Dynamic Range and Noise Floor

[dlaloum]

With dither there definitely is.

There we artificially create a higher noise floor....

 

[Blumlein 88]

There we artificially create a higher noise floor....

But after dithering you can have signals below the original hard lowest level.

 

[Cbdb2]

Digital zero is zero sound including noise. The noise comes with the analog.
In the digital world you can add thousands of tracks of zero level and still end up with no noise. Try that with tape.

 

[Cbdb2]

I read that:

" In its purest form, dynamic range in music production is simply the decibel (dB) difference between quietest and loudest sounds in a mix or audio file. Therefore, songs with a wide dynamic range will have a larger gap between the loudest sound and quiet sound compared to songs with more consistent sound quality.

Dynamic range can also refer to the loudest and quietest sounds hardware or audio systems can properly represent. The bottom range of this value is called the noise floor."

If you want to listen to the quietest sound like the sound of silence. Do you need high dynamic range? Or Is it possible to have quietest sound only without regards to loudest sound? Meaning is quietest sound always coupled to dynamic range or can the two be decoupled? How?

Theres a big difference between musical and hardware dynamic range, this relates to hardware.
How do you listen to silence? Even if your system (and room) has zero noise you still need a minimum amount of signal to hear it. This is the lower end of your gears dynamic range. The high end is your system full out. So yes its a range between lowest and highest levels.

 

[Art of sound]

I've been looking at real world dynamic range of music. most of hans zimmer's has 10-11 db and given that most movie and music soundtracks have around 21db clearance (99.9%) are we looking at raw engineering performance that doesn't translate into real world gain?

Album list - Dynamic Range DB

 

[MaxwellsEq]

I've been looking at real world dynamic range of music. most of hans zimmer's has 10-11 db and given that most movie and music soundtracks have around 21db clearance (99.9%) are we looking at raw engineering performance that doesn't translate into real world gain?

Album list - Dynamic Range DB

You should read posts 8 and 9 in this thread to understand why numbers like "10-11 db" are meaningless. I have pieces of music that start with a slow fade covering very wide dynamic ranges, but if I feed them through popular tools for "measuring music dynamic range", they give a figure of 12, which is clearly NOT a measure between the quietest sounds and the loudest sounds. Everyone here should be able to run a level meter in software that shows slow fades over 10s of dB.

The tools also can't account for frequency windows - I can perceive a very quiet piccolo being played, even though the double-basses are much louder. The dynamic range meter only detects the peak thing.

 

[Art of sound]

You should read posts 8 and 9 in this thread to understand why numbers like "10-11 db" are meaningless. I have pieces of music that start with a slow fade covering very wide dynamic ranges, but if I feed them through popular tools for "measuring music dynamic range", they give a figure of 12, which is clearly NOT a measure between the quietest sounds and the loudest sounds. Everyone here should be able to run a level meter in software that shows slow fades over 10s of dB.

The tools also can't account for frequency windows - I can perceive a very quiet piccolo being played, even though the double-basses are much louder. The dynamic range meter only detects the peak thing.

part of what you're saying is true. but i wouldn't go so far as to need the 16+ bits of headroom always to get better separation. maybe
Is there a study done specifically showing preference correcting for DR? or multitone distortion? and default attenuation filter ? we might need that data before making use of ASR's data but i'm glad its here.

On another note. More expensive hardware (iem, headphones) are usually far better with timbre and instrument separation given my 2 year experience . yet to purchase expensive speakers but i've heard expensive car audio tweeters that can be differentiated fairly easily on timbre itself if not for instrument separation and vocals (B&W on volvo and BMW). Don't own an expensive DAC or amp but the differences in frequency range is audible on the various products i bought over the 2 years into this hobby and all the reviewers are right about FR being the most important factor.

To add an example, I've owned hidizs s9 pro and the sound was tad bright and congested and when paired with hd560s and other warm sounding iems and akg k361 etc it was okay but i could clearly hear the difference is dynamics from a desktop dac (monoprice cavalli liquid spark dac) and topping a50 including better instrument separation and more details along with some minor FR difference ( all these devices scored high marks on most of the metrics tested at ASR)

 

[MaxwellsEq]

part of what you're saying is true. but i wouldn't go so far as to need the 16+ bits of headroom always to get better separation. maybe
Is there a study done specifically showing preference correcting for DR? or multitone distortion? and default attenuation filter ? we might need that data before making use of ASR's data but i'm glad its here.

