And now with Neal's help they're going to start ripping consumers off the same way! The horror!

I hope the above was just as facetious... That's the problem. Either someone is selling genuine snake oil like the $2000 USB cables or people go off the deep end the other direction and call sour grapes on modern digital audio. And the whole time miss that there are some really affordable really high quality products available these days. I don't know if this Pono is one of them or not. Like I said, too bad he's pushing ultra high research grade sample rates when the message should simply be 24 bit and good converters.

[brainwreck]

I re-read and re-watched Monty's stuff at xiph.org on digital audio, but I would be lying if I said that I understand all of it.

In the Digital Media Primer video, he says that a higher sampling rate allows for a much more gradual anti-aliasing filter (low pass for filtering off frequencies higher than Nyquist) than at lower sampling rates, further stating that steep filters are difficult (read: expensive) to build and are never completely successful. He goes on to say that it is essential to filter off anything above Nyquist after reconstruction. Ok, that seems to make sense, and it seems to be an argument for higher sampling rates at playback, not because of the presence of ultrasonics, but because of the anti-aliasing filter consideration.

On his page for Why 24/192 downlaods make no sense, he says that higher sample rates causes slight intermodulation distortion in amplifiers and speakers for some systems, which is an argument against higher sample rates:

Neither audio transducers nor power amplifiers are free of distortion, and distortion tends to increase rapidly at the lowest and highest frequencies. If the same transducer reproduces ultrasonics along with audible content, any nonlinearity will shift some of the ultrasonic content down into the audible range as an uncontrolled spray of intermodulation distortion products covering the entire audible spectrum. Nonlinearity in a power amplifier will produce the same effect. The effect is very slight, but listening tests have confirmed that both effects can be audible.

...

There are a few ways to avoid the extra distortion:

A dedicated ultrasonic-only speaker, amplifier, and crossover stage to separate and independently reproduce the ultrasonics you can't hear, just so they don't mess up the sounds you can.

Amplifiers and transducers designed for wider frequency reproduction, so ultrasonics don't cause audible intermodulation. Given equal expense and complexity, this additional frequency range must come at the cost of some performance reduction in the audible portion of the spectrum.

Speakers and amplifiers carefully designed not to reproduce ultrasonics anyway.

Not encoding such a wide frequency range to begin with. You can't and won't have ultrasonic intermodulation distortion in the audible band if there's no ultrasonic content.

They all amount to the same thing, but only 4) makes any sense.

Are there any benefits of higher sample rates at playback to consider for weighing against the con of possible audible intermodulation distortion?

On the same page, in the oversampling section, he goes on to say:

Sampling rates over 48kHz are irrelevant to high fidelity audio data, but they are internally essential to several modern digital audio techniques. Oversampling is the most relevant example [7].

Oversampling is simple and clever. You may recall from my A Digital Media Primer for Geeks that high sampling rates provide a great deal more space between the highest frequency audio we care about (20kHz) and the Nyquist frequency (half the sampling rate). This allows for simpler, smoother, more reliable analog anti-aliasing filters, and thus higher fidelity. This extra space between 20kHz and the Nyquist frequency is essentially just spectral padding for the analog filter.

...

That's only half the story. Because digital filters have few of the practical limitations of an analog filter, we can complete the anti-aliasing process with greater efficiency and precision digitally. The very high rate raw digital signal passes through a digital anti-aliasing filter, which has no trouble fitting a transition band into a tight space. After this further digital anti-aliasing, the extra padding samples are simply thrown away. Oversampled playback approximately works in reverse.

This means we can use low rate 44.1kHz or 48kHz audio with all the fidelity benefits of 192kHz or higher sampling (smooth frequency response, low aliasing) and none of the drawbacks (ultrasonics that cause intermodulation distortion, wasted space). Nearly all of today's analog-to-digital converters (ADCs) and digital-to-analog converters (DACs) oversample at very high rates. Few people realize this is happening because it's completely automatic and hidden.

