• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Benefits of capturing audio at higher sampling rates

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
That was an output of a concrete guitar amp described in the article, in its overdrive (clipping) mode. The article is pretty educational. Such guitar "amp" is not really a conventional amplifier per se, but a highly-non-linear electronic musical instrument, which is controlled by yet another highly-non-linear electromechanical musical instrument - the guitar.

So it was a clipped amplifier output signal and there are still no axis to the scope plot so we really don't know what the frequency was nor the amplitude.
Have you seen what the speaker will reproduce when fed with that amp output... something totally different !

The goal of hifi reproduction is to reproduce that faithfully. So far you have not been able to show that even 44.1 can't reproduce the sound of the electric guitar coming from a speaker that is limited to around 4kHz.
 

Cosmik

Major Contributor
Joined
Apr 24, 2016
Messages
3,075
Likes
2,180
Location
UK
...a simple way to escape the tyranny of imperfect speakers and uncontrolled room acoustics is to use studio-quality headphones.
But there's a very interesting timing-related aspect to that: headphones cannot reproduce Blumlein stereo. Blumlein stereo provides no interaural timing difference because the mics are coincident, and it is the acoustic crosstalk to the 'wrong' ear from speakers mixed with the direct that creates a real, measurable timing difference at the ears.

Purist recordings that are optimised for speakers will use Blumlein stereo mic pairs while those for headphones will use spaced omnis of some kind (I presume) in order to encode a timing difference directly in the recording. And the ordinary panpot mimics the Blumlein pair, so it is optimal for speakers not headphones.

So you can't just swap between speakers and headphones without timing related side effects.

And headphones, of course, don't present the recording in a natural way: it is 'in your head'. I don't particularly mind that, but for the most natural portrayal of a typical recording I prefer speakers.

But I guess what you're arguing - and this would apply to the multiway speaker case - is that while the system plays fast and loose with 'timing' between channels and frequency bands, it is important to maintain the integrity of waveforms within, at least, their own channel, and the driver's own frequency band (although it is becoming a bit of a tenuous argument by that stage :)). Probably true, but I'm not personally worried about anything above CD sample rates.
 
OP
Sergei

Sergei

Senior Member
Forum Donor
Joined
Nov 20, 2018
Messages
361
Likes
272
Location
Palo Alto, CA, USA
Couldn't we fairly describe this as a significant negative result and maybe move on?

Ah, boolean logic! Beautiful abstraction. Yet not always sufficient. To me, it is neither positive nor negative result. The experiment revealed important factors beyond the simple yes-or-no frame of inquiry.

For instance, Windows hidden resampling isn't the only potential reason for some casual listeners not hearing the differences between 192/24 and 44/16. For instance, naively using Sox to play a sound file on my Mac:
iMac:OutV2 sergei$ play A-S192.wav
A-S192.wav:
File Size: 6.91M Bit Rate: 9.22M
Encoding: Signed PCM
Channels: 2 @ 24-bit
Samplerate: 192000Hz
Replaygain: off
Duration: 00:00:06.00
In:100% 00:00:06.00 [00:00:00.00] Out:265k [ -===|=====-] Hd:2.7 Clip:0
Done.