On another note. More expensive hardware (iem, headphones) are usually far better with timbre and instrument separation given my 2 year experience . yet to purchase expensive speakers but i've heard expensive car audio tweeters that can be differentiated fairly easily on timbre itself if not for instrument separation and vocals (B&W on volvo and BMW). Don't own an expensive DAC or amp but the differences in frequency range is audible on the various products i bought over the 2 years into this hobby and all the reviewers are right about FR being the most important factor.

To add an example, I've owned hidizs s9 pro and the sound was tad bright and congested and when paired with hd560s and other warm sounding iems and akg k361 etc it was okay but i could clearly hear the difference is dynamics from a desktop dac (monoprice cavalli liquid spark dac) and topping a50 including better instrument separation and more details along with some minor FR difference ( all these devices scored high marks on most of the metrics tested at ASR)

I'm struggling to understand you and I don't think I understand the points you are making. This may be a language issue.

I'll break it down a bit:

part of what you're saying is true.

Thank you. Which part of what I say is true and which part is a lie?

but i wouldn't go so far as to need the 16+ bits of headroom

"Headroom" is generally taken to mean unused capacity that you could use, but don't need to. So you might record a piano with 6dB of headroom, just in case there are peaks faster than your meter, or the pianist hits the keys harder than during rehearsal. 6dB is roughly 1 bit. 16 bits of headroom would give us approximately 96db of headroom which is not necessary.

16+ bits of headroom always to get better separation. maybe

Separation (in two channel systems) usually refers to how much leakage there is between left and right channels - the better the separation, the lower the leakage. There are no "bits" associated with separation. I think you are throwing "HiFi" terms around out of context

Is there a study done specifically showing preference correcting for DR?

What do you mean by DR? This thread has established "DR" is a vague term. Why would anyone measure a "preference" for something this vague? Are you perhaps suggesting the average listener would chose music that is highly compressed over something that has an enormous swing from almost complete silence to shattering crescendo? We know that people are often more impressed by compressed sounds. Also if the full 16 bits were used for music, most of us would have to turn the volume up to hear the quiet bits because they would be below the noise floor of our listening rooms

Is there a study done specifically showing preference correcting for ... ... multitone distortion?

What is "multitone distortion"?

Is there a study done specifically showing preference correcting for ... ... default attenuation filter ?

I'm not sure what a "default attenuation filter" is. A filter passes through a range of frequencies, as a low-pass; band-pass; or high-pass filter - whilst attenuating other frequencies. There is no "default attenuation filter" - unless perhaps you mean RIAA time constants

More expensive hardware (iem, headphones) are usually far better with timbre and instrument separation given my 2 year experience

I'm pleased that you have amassed some experience and hopefully intend to keep learning. In general, I think you are right: as price or cost is increased, designers are freed from being bound by constraints that lower budgets enforce. This doesn't mean to say that expensive is necessarily always better. It is possible to dress a DAC in diamonds, but there would be no improvement.

Don't own an expensive DAC or amp but the differences in frequency range is audible on the various products i bought over the 2 years into this hobby and all the reviewers are right about FR being the most important factor.

I would expect well executed DACs and amplifiers operating in their designed range to not exhibit audible variations in frequency response.

To add an example, I've owned hidizs s9 pro and the sound was tad bright and congested and when paired with hd560s and other warm sounding iems and akg k361 etc

Transducers do sound different from each other because there are always compromises in design and construction. For loudspeakers, quite a lot of what we are hearing is the room that we are listening in. For headphones and IEMs, our external and internal ears and head shape vary from person to person, so I would expect some people to not get on with an IEM that sounds fine to someone else. For example, I wear glasses and so getting a reliable seal with on and over-ear headphones is trickier than for people with better sight.

 

[Art of sound]

I'm struggling to understand you and I don't think I understand the points you are making. This may be a language issue.