I'm not making much sense of that this section, other than oversampling and a digital anti-aliasing filter allows for a smaller transition band (lower sample rate). For reconstruction, he (over?)simply states that oversampled playback is approximately the reverse... Off to wikipedia:

Oversampling in reconstruction

The term oversampling is also used to denote a process used in the reconstruction phase of digital-to-analog conversion, in which an intermediate high sampling rate is used between the digital input and the analogue output. Here, samples are interpolated in the digital domain to add additional samples in between, thereby converting the data to a higher sample rate, which is a form of upsampling. When the resulting higher-rate samples are converted to analog, a less complex/expensive analog low pass filter is required to remove the high-frequency content, which will consist of reflected images of the real signal created by the zero-order hold of the digital-to-analog converter. Essentially, this is a way to shift some of the complexity of the filtering into the digital domain and achieves the same benefit as oversampling in analog-to-digital conversion
https://en.wikipedia.org/wiki/Oversampling

The conclusion seems to be that oversampling elliminates the need for higher sample rates, allowing for a less expensive analog filter. So then, why use higher sample rates (as opposed to oversampling) at all, for either the adc and dac stages? Why do higher end converters feature high sample rates if it is a waste? Are there any benefits to using more expensive analog anti-aliasing filters, rather than oversampling?

[plush2]

Originally Posted by brainwreck

The conclusion seems to be that oversampling elliminates the need for higher sample rates, allowing for a less expensive analog filter. So then, why use higher sample rates (as opposed to oversampling) at all, for either the adc and dac stages? Why do higher end converters feature high sample rates if it is a waste? Are there any benefits to using more expensive analog anti-aliasing filters, rather than oversampling?

Because we can, marketing, to stay in step with new formats (blue ray deliveries are higher sampling rate). The move to higher sampling rates was more of a 'what if...?" proposition than a scientifically provable problem in need of a solution.

As a result of all that development we now have much better and much cheaper DAC/ADC designs on the market. When the marketing hype is on getting 192 or 384 kHz to market the companies don't care so much about giving away the farm on those lower sample rates. Don't look at it as your soundcard having useless high sample rate features, look at it as your inexpensive soundcard having a really effective regular sample rate performance.

[babag]

i'd like to hear some commentary on another area in the discussion.

i heard an interview, or part of one, with neil young while i was driving. the point he seemed to be making about the player was that it would allow listeners to listen at the same sample rates/bit depths that music was originally recorded at. that seems, to me, to imply a political purpose of pushing distribution channels into maintaining the integrity of format of the original recordings. aren't most, if not all, commercially distributed recordings resampled today? wouldn't the maintenance of format require offering something not offered today? just curious. young's phrasing was something to the effect of playing back 'in the format it was recorded in.'

thanks,
BabaG

[rothchild]

IMO this is being sold to the kind of people who think a Class A amplifier is necessarily better than a Class B amplifier simply because an A grade is higher than a B grade.

[rothchild]

Originally Posted by babag

i'd like to hear some commentary on another area in the discussion.

i heard an interview, or part of one, with neil young while i was driving. the point he seemed to be making about the player was that it would allow listeners to listen at the same sample rates/bit depths that music was originally recorded at. that seems, to me, to imply a political purpose of pushing distribution channels into maintaining the integrity of format of the original recordings. aren't most, if not all, commercially distributed recordings resampled today? wouldn't the maintenance of format require offering something not offered today? just curious. young's phrasing was something to the effect of playing back 'in the format it was recorded in.'

thanks,
BabaG

It's probably a tie in with these guys: https://www.meridian-audio.com/news-...authenticated/

A proprietary format that is supposedly 'signed' as being 'as the artist intended' (an article of faith for HiFi Heads) - yes indeed it will put you in the same room, on the same drugs and the in the same mood, listening to the same speakers as the artist when they got bored / ran out of money on the mix......

[heda]

wow the video on the page is so lame...
the most worrying thing is that the kickstarter was funded with $6,225,354
pledged of $800,000 goal

[David Else]

Crazy high sample-rates for playback is moronic, please can we catch these con artists and force them into a public ABX test!!!

I can very much recommend the Samsung YP-R0 mp3 player if you can get one, they are out of production now. This plays FLAC at 44.1/48 great.

If my Benchmark DAC1 headphone out is 9/10 then the YP-R0 is 7.5 out of ten. It is just not quite as 'tight' or 'real' as the DAC1, but for the price it is amazing Blows away other mp3 players I have heard.