Looks OK, right? Until you try this (-V6 increases the verbosity of Sox informational messages):
iMac:OutV2 sergei$ play -V6 A-S192.wav
play DBUG sox: Looking for a default device: trying format `coreaudio'
play: SoX v14.4.2
time: Feb 22 2015 14:58:07
uname: Darwin iMac 18.6.0 Darwin Kernel Version 18.6.0: Thu Apr 25 23:16:27 PDT 2019; root:xnu-4903.261.4~2/RELEASE_X86_64 x86_64
compiler: gcc 4.2.1 Compatible Apple LLVM 6.0 (clang-600.0.56)
arch: 1288 48 88 L
play INFO formats: detected file format type `wav'
play DBUG wav: Searching for 66 6d 74 20
play DBUG wav: WAV Chunk fmt
play DBUG wav: Searching for 64 61 74 61
play DBUG wav: WAV Chunk data
play DBUG wav: Reading Wave file: Microsoft PCM format, 2 channels, 192000 samp/sec
play DBUG wav: 1152000 byte/sec, 6 block align, 24 bits/samp, 6912000 data bytes
play DBUG wav: 1152000 Samps/chans
play DBUG wav: Searching for 4c 49 53 54
play DBUG wav: WAV Chunk LIST
play DBUG wav: Type INFO
play DBUG wav: Attempting to seek beyond unsupported chunk `INAM' of length 2 bytes
play DBUG wav: Attempting to seek beyond unsupported chunk `id3 ' of length 48 bytes
Input File : 'A-S192.wav'
Channels : 2
Sample Rate : 192000
Precision : 24-bit
Duration : 00:00:06.00 = 1152000 samples ~ 450 CDDA sectors
File Size : 6.91M
Bit Rate : 9.22M
Sample Encoding: 24-bit Signed Integer PCM
Endian Type : little
Reverse Nibbles: no
Reverse Bits : no
play INFO formats: can't set sample rate 192000; using 44100
Output File : 'default' (coreaudio)
Channels : 2
Sample Rate : 44100
Precision : 25-bit
Duration : 00:00:26.12 = 1152000 samples = 1959.18 CDDA sectors
Sample Encoding: 32-bit Floating Point PCM
Endian Type : little
Reverse Nibbles: no
Reverse Bits : no
play DBUG effects: sox_add_effect: extending effects table, new size = 8
play DBUG effects_i_dsp: make_lpf(n=105 Fc=0.3850137 β=11.7712 ρ=0.5 dc-norm=0 scale=1)
play DBUG rate: fir_len=105 dft_length=1024 Fp=0.67625 Fs=1 Fn=2.17687 att=114.391 1/1
play DBUG effects_i_dsp: make_lpf(n=1763 Fc=0.005596301 β=11.2846 ρ=0.63 dc-norm=0 scale=147)
play DBUG rate: fir_len=12 phases=147 coef_interp=0 size=14.1k
play DBUG rate: 16|16 preload=16 remL=0
play DBUG rate: 0|0 preload=52 remL=0
play DBUG rate: 0|11 preload=5 remL=0
play DBUG rate: 16|16 preload=16 remL=0
play DBUG rate: 0|0 preload=52 remL=0
play DBUG rate: 0|11 preload=5 remL=0
play INFO sox: effects chain: input 192000Hz 2 channels (multi) 24 bits 00:00:06.00
play INFO sox: effects chain: rate 44100Hz 2 channels 32 bits 00:00:06.00
play INFO sox: effects chain: output 44100Hz 2 channels (multi) 25 bits 00:00:06.00
play DBUG sox: automatically entering interactive mode
play DBUG sox: start-up time = 0.035055
In:100% 00:00:06.00 [00:00:00.00] Out:265k [ -===|=====-] Hd:2.7 Clip:0
Done.

Holy cow! Unlike a professional DAW, which would make sure that the DAC is actually playing the file at 192/24, Sox just quietly resamples. Audacity behaves same way on my Mac. You need a DAC that indicates actual sampling rate and bit depth used at any given moment to catch this behavior.

How many people who say they could never hear difference between 192/24 and 44/16 fell into this trap? Especially on Windows: a laptop description could say that it supports 192/24, the Windows sound settings could show that 192/24 is engaged on the audio output, and yet Windows could resample anyway.
 

PierreV

Major Contributor
Forum Donor
Joined
Nov 6, 2018
Messages
1,437
Likes
4,686
Ah, boolean logic! Beautiful abstraction. Yet not always sufficient. To me, it is neither positive nor negative result. The experiment revealed important factors beyond the simple yes-or-no frame of inquiry.
.

Well, I'd rather avoid ignoring the sampling theorem and rejecting boolean logic in the same post (or even individually ;) ;) ).
Let's say that this is a protocol design issue then.

But yeah, I can accept that what you expect to receive isn't always what you actually receive in practical terms and that, yes, a DAC that explicitly displays what it gets and what it decodes remains useful in the context.
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,525
Likes
37,058
Ah, boolean logic! Beautiful abstraction. Yet not always sufficient. To me, it is neither positive nor negative result. The experiment revealed important factors beyond the simple yes-or-no frame of inquiry.

For instance, Windows hidden resampling isn't the only potential reason for some casual listeners not hearing the differences between 192/24 and 44/16. For instance, naively using Sox to play a sound file on my Mac:


Looks OK, right? Until you try this (-V6 increases the verbosity of Sox informational messages):


Holy cow! Unlike a professional DAW, which would make sure that the DAC is actually playing the file at 192/24, Sox just quietly resamples. Audacity behaves same way on my Mac. You need a DAC that indicates actual sampling rate and bit depth used at any given moment to catch this behavior.

