• 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!

Please Help Me Establish Some Perspective!

DuaneInHiding

Member
Joined
Jan 8, 2020
Messages
8
Likes
21
I’ve been lurking and absorbing so much great information from this forum, but before I get carried away with making purchases based on an incomplete understanding I was hoping to ask some stupid questions (while showing my math) in the hopes that some of the generous people I’ve seen posting around here will set me straight.

For ease of argument, let’s say:
A. my listening area has a consistent ambient noise floor of 30 dB. (It’s very consistent, uniform and broadband…like a white/pink noise generator I play to help my cats sleep.)
B. I put on my open headphones (for sake of argument, so open that I’m hearing all 30 dB of the room around me along with the music) and adjust the volume so that the peaks are hitting my eardrums at 90 dB.

Question #1: Is it accurate to say that I will then be able to perceive 60 dB of dynamic range (correct term?) in my music — the distance between the loudest peaks and the point where the quietest parts are swallowed up by my room’s ambient noise?

Question #2: According to the review, a Topping D90 has a SINAD of about 120dB. (Don’t know if that’s the correct way to say that…) In the hypothetical situation spelled out in A and B above, would it be accurate to say that any noise introduced by the D90 would be 60dB below the ambient noise floor established by the combination of my listening environment and the playback volume of my music?

Question #2b: 60dB is a lot, right?!

Question #3: Let’s say I’m weird and I actually want to hear the noise introduced by a poorly-engineered DAC. Would I need to buy one so terrible that it has a SINAD of 59 dB before I could really hear its contributions to my music over my room’s ambient background noise?

Question #4: I decide to compare the Topping D90 to the off-brand CrappyDAC, which has a measured SINAD of 61 dB. Assuming my room only allows for 60 dB of range between the loudest peaks and the ambient noise, will the D90 effectively sound less noisy than the CrappyDAC? Or will they both perform well enough that their introduced noise will be masked below the ambient sound in the room to effectively the same degree?

Can anyone help shed some light on these questions?

Thanks in advance,
Duane
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,521
Likes
37,050
This looks like your first post. Welcome to ASR. Glad you are here Duane.

The 30 db SPL noise level is over the entire bandwidth up to 20 khz. Your ear is relatively insensitive to low frequencies, most sensitive around 3-5 khz and somewhat less sensitive at higher frequencies. The sensitivity of your ear follows the Fletcher-Munson curves.

Your ear breaks up the total 20 khz bandwidth into something like 30 or so bands so it gets below the raw total noise level. Here is something I hope helps. I measured my room noise with a calibrated microphone. This is 37 db SPL. It is such that 105 db SPL would be at the max 0 db level in this FFT. I'll try not to be confusing.

1578468930077.png

This is broken into 64 bands. Notice the low frequencies are rather high, but at say 3 khz and up it is low. Without complicating it too much, this would show that in my room with 37 db SPL noise levels, in the 3-5 khz range as my ear would hear it the noise floor is about 8-10 db SPL. So where my ear is most sensitive I only have 10 db noise levels. This kind of shape to the noise floor is more or less how it will be anywhere.

So in the 3-5 khz range in your room (quieter than mine) you likely could hear to right around 0 db SPL. With peaks up to 90 db SPL you would have 90 db of dynamic range. In frequencies below 100 hz, you'd have less, but your ear is 30 or more db less sensitive there anyway. Hopefully this helps a little with question #1.

Question #2B, though you have effectively more than 60 db, yes it is a lot. In fact I've tried to find recordings with more dynamic range than that, and the most I've found is 65 db. But that too is using a single number. It would be lower in our most sensitive region, best I can tell around that 90 db range. The Topping would have its noise floor well below your ambient noise if you have good gain staging. If you had a super-sensitive amp, and super-sensitive speakers and had to turn down the volume dramatically you might not get the full potential of the Topping dynamic range and low noise level.

Question #3: You are limited more by noise in recordings than the DAC. So one as poor as having 59 db dynamic range might be audible or maybe not. Depends on the spectrum of its noise. Also depends upon whether you've matched gain levels of other components to your DAC well or not.

Question 4 takes some delving into more deeply. For starters you need to remember SINAD is noise+distortion. You could have a low noise DAC with low distortion, but with lots of 60 hz hum say around -60 db. It would get a rating of -60 db, and yes, you might hear the hum. It might have low noise and low hum, but high 2nd and 3rd THD causing SINAD to be -60 db. You'll probably never hear it at all. Noise is low and with music -60 db THD is probably not audible especially if only the lower harmonics. It would likely sound very clean to you.