The portable mp3 player market has been destroyed by the mobile phone, unfortunately usually the mobile phone sound quality is terrible.

Would have to hear this Pono to say if it sounds any good. Maybe the high sample-rate thing is bullshit marketing, but it might be worth it for the DAC and headphone amp...

[brainwreck]

Originally Posted by plush2

Because we can, marketing, to stay in step with new formats (blue ray deliveries are higher sampling rate). The move to higher sampling rates was more of a 'what if...?" proposition than a scientifically provable problem in need of a solution.

As a result of all that development we now have much better and much cheaper DAC/ADC designs on the market. When the marketing hype is on getting 192 or 384 kHz to market the companies don't care so much about giving away the farm on those lower sample rates. Don't look at it as your soundcard having useless high sample rate features, look at it as your inexpensive soundcard having a really effective regular sample rate performance.

I can understand r&d leading to better converters at lower sample rates, but why provide higher sample rates on end user devices? Why do video formats such as blue ray provide higher sample rates?

Outside of the discussion of sound quality for normal playback, higher sample rates are good for anything involving lower speed playback, such as studying recordings at half-speed.

[Nip]

Originally Posted by brainwreck

The conclusion seems to be that oversampling elliminates the need for higher sample rates, allowing for a less expensive analog filter. So then, why use higher sample rates (as opposed to oversampling) at all, for either the adc and dac stages? Why do higher end converters feature high sample rates if it is a waste? Are there any benefits to using more expensive analog anti-aliasing filters, rather than oversampling?

It's a different story to sample what is actually there in the analog signal - and to interpolate some kind of math average in between two samples when adding a sample.

So as I see it - you cannot replace higher samplerates by oversampling - even if oversampling improves a bit over running lower level samplerate, filter argument etc.

Do the experiment of thought and consider samplerate of 10kHz, and oversample this up to 40k or similar - then it's obvious you will loose information that cannot be recreated by math.

[plush2]

Originally Posted by brainwreck

I can understand r&d leading to better converters at lower sample rates, but why provide higher sample rates on end user devices? Why do video formats such as blue ray provide higher sample rates?

Outside of the discussion of sound quality for normal playback, higher sample rates are good for anything involving lower speed playback, such as studying recordings at half-speed.

The hypothesis that resulted in higher sample rates is eluded to in Nip's post. It's easily provable that there is sound that happens in those (+22-24 kHz) higher frequencies. The thought is that somehow, even though it is equally provable that no one yet discovered has the ability to hear those higher frequencies, there is something we are missing. The skeptics (of which I'm one) point to the growing pile of evidence that suggests the difference is entirely imperceptible to every human which is kind of the point of what we're doing. The proponents say there is anecdotal proof and it can't hurt. There's also a really easy assumption that more dots on that waveform line makes for a more accurate recording. It wouldn't be the first time an industry with reputable companies has tried to sell us something of questionable worth.

[Pollo]

Originally Posted by serr

Try this yourself:
A/Bing subtle differences well within the range of your perception to be a variable? Nope. Convert a file back and forth 100 times! If something starts to degrade, you'll hear it in the 100th pass.

Convert 48k to 44.1k and then back and forth 100 times. Obvious loss.
Convert 96k to 48k... loss.
Convert 96k to 44.1k... loss.
Convert 192k to 96k and then back and forth 100 times.
The first conversion to 96k and then back to 192k. Comparing the original and generational 192k gets you the faintest warble about 120db down when you phase cancel one from the other. I hear nothing on the A/B. But check this out: That 1st generation 192k file digitally cancels perfectly with the 100th generation file.

That tells me that the 96k container holds the complete music program with no loss. You can do the conversion/encoding math on it 100 times and it still phase cancels perfectly with the original.

Maybe I don't understand what you mean but to me this seems a misconception.
If you mean to do a samplerate conversion 100 times back and forth as in a mathematical process to convert one to the other, then your statement is simply not true.
Because it is a mathematical process there is no difference between converting from 48 to 44.1 or 192 to 96. The math is the same.
If you tried this and you get the results you describe than your sample rate converter is at fault.
I am not a methematician but I think that you can do lossless sample rate conversion if you do it right. That means you can do it a zillion times back and forth and still get exactly the same result.