How many people who say they could never hear difference between 192/24 and 44/16 fell into this trap? Especially on Windows: a laptop description could say that it supports 192/24, the Windows sound settings could show that 192/24 is engaged on the audio output, and yet Windows could resample anyway.
So now your rear guard action to defend the usefulness of hires is because Windows might resample it?

See you've sidestepped the issue of what benefit correctly done hires is. Until that is established the rest doesn't matter.
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,525
Likes
37,058
But there's a very interesting timing-related aspect to that: headphones cannot reproduce Blumlein stereo. Blumlein stereo provides no interaural timing difference because the mics are coincident, and it is the acoustic crosstalk to the 'wrong' ear from speakers mixed with the direct that creates a real, measurable timing difference at the ears.

Purist recordings that are optimised for speakers will use Blumlein stereo mic pairs while those for headphones will use spaced omnis of some kind (I presume) in order to encode a timing difference directly in the recording. And the ordinary panpot mimics the Blumlein pair, so it is optimal for speakers not headphones.

So you can't just swap between speakers and headphones without timing related side effects.

And headphones, of course, don't present the recording in a natural way: it is 'in your head'. I don't particularly mind that, but for the most natural portrayal of a typical recording I prefer speakers.

But I guess what you're arguing - and this would apply to the multiway speaker case - is that while the system plays fast and loose with 'timing' between channels and frequency bands, it is important to maintain the integrity of waveforms within, at least, their own channel, and the driver's own frequency band (although it is becoming a bit of a tenuous argument by that stage :)). Probably true, but I'm not personally worried about anything above CD sample rates.
You know Blumlein recordings over headphones don't sound too bad actually. They aren't the optimum as you say. ORTF would be better for phones.
 
OP
Sergei

Sergei

Senior Member
Forum Donor
Joined
Nov 20, 2018
Messages
361
Likes
272
Location
Palo Alto, CA, USA
Purist recordings that are optimised for speakers will use Blumlein stereo mic pairs while those for headphones will use spaced omnis of some kind (I presume) in order to encode a timing difference directly in the recording.

Good point. Both would benefit from a more contemporary recording technique. For instance, capturing waveforms at "generalized ears" for headphones, or sound-field in the vicinity of generalized listener's head with enough precision to later map it onto individual headphones and HRTFs. As for the speakers, properly mixed and mastered multi-channel sound is a huge improvement over stereo IMHO.
But I guess what you're arguing - and this would apply to the multiway speaker case - is that while the system plays fast and loose with 'timing' between channels and frequency bands, it is important to maintain the integrity of waveforms within, at least, their own channel, and the driver's own frequency band (although it is becoming a bit of a tenuous argument by that stage :)).

What you described is part of a sufficient condition (another part is putting the listener's head into a vice :)). Not easily achievable. What I'm after is necessary condition. Which, I argue, depends on specifics of the music and the listener. You see, some people here weren't impressed with my demonstration of 48/24 grossly smoothing out the transients present on 192/24. I agree that for this particular 6-second music fragment, and I guess for the vast majority of listeners, 48/24 is already sufficient.

For me though, the experiments demonstrated once again why gamelan sound may be so hard to capture on CD. Instead of one percussionist, gamelan orchestra has ~40, in total generating between 80 and 240 transients per second. When the transients generated by one percussionist are taken away, I perceive some of his hits on cymbals, sometimes, as being inaccurate - not in synch with guitarists. If gamelan transients are taken away, complex pieces of gamelan music turn into undifferentiated mush - a much more significant enjoyment-constraining effect.
 
Last edited:
OP
Sergei

Sergei

Senior Member
Forum Donor
Joined
Nov 20, 2018
Messages
361
Likes
272
Location
Palo Alto, CA, USA
So now your rear guard action to defend the usefulness of hires is because Windows might resample it?

To me, it epitomizes the confirmation bias: the thinking perhaps goes like this: since it is "widely known" and "scientifically proven" that hires audio is irrelevant, what harm could be done by resampling the hires material to a lower resolution, without telling the user?

I'm not defending the universal usefulness of hires in this instance, I'm pointing out that this could be one of the reasons for the rock-solid belief of millions of Windows users that hires audio doesn't matter: "I dowloaded it, I listened to it, there is absolutely no difference, this is all snake oil".