I suspect I've been confusing. So if the case, don't hesitate to ask questions of various parts of my post.
 
OP
D

DuaneInHiding

Member
Joined
Jan 8, 2020
Messages
8
Likes
21
First of all: Blumlein 88, thank you so much for your incredibly thorough and generous response! I genuinely appreciate the time it must have taken, and it has already contributed so much to my understanding of things.

I can see that I’ve stepped into really deep waters here, so I’ll start moving even slower:

I don’t have the equipment or confidence in my abilities to make such a precise measurement of my own room. So in the interest of keeping this fascinating conversation going, I would like to revisit some of my questions while assuming that we’re talking about Blumlein 88’s room instead of mine. (Based on some measurements taken this morning, it seems safe to assume that my room is noisier.)

Question #1, now answered! If this is true...

I measured my room noise with a calibrated microphone. This is 37 db SPL. It is such that 105 db SPL would be at the max 0 db level in this FFT.

...and this...

Without complicating it too much, this would show that in my room with 37 db SPL noise levels, in the 3-5 khz range as my ear would hear it the noise floor is about 8-10 db SPL.

...then it seems we are dealing with an effective dynamic range of ~ 95 dB in the most sensitive areas of human hearing, considered inside a real room. Trying to extrapolate that out into answers to my previous questions, I get this:

#2: looking at the Topping D90 review in more detail, I focus on the top-right quadrant and the spikes between 3-6k in order to make sure I’m concentrating on the noise that “matters” given where our hearing is most sensitive. In that case, the highest spikes are below -130dB. Given these numbers, the noise the D90 will be contributing to any recording I play through it will be ~35 dB below the ambient noise of the room, and therefore effectively masked.

#3: assuming my hypothetical CrappyDAC had a FFT readout with the same shape as the D90 only with the noise spikes much higher,I assume that those spikes would need to exceed -95 dB before the CrappyDAC was contributing any audible noise in my given listening situation.

I’ll skip #4 for now, because this all leads me to one important new question — probably the big one: based on the collective wisdom represented on this board, am I thinking about these things in the right way? Or is this whole line of inquiry missing some bigger point?

For background, I should explain that I’m a musician who does a lot of home recording and mixing. In studying mastered commercial releases in order to improve my own mixes, I see very little effective dynamic range. For example, a recent rock song I like…

“Lazarus,” by David Bowie:
Peak: 0 dBFS
Average loudness: -8 dBFS
Dynamic range, excluding loudest 5% and quietest 10%: 5.8 dB

That’s very compressed! But I mostly only listen to classical music at night with headphones, so let’s take the first movement of Mozart’s Requiem (Karajan for Deutsche Grammophon in the ‘70s):

Peak: -0.6 dBFS
Average loudness: -16.1 dBFS
Dynamic range, excluding loudest 5% and quietest 10%: 18.5 dB

With those numbers, is it safe to say that even though Blumlein’s room allows for 95 dB of range between absolute peaks and the noise floor at 3-5k, even the Mozart recording is only using about 18.5 dB of it: assuming I set its peak at 105 dB, then the music would be happening between 86.5 dB and 104.5 dB — in other words, more than 75 dB above the ambient noise floor in our most sensitive area of hearing. (And even higher than the noise generated by something like the D90 in those frequencies.)

Hence my obsession so far with ambient noise floor: I keep seeing that a component with a noise floor below -120 dB is “provably transparent,” but that seems far, far below the real-world conditions represented by average recordings played in even Blumlein’s quiet listening environment.

Or…am I thinking about all of this the wrong way?

Again, I want to say thank you for the time you all are spending on a newcomer’s posts. I’m sorry to be writing novels right away, and I hope that doesn’t come across as rude — I'm just excited by these concepts, and I genuinely appreciate any wisdom anyone has time to share.

Best,
Duane
 

Hipper

Addicted to Fun and Learning
Joined
Jun 16, 2019
Messages
753
Likes
625
Location
Herts., England
Duane, do you actually know the ambient noise in your room?

It can be measured using a free software called Room EQ Wizard (REW), plus you need a microphone, stand and cable, and a laptop. REW is complicated but there's plenty of help here and elsewhere.
 