If you meant re-recording at a different samplerate as in going from digital to analog and back to digital than it is a different story. That is not a lossless process. But it won't be for 192 to 96 either.

[brainwreck]

Originally Posted by Nip

It's a different story to sample what is actually there in the analog signal - and to interpolate some kind of math average in between two samples when adding a sample.

So as I see it - you cannot replace higher samplerates by oversampling - even if oversampling improves a bit over running lower level samplerate, filter argument etc.

Do the experiment of thought and consider samplerate of 10kHz, and oversample this up to 40k or similar - then it's obvious you will loose information that cannot be recreated by math.

What get's me on this topic is that the sampling theorem says that all frequencies falling below half of the sampling rate will be represented accurately. Is that for all wave shapes, or only sine waves? In real-world use cases, is this an issue?

Also, it seems that there would be some intermodulation distortion (in the playback system) when playing back tape. If that tape master is sampled at a low sample rate, it seems that 'defect' would be removed, resulting in a different sound between the original tape and the sampled copy. Is this a practical concern when choosing sample rates for distribution?

[brainwreck]

Originally Posted by plush2

The hypothesis that resulted in higher sample rates is eluded to in Nip's post. It's easily provable that there is sound that happens in those (+22-24 kHz) higher frequencies. The thought is that somehow, even though it is equally provable that no one yet discovered has the ability to hear those higher frequencies, there is something we are missing. The skeptics (of which I'm one) point to the growing pile of evidence that suggests the difference is entirely imperceptible to every human which is kind of the point of what we're doing. The proponents say there is anecdotal proof and it can't hurt. There's also a really easy assumption that more dots on that waveform line makes for a more accurate recording. It wouldn't be the first time an industry with reputable companies has tried to sell us something of questionable worth.

What about those higher frequencies being folded back, say for tape?

[plush2]

Originally Posted by brainwreck

What get's me on this topic is that the sampling theorem says that all frequencies falling below half of the sampling rate will be represented accurately. Is that for all wave shapes, or only sine waves? In real-world use cases, is this an issue?

Also, it seems that there would be some intermodulation distortion (in the playback system) when playing back tape. If that tape master is sampled at a low sample rate, it seems that 'defect' would be removed, resulting in a different sound between the original tape and the sampled copy. Is this a practical concern when choosing sample rates for distribution?

I like the way you put the question to Nip so I'll answer from that. It can represent only a sine wave in that very top nyquist frequency. As I understand it those other wave shapes necessarily contain harmonics which exist above their fundamental frequency which puts them up above perception anyways if you hold to the available evidence. If that explanation works for you then a sine wave at that frequency is the accurate representation of the part of the wave captured.

The audible part of the intermodulation would be captured. Basically, the part of any sound that you can hear is captured using the 44.1 or 48 kHz sampling rates. Any interplay that has gone on while capturing to tape has already happened, it is part of the recorded sound. It will not cease to happen or continue to happen after sampling. It is simply captured.

If you believe that ultra-sonic audio content is somehow working down into the audible realm through tape distortion then you may have a problem when laying back to tape. I don't know.

[brainwreck]

Originally Posted by plush2

I like the way you put the question to Nip so I'll answer from that. It can represent only a sine wave in that very top nyquist frequency. As I understand it those other wave shapes necessarily contain harmonics which exist above their fundamental frequency which puts them up above perception anyways if you hold to the available evidence. If that explanation works for you then a sine wave at that frequency is the accurate representation of the part of the wave captured.

That makes perfect sense.

Originally Posted by plush2

The audible part of the intermodulation would be captured. Basically, the part of any sound that you can hear is captured using the 44.1 or 48 kHz sampling rates. Any interplay that has gone on while capturing to tape has already happened, it is part of the recorded sound. It will not cease to happen or continue to happen after sampling. It is simply captured.

If you believe that ultra-sonic audio content is somehow working down into the audible realm through tape distortion then you may have a problem when laying back to tape. I don't know.

