• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). There are daily reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Understanding THD+N Charts

Hiten

Active Member
Joined
Nov 24, 2019
Messages
291
Likes
359
Location
India
Hi all,
trying to understand thd+n charts. Pl. take a look at following. (Charts are from audio poland website of Denon and Emotiva amplifiers) Am curious why some charts starts (0 or 0.1 Pout) higher and some lower. Is it because of Noise measurements are combined with THD measuremens ? what contributes more at such low level circuit Noise, layout etc. or distortion ?
sample 1.jpg
sample 2.jpg

as mostly we listen at low 1 watt level and if I suppose when speakers are highly sensitive the less the thd+N (lower as in left chart) better ?
thanks and regards.
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
17,163
Likes
29,884
It is because of noise. THD+N obviously includes noise. More than likely the THD part is the same percentage or lower as at the higher power levels. Noise might be constant and therefore at lower levels is a larger part of THD+N. These remember are ratios. So same noise in terms of watts is a larger percentage at low powers and a lesser percentage at higher powers.
 

DonH56

Master Contributor
Technical Expert
Forum Donor
Joined
Mar 15, 2016
Messages
6,929
Likes
13,688
Location
Monument, CO
I wrote a long thread that has some examples, but it is not an easy read.... It may be easier to skip through at first and check some of the plots and examples.

 
OP
Hiten

Hiten

Active Member
Joined
Nov 24, 2019
Messages
291
Likes
359
Location
India
Thanks gentlemen.
Though technically challenging for me I did go through DonH56 informative thread. Post #5 is interesting. from which I understand the source should be utilized fully (not clipping) so as to voltages remain as far away as possible from noise floor.

Here is interesting amplifier. link : Denon Hope the link translates to english.
Kindly see measurements at bottom of the page. It has input senstivity of 0.13v for full power. The CD player has capacity to output 2V. Typical power amplifier does not have such high sensitivity. Wouldn't that amplify noise also ? On the same website Accuphase amp shows much much less noise. Dont know why denon has high input sensitivty either by having low value volume control or too much gain ?
regards.
 

AnalogSteph

Major Contributor
Joined
Nov 6, 2018
Messages
2,522
Likes
2,405
Location
.de
An input sensitivity of 150-200 mV is typical for integrated amplifiers. They traditionally arrived at this value in order to accommodate the customary nominal 300 mV consumer line level with several dB of headroom to spare. If you do the math for a typical phono preamp and cartridge (say, 2.5 mV nominal output for an MM job at 40 dB gain), that's right in the ballpark.

And yes, the extra gain comes with a corresponding noise penalty, though the specifics depend on the number of volume controls and gain distribution. Common concepts with 44-46 dB in a row following the volume pot (whether pre + power or all in one) struggle to make it to 110 dB worth of dynamic range. You can do better with a decent DAC feeding a power amp these days. Fancy preamps back in the day would be employing a two-stage volume control. Decent PGAs are doing the same internally.

The 2 Vrms CD player level isn't nominal output, it's maximum output. Pre-loudness war, CDs were generally mastered to hit average levels of somewhere around 300 mVrms (which would be -16.5 dBFS relative to 2.0 Vrms). This would provide somewhat more headroom than existing sources at the time.

Now we all know what happened later. CDs got substantially hotter (8-10 dB is not uncommon), but not much about the equipment changed. Many people found themselves in a part of the volume pot range where things would get super touchy with worsening channel balance. No doubt one of the most common complaints before streaming popularized volume normalization features (which, mind you, had been existing for years at that point, see e.g. ReplayGain).
 
OP
Hiten

Hiten

Active Member
Joined
Nov 24, 2019
Messages
291
Likes
359
Location
India
excellent and ellaborative post. Thanks.
CDs were generally mastered to hit average levels of somewhere around 300 mVrms (which would be -16.5 dBFS relative to 2.0 Vrms). This would provide somewhat more headroom than existing sources at the time.
at 300mV still the Denon amplifier requires almost 1/3 of that to reach maximum power. Dont know what are benefits to have such high sensitivity ?
 

AnalogSteph

Major Contributor
Joined
Nov 6, 2018
Messages
2,522
Likes
2,405
Location
.de
I am guessing they were shooting for 150 mV (300 mV + 6 dB) and the amp ended up with somewhat lower maximum power than the original designer may have envisaged (who may not even be present any more). It is common to "recycle" established preamp and power amp circuitry without changing the gain. In this case, total gain seems to be about 45.5 dB (16.5 dB preamp + 29 dB power amp, very common values). That ends up being a bit higher than what a ca. 75 Wpc amp really needs, you could still do 100 Wpc like that. Close enough, I suppose.
 
OP
Hiten

Hiten

Active Member
Joined
Nov 24, 2019
Messages
291
Likes
359
Location
India
I am guessing they were shooting for 150 mV (300 mV + 6 dB) and the amp ended up with somewhat lower maximum power than the original designer may have envisaged (who may not even be present any more). It is common to "recycle" established preamp and power amp circuitry without changing the gain. In this case, total gain seems to be about 45.5 dB (16.5 dB preamp + 29 dB power amp, very common values). That ends up being a bit higher than what a ca. 75 Wpc amp really needs, you could still do 100 Wpc like that. Close enough, I suppose.
I miss old school products. If the amplifier had pre out, main in etc. things would have been different. And people would have buying choices (more product sales too). I am sure it doesnt add significant cost to the amplifier manufacturing.
 