OP
D

DuaneInHiding

Member
Joined
Jan 8, 2020
Messages
8
Likes
21
Duane, do you actually know the ambient noise in your room?

It can be measured using a free software called Room EQ Wizard (REW), plus you need a microphone, stand and cable, and a laptop. REW is complicated but there's plenty of help here and elsewhere.

Hey Hipper, thanks for your reply!

I’ve used REW extensively when working on the sound treatment of my office/“studio” - used the waterfall plots to figure out which frequencies were ringing, placing treatment so that the mixing position has an even decay, etc. Unfortunately the measurement microphone I used for that has since broken, and there isn’t an easy way for me to use my more accurate recording equipment in the bedroom where I mostly listen to my headphone rig.

I’m not particularly worried about establishing the specific noise floor of my listening space, so much as I’m interested in figuring out…I guess it’s time to figure out what my real, actual question is. I think it’s this:

I have a Fiio M11 (non-Pro version), which I mostly listen to with a pair of Sennheiser HD580s through the M11’s balanced output. To a novice like me, Fiio’s stated specs for the M11 seem pretty great: its dual AK4493 DACs running through the balanced output produces a reported THD+N of less than .002%, more than 88.5mW of output into 300 ohms at under 1%THD+N, an SNR of under 118 (A weighted, though).

There aren’t any ASR tests for this device, though, and I’m naturally a little skeptical about manufacturer specs. Elsewhere on this site I’ve been reading these remarkable, clear, explicit tests of what a person can buy for ~$500 dollars now — say a Topping D50s into an SMSL SP200 — and I can’t help but start dreaming about that rig and relegating the M11 to use as a digital transport when I’m at home.

But then…I take a step back, and I begin to wonder what the actual difference would be in a real room, when considering the actual dynamic range of the recordings I regularly play. That’s why I’m so hung up on ambient noise — in my bedroom, will the difference between the M11 I have and the D50s/SP200 I want even be detectable?

That’s why I’m stumbling around with these definitions of “dynamic range” and all that — metaphorically, am I agonizing over a few centimeters in difference when both things are kilometers too far away for me to see clearly anyway?

The bottom line for me is that this is all super interesting, and the pursuit of the limits of precision for $X is just pure fun — it’s no different than people hot-rodding their cars. So I’m not so much interested in “is a lower SINAD really worth the money” because this is the hobby I love, so my disposable money will probably flow in this direction anyway… The question is more if a DAC/amp upgrade will give me as many giggles as a new pair of headphones, considering the (unmeasured) quality level I’m already enjoying right now…
 

garbulky

Major Contributor
Joined
Feb 14, 2018
Messages
1,510
Likes
827
Let me make it simple. Theoretically, the worst performing dacs and the best performing dacs should sound THE SAME. (Theoretically.) Do you find that to be true or no?
 
OP
D

DuaneInHiding

Member
Joined
Jan 8, 2020
Messages
8
Likes
21
Let me make it simple. Theoretically, the worst performing dacs and the best performing dacs should sound THE SAME. (Theoretically.) Do you find that to be true or no?

Please, make it simple!

Today, the only honest answer I can give you your question is that I lack the experience to know what percentage of a song’s sonic signature is being contributed by the DAC.

The same could also be said for every other link in the chain — amp, cables, etc. except for two: my experience tells me that the source file and the speakers/headphones/IEM are more important than the rest. Between those two points I can recognize subjective differences in entire systems (“man, violin’s sound bad in my wife’s car”) but I don’t feel capable of saying “her car’s stereo sounds bad because it has a great DAC being wasted on a bad amp,” versus the opposite.

So if you tell me that the worst-ranked DAC measured on this site sounds exactly like the current best, then…I’m certainly surprised, but also humble enough to admit that I don’t know anything about this stuff yet.

I'm sure some of you can recognize what I'm feeling right now: I sit and listen to my M11 through my HD580s and it sounds glorious. Absolutely the best listening experience I've ever had. But according to this tests on this site, there are amps/DACs/whatever that might make it sound EVEN BETTER...and having not heard 'even better' for myself, I can't say for sure whether the gap between where I am and where I could be is big, small, detectable, purely theoretical...
 

garbulky

Major Contributor
Joined
Feb 14, 2018
Messages
1,510
Likes
827
Please, make it simple!

Today, the only honest answer I can give you your question is that I lack the experience to know what percentage of a song’s sonic signature is being contributed by the DAC.