Yea, just to clarify, my question here concerns distortions introduced via ultrasonic frequencies affecting audible frequencies during playback of a master tape vs. the absence of ultrasonic frequencies in the playback of a low sample rate digital copy of the same tape. To clarify further, I'm posing a question here, not stating any opinion or fact:

Master tape playback (including ultrasonics) ---> playback system; results in intermodulation distortion and/or other 'defects' in the playback system (amps and speakers).

Digital copy of master tape (lacking ultrasonics) ---> playback system; lacking intermodulation distortion and/or other 'defects' which are introduced (in amps, speakers) during playback of the master tape.

Hopefully that makes more sense of my question. Edit: It might help if I outright state my actual question: Might a higher sample rate better represent the resulting audio of a master tape, considering the effects of ultrasonics on the audible range? If that were the case, an even higher sample rate might be needed to prevent aliasing of the ultrasonics. No? Again, this is all 'would if' stuff. Also, might something along these lines be behind statements by producers, engineers, and musicians that master tapes sound different[insert phrase: more real, organic, etc.] than digital copies?

I don't know if research has been done on this. I would think that, surely, someone has covered this. It may even be a reason for why ultra high sample rates have been explored, for example, 384k.

[David Else]

I think this post by Fabian from TDR might hold the answers you are looking for

https://www.gearslutz.com/board/mast...rocessing.html

[Nip]

Originally Posted by brainwreck

What get's me on this topic is that the sampling theorem says that all frequencies falling below half of the sampling rate will be represented accurately. Is that for all wave shapes, or only sine waves? In real-world use cases, is this an issue?

I'm not sure how Fourier analysis is defined, if as many physics phenomena - if a constonuous signal wave so and so - or applied to any music.

But if we say that Fourier is valid, and any music can be described as a series of sine wave frequencies summed - then Nyquist is probably valid for any music too.

What I know is a perfect square wave can be described as a series of odd harmonics with reduced amplitude according to a factor depending on order of harmonic.

So if a continuous signal any wave shape would be valid for Nyquist is my assumption. One sample each half period is enough according to Nyquist - to represent that frequency and below.

Where actuall real world case applies to be audible I don't know.

Maybe there are musical content to fall outside of Nyquist theorem.

Also, it seems that there would be some intermodulation distortion (in the playback system) when playing back tape. If that tape master is sampled at a low sample rate, it seems that 'defect' would be removed, resulting in a different sound between the original tape and the sampled copy. Is this a practical concern when choosing sample rates for distribution?

Any mechanical movement such as capstan for tape transport would introduce some irregularities and intermodulation as you say. I would compare it to jitter in digital world.

About your question about practical concern - it might make a difference as you say - if intermodulation is known to be in a certain frequency range - but wonder if this ever is the case.

If knowing that above 48k there is serious intermodulation distortion - one would stay at 48k sample rate - maybe and avoid getting that into digital domain.

As I read your question - better to oversample to 96k than sample that actual content from tape - could be so.

[karbomusic]

What I know is a perfect square wave can be described as a series of odd harmonics with reduced amplitude according to a factor depending on order of harmonic.

Probably more accurately stated that a perfect square wave creates an infinite series of odd order harmonics. Harmonics are a result of the shape of the wave, recreate the wave you get the same harmonics. Modify the wave, the harmonic structure follows. This can be demonstrated with a signal generator and an FFT in Reaper (or via analog gear) for those interested. Just to use guitar tones for some examples...

The distortion one hears in a distorted guitar tone for example, is in fact the harmonics created by the shape of the wave. Remove the harmonics, it would be a clean tone but to remove the harmonics you'd just change the shape of the waveform. For most any distorted or overdriven effect that I'm aware of (analog or otherwise), those harmonics are purposely rolled off quite a bit to remove the metallic qualities (usually all is gone by 7 kHz). So from that perspective, we don't even care about the 10-20kHz range for a overdriven guitar for those worried about sampling what they create.

As far as IM distortion per any medium, I can't think of any reason anyone would like how that sounds and most go to great lengths to make sure it doesn't occur. Using a guitar tone again as an example, that splatty nasty distortion which is the kind we don't like (usually heard as ugly splats bouncing around in the decaying note) is typically due to IM and we tend to run from it like the plague. I have a Fender Hot Rod Deluxe which has a really nice clean channel, but one of the most horrid drive/distortion sounds I have ever heard in my life . Much of that is IM if I remember my testing and is a completely analog process. Recording it digitally would be no problem and that nasty/horrid IM would most definitely be in the recording.