I'll break it down a bit:

Thank you. Which part of what I say is true and which part is a lie?


"Headroom" is generally taken to mean unused capacity that you could use, but don't need to. So you might record a piano with 6dB of headroom, just in case there are peaks faster than your meter, or the pianist hits the keys harder than during rehearsal. 6dB is roughly 1 bit. 16 bits of headroom would give us approximately 96db of headroom which is not necessary.


Separation (in two channel systems) usually refers to how much leakage there is between left and right channels - the better the separation, the lower the leakage. There are no "bits" associated with separation. I think you are throwing "HiFi" terms around out of context


What do you mean by DR? This thread has established "DR" is a vague term. Why would anyone measure a "preference" for something this vague? Are you perhaps suggesting the average listener would chose music that is highly compressed over something that has an enormous swing from almost complete silence to shattering crescendo? We know that people are often more impressed by compressed sounds. Also if the full 16 bits were used for music, most of us would have to turn the volume up to hear the quiet bits because they would be below the noise floor of our listening rooms


What is "multitone distortion"?


I'm not sure what a "default attenuation filter" is. A filter passes through a range of frequencies, as a low-pass; band-pass; or high-pass filter - whilst attenuating other frequencies. There is no "default attenuation filter" - unless perhaps you mean RIAA time constants


I'm pleased that you have amassed some experience and hopefully intend to keep learning. In general, I think you are right: as price or cost is increased, designers are freed from being bound by constraints that lower budgets enforce. This doesn't mean to say that expensive is necessarily always better. It is possible to dress a DAC in diamonds, but there would be no improvement.


I would expect well executed DACs and amplifiers operating in their designed range to not exhibit audible variations in frequency response.


Transducers do sound different from each other because there are always compromises in design and construction. For loudspeakers, quite a lot of what we are hearing is the room that we are listening in. For headphones and IEMs, our external and internal ears and head shape vary from person to person, so I would expect some people to not get on with an IEM that sounds fine to someone else. For example, I wear glasses and so getting a reliable seal with on and over-ear headphones is trickier than for people with better sight.

I do agree with you that DR of 16 bits not being necessary for a DAC to be considered as well engineered.
multitone distortion is a better indicator for audibility of distortion. By attentuation filter i meant nyquist reconstruction filter.

most humans are worried about the music rather than the quality and i wouldn’t blame them but once you show them that they can get better quality they will identify it as better as in obviously vocals/instruments get clearer and more discernable when you improve quality of source compression itself but beyond a point it becomes less obvious and sound is subjective just like other forms of perception it likely varies a lot amongst humans. i’ve explained some basic principles to many people and if it doesn’t involve behavioral change they will entertain and adapt to the changes as long as the software/products within their reach and provides similar experience but they won’t jump through hoops for that quality so goes with anything about us.

’’Human perception is so layered and complex that our brains are constantly responding to stimuli that aren’t real in any physical, concrete sense, but are just as vitally important: our thoughts…….In other words, ideas and thoughts and concepts have lives inside us.”
“The Rubber-Hand Illusion is what’s called a body transfer, but our brain’s way of processing reality… rather than giving it to us directly… also opens us up to other slightly freaky “mix-ups” of the senses. Researchers have shown, for example, that we are capable of hearing “phantom words.” When listening to nonsense sounds, our brains pick out clear-seeming words that aren’t actually there in the audio. There is also the “Barber’s Shop Illusion,” in which a recording of snipping scissors gives the impression of the sound getting closer or farther depending on the volume rising or lowering, when the sound isn’t changing position at all. Or think of the common experience of sitting in a stationary car or plane, and when the car or plane next to you begins to move, at first you think you are the one in motion. There are many, many more phenomena like these.”
-deviate by beau lotto


Good music matters a lot and i mean not just mastering. good music is only a tiny fraction of the 100 million songs and its getting more muddier with technology and easy of existing samples and use of machine learning in production.
In my experience audiophile community is not better at picking good music vs regular people and maybe skewed to the latter being better as the numbers are larger there.

Humans end up rewarding creativity so the very few get most of the money as goes with all forms of creativity it follows a power law distribution. Rightfully so

 

[MaxwellsEq]

You avoided answering most of my questions. Then came up with some questionable statements.