The same could also be said for every other link in the chain — amp, cables, etc. except for two: my experience tells me that the source file and the speakers/headphones/IEM are more important than the rest. Between those two points I can recognize subjective differences in entire systems (“man, violin’s sound bad in my wife’s car”) but I don’t feel capable of saying “her car’s stereo sounds bad because it has a great DAC being wasted on a bad amp,” versus the opposite.

So if you tell me that the worst-ranked DAC measured on this site sounds exactly like the current best, then…I’m certainly surprised, but also humble enough to admit that I don’t know anything about this stuff yet.

I'm sure some of you can recognize what I'm feeling right now: I sit and listen to my M11 through my HD580s and it sounds glorious. Absolutely the best listening experience I've ever had. But according to this tests on this site, there are amps/DACs/whatever that might make it sound EVEN BETTER...and having not heard 'even better' for myself, I can't say for sure whether the gap between where I am and where I could be is big, small, detectable, purely theoretical...
I’m not saying they sound the same at all. I find myself hearing differences where double blind tests and even measurements tell me there aren’t any. I go by what I hear during regular use. take the Oppo 105 it had some great measurements. My dc1 has worse measurements but both species to inaudible levels of
Distortion. Nevertheless I thought the Oppo sounded too bright and thin and I happily preferred the dc-1 dac. you couldn’t pay me to use the oppo.


Most people here use double blind testing as what to go by instead. If you go by double blind testing, you’ll find that you would be hard pressed to reliably identify any differences between most electronics including that of your phone or laptop output and the worlds best dac. Even if by some off chance you did identify a difference in the test it would be so small and trivial that you may think it’s not even mentioning. Does that mean you imagined all those big differences ? That’s not for me to say. That’s for you to decide.
 

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,198
Likes
16,981
Location
Riverview FL
That’s why I’m stumbling around with these definitions of “dynamic range” and all that — metaphorically, am I agonizing over a few centimeters in difference when both things are kilometers too far away for me to see clearly anyway?

You mentioned "perpsective...

Ponder this...

https://www.audiosciencereview.com/...tellar-gain-cell-dac.9273/page-29#post-273791

Or this:

https://www.audiosciencereview.com/forum/index.php?threads/the-shoutometer.2555/
 
Last edited:

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,521
Likes
37,050
Well remember up thread where Amir said he had a reel to reel machine with 70 db SNR and it was an audible amount of noise. I hope I'm not putting words in his mouth. If you were to play just the noise at that level you most likely would hear it. But I also said the best recordings I'd found only had dynamic range around 65 db. Well, if you play music on that reel to reel, especially modern recordings of music you likely won't hear any noise as the music will mask it. If they haven't ramped down the noise in silent portions before the music starts you might catch the noise for second or two.

I don't think the files are up now, I posted some once where I added different amounts of noise to recordings asking people at what level they could no longer hear it at normal listening volumes. The average for most people was around -75 db noise. Considering normal listnening levels are around 75-80 db SPL average with a bit higher peaks for most people and noise in room that makes sense.

Looks like the Fiio is quiet enough and low enough in THD you are unlikely to benefit going to another device. Or if there is a difference it would be very small. This is assuming the FR and distortion connected to your other gear is fine. Which it probably is.

In short DACs are generally a solved problem. Though some of Amir's tests show people can still mess it up sometimes.

You are correct transducers are the worst bottlenecks in audio. Speakers and microphones. And more so due to over-processing the recordings. Most other parts can be a non-problem.
 

MZKM

Major Contributor
Forum Donor
Joined
Dec 1, 2018
Messages
4,240
Likes
11,462
Location
Land O’ Lakes, FL
I’ve used REW extensively when working on the sound treatment of my office/“studio” - used the waterfall plots to figure out which frequencies were ringing, placing treatment so that the mixing position has an even decay, etc.
Heads up, Waterfall plots are not the best tool as the decay is in reference to initial SPL, so if you have a room mode causing a peak at a certain frequency, a waterfall plot would make it look like it is has a long decay, which isn’t what’s happening and a simple EQ adjustment would fix it.
Video quickly talking on this (time stamped, but skip to 23min for waterfall):
 
OP
D

DuaneInHiding

Member
Joined
Jan 8, 2020
Messages
8
Likes
21

Oh man...these two posts are EXACTLY what my befuddled brain needed. Thank you so much! (And did you also draw that Everest vs. sand castle illustration? Because I like the way your brain works...) I want to quote something I read once about how ill-suited our primate brains are to conceptualizing very large or very small numbers...and I think I got a double helping of whatever defective gene causes that. Simple, clear comparisons like "60 dB down is like your house sitting next to Mt. Everest" or "it's like somebody shouting at you from a kilometer away" helps me sneak up to these ratios in a way that log displays can't. (A flaw in me, not the charts.)