[karbomusic]

Is that for all wave shapes, or only sine waves? In real-world use cases, is this an issue?

Not an issue, sine wave is only used as an example to simplify the explanation and the fact that everything is a sum of sine waves. I do see a lot of discussion about harmonics but we should remember those are a result of the shape of the wave and should be thought of that way.

[plush2]

Since Fourier and decomposing audio into sine waves has come up here I wanted to share this video.

https://www.youtube.com/watch?v=8KmVDxkia_w

It's a 100+ year old mechanical fourier transform calculator. It basically demonstrates in a mechanical way how that theory actually works. I found it hugely helpful in actually grasping the idea.

[brainwreck]

Originally Posted by plush2

Since Fourier and decomposing audio into sine waves has come up here I wanted to share this video.

https://www.youtube.com/watch?v=8KmVDxkia_w

It's a 100+ year old mechanical fourier transform calculator. It basically demonstrates in a mechanical way how that theory actually works. I found it hugely helpful in actually grasping the idea.

Thanks for posting that link...crazy and very cool. I used to have a video link for a very simple machine that draws a single sine, but the video was taken down.

Edit for fun: https://www.youtube.com/watch?v=_EED8F7P_q4

[karbomusic]

very simple machine that draws a single sine

A speaker, duct tape, a pencil, 1hz signal and scrolling paper will do it.

[ernzo]

What mistifyes me the most with Digital audio and the Samplerate topic, is the fact that, even when the theory behind Sampling seems to be so simplistic?, Timbre information seems to be represented very well.. so well it's even spooky to think about it!


Let's take a classical guitar for the example..

What makes a guitar sound better/richer (timbre) than another, or a flamenco guitar sound different (timbre) than a regular Classical guitar?

Materials (choice of woods is super important), design, quality of work/craftmanship, physical synergy, chance/luck/serendipity?...
(ignoring the player or performance here, even tho it's another infinitely wide factor too)


Given the complexity of the instrument, I imagine one single classical guitar note, will not be made of one single sampled sinewave.. but MANY sinewaves will be needed to represent its harmonic and timbrical richness, right?

And all those digital sinewaves will have crazy modulations and intermodulations themselves as the instrument materials resonate along, and the note goes on and decays...

And even further, those sinewaves will not be of a fixed nature, where they forcibly have to correspond to the actual start and end of the note... but rather they will be continously and dynamically generated in Time, as some kind of lake-effect representation /reduction/simplification of the actual sound/timbre being played.... right?


With all of it tho, Im still mistifyed as how is it possible that strongly simplifyed/cut and mathematically re-generated group of sine waves are able to represent Timbre information so well, or music at all... when you are trying to record or reproduce not just one single classical guitar note, but actual guitar chords made of many notes, resonating in a very unique and deeply rich way; and 20 other instruments, each with its own unique timbre quality, resonating in its own equally complex/rich way and so on...

How on earh can it be so nicely represented?


And how does the limitations of recording hardware (mics? gear..), or filtering (highpass/lopass), affect the representation of those Freqs/sounds that are in origin the result of infinitely wide and detailed physical interactions of the instrument/materials/notes being played...?

We capture a fawlty and very filtered version of something that has been generated by something almost infinitely rich in spectrum...

So, wont it be beneficial in some cases to be capturing the sound SOURCE with a wider Freq spectrum and faster sampling resolution from the beginning, and try to include more of that microscopic information too?


I know unlike with Bit Depth, wich has a very noticeable effect on audio quality, the usual usual answer is NO..

But seeing it from this perspective, wont higher Sample Rates HELP at giving a better/clearer representation of all the crazy (inter)modulations and stuff going on that actually make up Timbre and all other nuances of sound?

wont higher sample rates offer a better resolution matrix for sound mixing, given the amount of detail and information of the MANY sampled sinewaves we're trying to mix?

Perhaps thats part of what we percieve as "lacking" in recorded audio vs the real thing..