Samsung tried to pull a similar trick with its "4K" tablets. They didn't actually include a 4K-capable display, but rather a hardware+software subsystem able to resample 4K video to a lower resolution in real-time. This marketing move didn't work so well in the visual domain.

See you've sidestepped the issue of what benefit correctly done hires is. Until that is established the rest doesn't matter.

Sidestepped? Doesn't the experiment demonstrate that even if a studio has to release in, say, CD format, the quality of later-time resampling from 192/24 to lower resolutions, done with a multi-thousands-taps filters on a powerful CPU, is significantly better than the quality of real-time downsampling done in ADC, using filters with a handful of dozens of taps?

To me, such experiments demonstrated a qualitative difference. In the former case (filtering done in later-time), I couldn't hear transient distortions in the S192 null vs the CD version at all, at any amplification available to me. So, this criteria of the filtering quality likely is what back then was believed to be sufficient for absolute inaudibility of the transient distortions.

In the latter case (filtering done in real-time), the transient distortions were at about -30 dB SPL of peak at times. I could hear them in the null. I can only guess that the criteria of the acceptability of the real-time filtering was that the transient distortions are low enough to be masked by the signal: essentially a shortcut. Since the capability to capture in 96/24 and 192/24 was there for more complex signals, the shortcut was deemed harmless.

Well, the "masking is good enough" hypothesis may be true most of the time, on most of music materials, yet the burden of proving that it is universally true all the time, for all the music, and all listeners, is on those who insist on capturing everything in 48/24, even though 192/24 gear isn't much more expensive these days, and data storage is cheap and abundant.
 
OP
Sergei

Sergei

Senior Member
Forum Donor
Joined
Nov 20, 2018
Messages
361
Likes
272
Location
Palo Alto, CA, USA
So it was a clipped amplifier output signal and there are still no axis to the scope plot so we really don't know what the frequency was nor the amplitude.
Have you seen what the speaker will reproduce when fed with that amp output... something totally different !

No, I haven't seen it for the amp described in the article.
The goal of hifi reproduction is to reproduce that faithfully. So far you have not been able to show that even 44.1 can't reproduce the sound of the electric guitar coming from a speaker that is limited to around 4kHz.

I consider your statement fair. I'm going to follow up on this when I have an opportunity to either experiment with a similar vintage guitar amp, or to speak with a hands-on expert in this area.
 

Tks

Major Contributor
Joined
Apr 1, 2019
Messages
3,221
Likes
5,494
Maybe there is something going over my head utterly. But what are these mics used to record audio at 192kHz exactly? Like I can imagine a microphone recording ultrasonics, but not really a mic that does 1Hz to 1millionHz (or if it does, does it remotely well.

I just don't understand this thread (partially because it's over my head) secondly because of the seemingly puzzling point trying to be made by further puzzling diction deliberations over the meaning of words like "transients" and whatnot.
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
There are a few mics that go to 40-50k range. Also to 100kHz.
These are specialized mics usually for research or specific tasks etc. but they can be used for recording music.

You can also use normal mics (say with 20-25kHz bandwidth) and record with 192kHz. Higher frequencies can be 'described' about 4x more 'accurate' with respect to frequency/time domain.
During recording/mixing stage this has some advantages over recording/mixing at lower bitrates.

Aside from some overheads and drum mics quite a few mics are band limited even below 20kHz.
Other properties of purpose made mics are more important than FR extension and accuracy in some cases.
They could have a certain tone or overload properties or are lower noise for instance.

What one considers a transient is mostly under discussion.
What I get from J_J (one of the most knowledgeable members in this field at ASR) is that the first wavefront (its amplitude and steepness) is of importance for perception.
Research seems to show that the first wavefront is handled differently (and not per se in the frequency domain) by the brain.
What is debated about all of this (by the OP) is whether or not this can accurately be recorded and reproduced with 44.1 CD quality.
Most members are of the conviction anything above the audible range is moot so while recording at higher bitrates gives advantages what is questioned here is the final delivery for the masses.
CD quality is sufficient for 99% of the consumers.
The whole 'hires' camp often base their theories on what they read and heard (mostly under suspect conditions)
What makes this worse is marketing, different masters being used and DAC's using incorrect reconstruction filters that are 'said' to be better.

Listen for yourself... if you think you need hires in your life.. use it.
When you are perfectly happy with CD or MP3 its all good as well.
There is no need to break the bank when it comes to audio electronics and more than adequate performance.
Those that want 'audio jewelry' should buy it.
 