I'm very comfortable with overkill; I have a 100 watt Marshall halfstack sitting beside me in my 10x12 office. I know it's too much power for the space, too much quality for my poor playing, too big and excessive for the situation I'm in. But I've been playing guitar for 35 years, I've owned a dozen amps and played hundreds... I understand exactly how much amp I really need, and therefore I can blow some money and incur some unnecessary hassle to live with a thing that just makes me smile every time I look at it.

In audiophile world, though, I feel like a kid wandering around in a Guitar Center with a handful of money and absolutely no clue where to start. Even if I managed to land on exactly the same amp I'm enjoying today, I'd still wonder if I'd made the right choice -- "Could I have gotten by with less power? Can I even hear the difference between this one and any of the others?"

I'll probably still end up splashing out for a rocketship DAC/amp combo in that same spirit, but I think I want to really understand just how crazy over-spec'ed it is before I do. Ray, your posts brought me a lot closer to that goal -- thank you!
 
OP
D

DuaneInHiding

Member
Joined
Jan 8, 2020
Messages
8
Likes
21
I’m not saying they sound the same at all. I find myself hearing differences where double blind tests and even measurements tell me there aren’t any. I go by what I hear during regular use. take the Oppo 105 it had some great measurements. My dc1 has worse measurements but both species to inaudible levels of
Distortion. Nevertheless I thought the Oppo sounded too bright and thin and I happily preferred the dc-1 dac. you couldn’t pay me to use the oppo.


Most people here use double blind testing as what to go by instead. If you go by double blind testing, you’ll find that you would be hard pressed to reliably identify any differences between most electronics including that of your phone or laptop output and the worlds best dac. Even if by some off chance you did identify a difference in the test it would be so small and trivial that you may think it’s not even mentioning. Does that mean you imagined all those big differences ? That’s not for me to say. That’s for you to decide.

Thanks, I think I get what you're saying. On the music recording side of things, I struggle with this too: I can tell you exactly why I prefer one mic preamp over another for XYZ reason, but I've also left the wrong device patched in and not realized it more than once... Even knowing that, I wouldn't get rid of the gear I love because SOMETHING helped form those opinions over years of use. Fascinating.
 
OP
D

DuaneInHiding

Member
Joined
Jan 8, 2020
Messages
8
Likes
21
In short DACs are generally a solved problem. Though some of Amir's tests show people can still mess it up sometimes.

Thanks for that -- my listening experiences with the M11 definitely suggest that they got it right, so I'll probably start directing my search for audio grails in another direction. It's crazy how something as utilitarian as a DAC can become so fascinating...
 

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,317
Location
Albany Western Australia
Heads up, Waterfall plots are not the best tool as the decay is in reference to initial SPL, so if you have a room mode causing a peak at a certain frequency, a waterfall plot would make it look like it is has a long decay, which isn’t what’s happening and a simple EQ adjustment would fix it.
Video quickly talking on this (time stamped, but skip to 23min for waterfall):

Im not sure thats quite right :)

The long decay appears to be really happening. What he says in the video is that there is not much point in examining the dacay because if a peak in FR occurs at time 0 then it will be occurring at any further time X. Yes this is correct, and yes it will be corrected by eq. However there can also be peaks in the FR which dont have long decays which seems to sort of prove the point that the long decay is really happening. Maybe just plain speaker frequency response peaks. So the waterfall is still useful for sorting out what is going on, even though the solution (eq) might be the same.


Peak and long decay (room mode) at 46Hz. Peak at 134Hz doesnt have long decay even though close in amplitude.
1578552173242.png

1578552509268.png


Yes the peak at 46 Hz is about 5dB louder than the peak at 134Hz, but as you can see it takes 4 times as long to fall by the same amplitude. If it was just because it is louder it would fall at the same rate.

I do agrre with the guy that the wavelet display is very useful.
 
Last edited:
Top Bottom