PS: I can see the Nyquist 20khz thing, but beyond Freq, i'm concerned about granularity or resolution in time and depth..

And btw.. Aren't microphones outdated? Its aproach? How could it be made better?

[karbomusic]

Given the complexity of the instrument, I imagine one single classical guitar note, will not be made of one single sampled sinewave.. but MANY sinewaves will be needed to represent its harmonic and timbrical richness, right?

Until you hit a pure harmonic at fret 12 which gets very close to a pure sine wave because well that's what it is.

wont higher sample rates offer a better resolution matrix for sound mixing, given the amount of detail and information of the MANY sampled sinewaves we're trying to mix?

Broadly stated... No because that is exactly what nyquist solves, better said as "what do we need to do in order to capture the sound in its entirety". See video in my sig.

[lxm]

I love how all you guys are like "it's BULLSHIT"!!!!!... When you probably haven't even heard one. At least know what you are talking about BEFORE you give your ignorant opinion.

[ernzo]

Originally Posted by karbomusic

Until you hit a pure harmonic at fret 12 which gets very close to a pure sine wave because well that's what it is.

Well.. Harmonics may sound clean to us dipending on the guitar quality, but I dont think any guitar will be ever giving you a pure sine wave at 12fret harmonic, or any other..

It dosent matter how perfect it is made, it will never be perfect enough, or perfectly intonated enough.. and it will always have added timbrical information, as any note is made of many harmonics and subharmonics produced by the resonance of the materials...

So to represent that 12 fret harmonic I ¿guess? many sampled sinewaves will be necesary, wich again will be dynamically generated in time, as the sound goes on and decays..

And well, afterall all this imperfection/characrter is something necesary/desirable, or your guitar will be sounding more like a Synth than the real thing..!

Originally Posted by karbomusic

Broadly stated... No because that is exactly what nyquist solves, better said as "what do we need to do in order to capture the sound in its entirety". See video in my sig.

I sure know the video.. Having 44100 samples per second and 16bit is sure a lot, but still that is not telling us how is Timbre information preserved so well...

Again, beyond the nyquist 20khz thing - just as 24bit or 32bit-fp sounds much better than 16 bit, perhaps sample rate is also a factor when dealing with very complex audio material, or with very dense mixing afterall...

[karbomusic]

Originally Posted by ernzo

Well.. Harmonics may sound clean to us dipending on the guitar quality, but I dont think any guitar will be ever giving you a pure sine wave at 12fret harmonic, or any other..

I said that because I tested it on an oscilloscope myself many times and didn't use my ears.

It dosent matter how perfect it is made, it will never be perfect enough, or perfectly intonated enough.. and it will always have added timbrical information, as any note is made of many harmonics and subharmonics produced by the resonance of the materials...

So to represent that 12 fret harmonic I ¿guess? many sampled sinewaves will be necesary, wich again will be dynamically generated in time, as the sound goes on and decays..

You might try testing it yourself. The differences are more related to difficulty being able to create the harmonic cleanly yourself. What you may be missing is that a harmonic in this sense is isolating away all the timber which is the point. In other words timbre is the harmonic structure so a single harmonic is a pure sine wave. Something that is not a pure sine wave creates the harmonics/timbre.

I sure know the video.. Having 44100 samples per second and 16bit is sure a lot, but still that is not telling us how is Timbre information preserved so well...

Ah but it does! Because harmonics and timber are a direct result of the shape of the waveform.. properly recreate the waveform, you get the proper harmonic structure. They are not exclusive but dependent on one another, harmonics are a result of the waveform. See my other post you can test this for yourself in Reaper with a tone generator and FFT plugins.

[plush2]

Originally Posted by ernzo

But seeing it from this perspective, wont higher Sample Rates HELP at giving a better/clearer representation of all the crazy (inter)modulations and stuff going on that actually make up Timbre and all other nuances of sound?

wont higher sample rates offer a better resolution matrix for sound mixing, given the amount of detail and information of the MANY sampled sinewaves we're trying to mix?

Perhaps thats part of what we percieve as "lacking" in recorded audio vs the real thing..