Rja4000

Major Contributor
Forum Donor
Joined
May 31, 2019
Messages
2,682
Likes
4,220
Location
Liège, Belgium
The funny thing with that discussion is that, as far as I know, a lot of the guitar effect pedals nowadays are applying some kind of dirty AD/DA conversion to the signal. In a typical board, at least one is doing that. And they've been build with little care of what happens outside of the frequency range of interest.
They are meant to run on limited-bandwidth loudspeaker cabinets, to begin with.
Even the expensive Kemper Profiler is said to be using internally 44.1kHz (24 bits).

So what's actually the point ?
That we can't keep the shit they produce outside of the hearing range if we record above 48kHz ?
Anyway, it seems wise to use a serious low pass filter when you record that.

I think -but that's indeed not very scientific- that most of the content above audible range is coming from those digital conversions and is, therefore, unwanted.

That, internally, the mixer or effects should use a higher bandwidth and bit rate to manipulate the signal, that's for me obvious, for engineering reasons. Not psychoacoustics reasons.
So that's probably why studios that can afford it are using 192kHz: to avoid resampling between equipments, and still allow serious signal manipulation with little audible effects.
And, maybe, also because it allows lower overall latency.
 
Last edited:

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,525
Likes
37,058
Maybe there is something going over my head utterly. But what are these mics used to record audio at 192kHz exactly? Like I can imagine a microphone recording ultrasonics, but not really a mic that does 1Hz to 1millionHz (or if it does, does it remotely well.

I just don't understand this thread (partially because it's over my head) secondly because of the seemingly puzzling point trying to be made by further puzzling diction deliberations over the meaning of words like "transients" and whatnot.
Lots of condenser mikes have some response to 30 khz or even a little higher. It is drooping quite a lot by then. Some few even have reasonably flat response to 40 or 50 khz. Which is why for any edge cases, 88.2 or 96 khz should be enough. Very nice response to 40 khz at least. There just is nothing worth worrying about above that, and you literally cannot hear it or any of its effects (except in poorly designed gear) anyway.

People who start talking super DSD or any PCM sampling rate above that are just talking garbage. Plain and simple. And any difference in 96 vs 44 is damned small. Would you really be turned off if music you like is recorded with 99% full possible perceptible fidelity instead of 100%? And it is a stretch to say it is even 1%. Maybe it is 1%, maybe it is 2% less than perfect. Probably is less than that. So we are truly down to a difference well into the weeds. Anyone who tells you the difference is big, large, night and day, that person is incredibly mistaken ( or lying to you). It really is very simple. If the difference were much of anything at all it would have been conclusively and with no doubt been shown by now. It hasn't, and the reason is any difference is somewhere between nothing, and not much of anything.
 

Cosmik

Major Contributor
Joined
Apr 24, 2016
Messages
3,075
Likes
2,180
Location
UK
Also, what about the effects of transducer intermodulation distortion in the ultrasonics region coming down into the audible band? If you hear a difference it isn't necessarily a 'real' difference.
 

Rja4000

Major Contributor
Forum Donor
Joined
May 31, 2019
Messages
2,682
Likes
4,220
Location
Liège, Belgium
Also, what about the effects of transducer intermodulation distortion in the ultrasonics region coming down into the audible band? If you hear a difference it isn't necessarily a 'real' difference.
I agree.
If there is an audible difference, there is a bigger chance that you'll hear problems caused by the higher bandwidth than any positive aspect.
 
OP
Sergei

Sergei

Senior Member
Forum Donor
Joined
Nov 20, 2018
Messages
361
Likes
272
Location
Palo Alto, CA, USA
Also, what about the effects of transducer intermodulation distortion in the ultrasonics region coming down into the audible band? If you hear a difference it isn't necessarily a 'real' difference.

I agree. A "professional" studio monitor I once evaluated comes to mind. It had attractively flat and extended frequency response. Decent price too. Boy, did it "sound different"! As far as I could tell, something with the equalization DSP wasn't quite right. Or maybe its designer also saved on tweeter amp's quality. Highs and transients were over-emphasized and distorted, resulting in "hot" and "brittle" sound (some would call it "revealing").

To me, the proper contribution of transients is more accurate approximation of a real-life performance. Like: the material the cymbals are made of is perceived in same or very similar way, not magically transformed into aluminum or tin. The percussionist's strikes are on time, not randomly time-shifted, which makes him sound like an amateur. Grand piano sounds like it is right there in my room. And so on.