I do agree with you that DR of 16 bits not being necessary for a DAC to be considered as well engineered.

I don't think you can agree with me, since I've never stated that at all. Clearly, a DAC MUST have 16 bits if it's converting Red Book CDs into audio, since that's what's on the disc. But I also believe that 16 bits is adequate for ensuring the noise floor is sufficiently low

multitone distortion is a better indicator for audibility of distortion.

You have still not explained what multitone distortion is. It does not sound like something that any of us want - we want all our tones undistorted.

most humans are worried about the music ... ... won’t jump through hoops for that quality so goes with anything about us.

OK I get what you are saying. Many people (perhaps a majority of the world) enjoy music, but can get pleasure from a small radio. If they hear it better reproduced, they can enjoy that as well, but don't care enough to spend money or time on it.

’’Human perception is so layered ... ...There are many, many more phenomena like these.”
-deviate by beau lotto

It's ironic that you quote something on perception. I certainly feel you would benefit by also studying Daniel Kahneman's "Thinking Fast and Slow" and study heuristics and biases

Good music matters a lot and i mean not just mastering. good music is only a tiny fraction of the 100 million songs and its getting more muddier with technology and easy of existing samples and use of machine learning in production.

Depending on your point of view, and sampling across the population, I would argue that good music makes up the majority of the 100 million songs, but not for any of us, there is some music we don't like. If you are talking about sound production, I think things are steadily improving. When I started in studios, everything used analogue tape, which has significant limitations, modern ADCs and computers are leagues ahead in terms of capability to capture an event

In my experience audiophile community is not better at picking good music vs regular people and maybe skewed to the latter being better as the numbers are larger there.

Really? I think that's rather a sweeping statement.

Humans end up rewarding creativity so the very few get most of the money as goes with all forms of creativity it follows a power law distribution. Rightfully so

Do you know what a power law is? Why do you think creativity follows a power law? Next up, you need to study Network Effects and also basic economics to understand why money is distributed the way it is.

 

[Art of sound]

You avoided answering most of my questions. Then came up with some questionable statements.


I don't think you can agree with me, since I've never stated that at all. Clearly, a DAC MUST have 16 bits if it's converting Red Book CDs into audio, since that's what's on the disc. But I also believe that 16 bits is adequate for ensuring the noise floor is sufficiently low


You have still not explained what multitone distortion is. It does not sound like something that any of us want - we want all our tones undistorted.


OK I get what you are saying. Many people (perhaps a majority of the world) enjoy music, but can get pleasure from a small radio. If they hear it better reproduced, they can enjoy that as well, but don't care enough to spend money or time on it.


It's ironic that you quote something on perception. I certainly feel you would benefit by also studying Daniel Kahneman's "Thinking Fast and Slow" and study heuristics and biases


Depending on your point of view, and sampling across the population, I would argue that good music makes up the majority of the 100 million songs, but not for any of us, there is some music we don't like. If you are talking about sound production, I think things are steadily improving. When I started in studios, everything used analogue tape, which has significant limitations, modern ADCs and computers are leagues ahead in terms of capability to capture an event


Really? I think that's rather a sweeping statement.


Do you know what a power law is? Why do you think creativity follows a power law? Next up, you need to study Network Effects and also basic economics to understand why money is distributed the way it is.

multitone distortion tests that amir does for each and every device correlate highly with SINAD but the opposite is less true. this is subjective though i haven't compared the numbers but i've seen this pattern many times.

I have read 'Thinking fast and slow" but i prefer Thaler's 'nudge' in behavioral economics. in any case this is out of topic. I'm not making any sweeping statements. good music as defined generally and not just well mastered. human perception is mostly the brain and very less dependent on sensory perception which by itself varies so much in humans. example: we have 100k-1million taste buds range. flavor needs more things like smell (ortho and retro), how it looks combined with the ambience collectively give us the impression of food. just imagine the variation of perception but we can definitely measure reviews of restaurants and have a metric. BTW i am very picky with food as well as music but i have no measurements to prove this other than being very selective.

everything on creativity is out of topic and it is my theory so my bad.