It's that intuitive leap that explains the push for HD audio. There is no evidence to suggest we need it, just a hunch that it must make some positive difference. The problem is the intuition is based on misconceptions about how digital audio works. Fourier and Nyquist proposed these theories well before digital audio existed as a practical reaility. The point being they are not arguments for digital audio but rather the framework that makes it possible.

[ernzo]

Right.. it's quite a paradox in itself, almost magical in a sense, thats why i'm asking!

So was I right when assuming digital audio is recorded as different sine waves..
(I had ApTuner Harmonics analysis tool in mind when thinkin about it..
http://s50.photobucket.com/user/Regi...er_hg.jpg.html)

or is it rather one single sinewave representing the average sum of all the minute details of the audio being recorded?


Perhaps that is also part of the problem.. the linear nature of the Framework, as you say, we're working in.. from electric current, or microphone conception to digital audio in itself..!

[karbomusic]

Right.. it's quite a paradox in itself, almost magical in a sense, thats why i'm asking!

But it isn't, not sure why you keep going down the magical road when everyone keeps providing you explicit info and some of it even easily testable.

[ernzo]

I was not taking the Magical (or the Ego road, but rather wondering how all of it works in a Philosophical sense...

It's very easy to say "Nyquist!", but how much of an answer is that?

[brainwreck]

Originally Posted by ernzo

Right.. it's quite a paradox in itself, almost magical in a sense, thats why i'm asking!

So was I right when assuming digital audio is recorded as different sine waves..
(I had ApTuner Harmonics analysis tool in mind when thinkin about it..
http://s50.photobucket.com/user/Regi...er_hg.jpg.html)

or is it rather one single sinewave representing the average sum of all the minute details of the audio being recorded?


Perhaps that is also part of the problem.. the linear nature of the Framework, as you say, we're working in.. from electric current, or microphone conception to digital audio in itself..!

I get what you're saying. It's pretty amazing stuff.

Each output channel would be a sum of everything in the left or right channel. I guess technically, it is two waveforms representing everything.

[karbomusic]

From an academic perspective you should be able to setup up three or four tracks in reaper, all with JS tone generator. Set different sine wave frequencies for each. Then throw a VST oscilloscope (JS gfxscope) on the master. Start with a single track soloed, then add the next, next, next. You should be able to see the waveform sum and become complex.

[plush2]

Originally Posted by ernzo

Right.. it's quite a paradox in itself, almost magical in a sense, thats why i'm asking!

So was I right when assuming digital audio is recorded as different sine waves..
(I had ApTuner Harmonics analysis tool in mind when thinkin about it..
http://s50.photobucket.com/user/Regi...er_hg.jpg.html)

or is it rather one single sinewave representing the average sum of all the minute details of the audio being recorded?

You are correct in that digital audio can be viewed stored and manipulated as a series of sine waves and amplitudes (according to fourier) but it can also be viewed stored and manipulated as a complex waveform with a high enough sampling rate and amplitude values for each sample. They are both correct and, as far as known human hearing is concerned, lossless. The incorrect assumption is that this is because it's digital audio. Analog audio can also be broken down and analyzed as a series of sine waves as it can be 'sampled' measured and analyzed based on amplitude and frequency. Digital is simply an easy and powerful means of doing this. It was done in analog before digital came along.

[Badmotor]

I'm keen to actually physically try one. I've had acquaintances say that it sounds really good, so that's enough to make me interested.

[ernzo]

Originally Posted by plush2

You are correct in that digital audio can be viewed stored and manipulated as a series of sine waves and amplitudes (according to fourier) but it can also be viewed stored and manipulated as a complex waveform with a high enough sampling rate and amplitude values for each sample. They are both correct and, as far as known human hearing is concerned, lossless. The incorrect assumption is that this is because it's digital audio. Analog audio can also be broken down and analyzed as a series of sine waves as it can be 'sampled' measured and analyzed based on amplitude and frequency. Digital is simply an easy and powerful means of doing this. It was done in analog before digital came along.

So digital audio data can be read/represented both as series of sinewaves or as a complex waveform.. interesting!

I will have to take a closer look at Fourier, the sinewaves idea is somehow pleasing, but I guess the final result is exactly the same..

I was not questioning the validity of either method tho, just asking..