Transients are not necessarily real-life ultrasonics. One of the best explanations I've seen in literature: on a high-resolution spectrogram of a sophisticated enough music fragment, you will see horizontal lines and vertical lines; the horizontal lines are slowly evolving sinusoids; the vertical lines are transients.

The vertical lines tend to spread over all frequencies in the hearing range, and are often seen extending into ultrasonics territory. This doesn't mean they actually contain ultrasonic sinusoids. This means that they are significantly more time-limited than the rest of the signal components, and thus they Fourier-transform into a wider spectrum.

For popular treatment of the wide spectrum characteristics of short-duration signals please see "Bandlimited versus timelimited" section in https://en.wikipedia.org/wiki/Bandlimiting. A more mathematically rigorous treatment can be found in https://pdfs.semanticscholar.org/859f/b7a193007e2622ae2821db6a20ea58ee7dbb.pdf
 
Last edited:
OP
Sergei

Sergei

Senior Member
Forum Donor
Joined
Nov 20, 2018
Messages
361
Likes
272
Location
Palo Alto, CA, USA
I'm going to follow up on this when I have an opportunity to either experiment with a similar vintage guitar amp, or to speak with a hands-on expert in this area.

Yesterday I spoke with United Audio Chief Scientist Dr. David Berners about the behavior of guitar amplifiers. He confirmed that the cone tighter suspension and surface oscillatory modes are significant non-linear characteristics of guitar amp speakers. In addition, shorter magnetic gap, damping the cone at high excursions because of the speaker coil getting out of the gap. He confirmed that studio monitors are designed to minimize these factors.

A new piece of information for me was how high the thermal compression of guitar amp speakers could be. Cold-coil to steady-state difference can be as large as 6 dB. Dr. Berners believes that most all guitar speakers use acoustically transparent grilles, so whistles and grille reflections are not significant factors. Neither he believes that bone fide ultrasonic radiation is a factor in steady state.

He was ambivalent about the reports of guitar amps speakers occasionally hitting hard stops. He attributed it to the behavior of cold speaker coils - that's why a sound check involving a guitar amp shall last for an extended period of time. United Audio developed capability to model the thermal compression in guitar amps, yet it was not in demand by their target market: I guess professionals do the sound check correctly.

Dr. Berners explained in length why only some of the United Audio plugins use algorithms based on the theory of Linear Time Invariant systems, yet most of them don't. Correspondingly, instead of using Fourier transforms or filters calculated based on LTI theory, most UA plugins iteratively solve systems of differential equations, in real time, to achieve faithful simulation of iconic non-linear analog gear.

He said that a constructive way to think about the shape of an acoustic signal that a guitar amp speaker emits is smoothed triangle or square wave. He also said that what they are concerned with is the signal after it is captured by a microphone. According to Will Shanks, product manager for UA plugins, their engineers sometimes consider 96 KHz sampling rate beneficial, yet not higher for the types of microphones they model.

To summarize: the experts don't believe that guitar amps emit bone fide ultrasonics. They are certain that the amps emit signals which are difficult to model as a sum of slowly evolving sinusoids. They believe that the sound of music instruments incorporating dynamic transducers can be adequately captured at the sampling rate of 96 KHz.
 

Frank Dernie

Master Contributor
Forum Donor
Joined
Mar 24, 2016
Messages
6,445
Likes
15,781
Location
Oxfordshire
How many people who say they could never hear difference between 192/24 and 44/16 fell into this trap? Especially on Windows: a laptop description could say that it supports 192/24, the Windows sound settings could show that 192/24 is engaged on the audio output, and yet Windows could resample anyway.
One of the reasons I compared a 24/96 file to a file downsampled to 44/16 then recomputed to 24/96 was that the second file could not have any of the "extra" data in it but would be handled identically by the DAC, however it treated a file.
I heard no difference.
The first time I heard a recorder where I could not detect any difference between the microphone feed and the recorder output was my first digital recorder, a StellaDAT, so 16/48. No tape recorder I used could do that.
 

Soniclife

Major Contributor
Forum Donor
Joined
Apr 13, 2017
Messages
4,500
Likes
5,417
Location
UK
That's a lot of words to say guitar amps are not anything like hifi amps on virtually all fronts. :)
And yet Hi-Fi amps and speakers can properly reproduce recordings of them, even from low rez vinyl.

This thread needs some extreme lo-fi guitar noise.
 
Top Bottom