  • Like
Reactions: MAB
OP
Hiten

Hiten

Active Member
Joined
Nov 24, 2019
Messages
291
Likes
359
Location
India
Was browsing some more amps and got confused.
Here are accuphase measurements LINK The THD+N graph at 0.1w show much much low levels and Signal to noise ratio is written as (1W A weighted) as 85dB all good.
But here is budget Denon amplifier LINK. This Denon graph starts at 0.1 watt much higher level but the signal to noise ratio is tiny bit better at 87dB ? How come this anomaly ?
Note : Measurements are at bottom of the linked pages
thanks
 

NTK

Major Contributor
Forum Donor
Joined
Aug 11, 2019
Messages
1,893
Likes
3,959
Location
US East
Was browsing some more amps and got confused.
Here are accuphase measurements LINK The THD+N graph at 0.1w show much much low levels and Signal to noise ratio is written as (1W A weighted) as 85dB all good.
But here is budget Denon amplifier LINK. This Denon graph starts at 0.1 watt much higher level but the signal to noise ratio is tiny bit better at 87dB ? How come this anomaly ?
Note : Measurements are at bottom of the linked pages
thanks
The SNR number of 87 dB they gave for the Denon was incorrect. Let's look at the 8 ohm measurements.

THD+N = 0.8% (all from noise) at 0.1 W gives noise voltage of 0.008 * sqrt(0.1 * 8) = 0.007155 V.
Max power at 8 ohm is 52 W, which means v_out @ max power = sqrt(52 * 8) = 20.396 V.
The SNR of the Denon is therefore: -20 log10(0.007155/20.396) = 69 dB
 

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
12,308
Likes
14,605
Location
Riverview FL
It is because of noise. THD+N obviously includes noise. More than likely the THD part is the same percentage or lower as at the higher power levels. Noise might be constant and therefore at lower levels is a larger part of THD+N. These remember are ratios. So same noise in terms of watts is a larger percentage at low powers and a lesser percentage at higher powers.

Measuring THD on my speakers, with 24 sweeps 2dB apart from about 46 to 92dB SPL (at 1kHz) at the listening position.

Basically, only the last four sweeps, the highest signal levels, show the speaker distortion begin to rise out of the measurement noise floor (the dense part of the graph).

1671549864050.png


SPL of the sweeps. Low Frequency ambient noise messes with the low frequencies at the lowest levels.

1671550070734.png
 
OP
Hiten

Hiten

Active Member
Joined
Nov 24, 2019
Messages
291
Likes
359
Location
India
The SNR number of 87 dB they gave for the Denon was incorrect. Let's look at the 8 ohm measurements.

THD+N = 0.8% (all from noise) at 0.1 W gives noise voltage of 0.008 * sqrt(0.1 * 8) = 0.007155 V.
Max power at 8 ohm is 52 W, which means v_out @ max power = sqrt(52 * 8) = 20.396 V.
The SNR of the Denon is therefore: -20 log10(0.007155/20.396) = 69 dB
Now it makes sense discrepancies between chart and the written measurements. Thanks. 69dB is not good measurement for modern amplifiers.
regards
 
OP
Hiten

Hiten

Active Member
Joined
Nov 24, 2019
Messages
291
Likes
359
Location
India
if I may, one more newbie query out of curiosity ...

A typical THD+N chart has tapering down mesurements from low watts to max allowed power where it then suddenl;y shoots up.
what if..
(1) It has very low noise but high distorion at low level will it give same tapering graph ?
Thanks.
 

wwenze

Addicted to Fun and Learning
Joined
May 22, 2018
Messages
941
Likes
1,287
The final THD+N chart is just the result of adding the components together, and on a logarithmic graph it will look almost like the graph simply follows the worst component at any point.

KRDvM38.png


So if distortion is high but the noise is low then you'll just see the flatter distortion-dominated part earlier.
index.php


The distortion of DACs and headphone amps (properly-design ones) are so low today that we are used to seeing that triangular graph. With speaker amps we still can see the point where distortion starts to dominate.
index.php
 
OP
Hiten

Hiten

Active Member
Joined
Nov 24, 2019
Messages
291
Likes
359
Location
India
So if distortion is high but the noise is low then you'll just see the flatter distortion-dominated part earlier.
Exactly what I wanted to know.
Will be searching for any amp measurements where all three charts are shown. i.e. Distortion, Noise and Distortion + Noise. If any one remembers posted. Please post link.
regards.
 
OP
Hiten

Hiten

Active Member
Joined
Nov 24, 2019
Messages
291
Likes
359
Location
India
Update :
I wrote a long thread that has some examples, but it is not an easy read.... It may be easier to skip through at first and check some of the plots and examples.

I revisited this link. And answers most queries
thanks.
 
OP
Hiten

Hiten

Active Member
Joined
Nov 24, 2019
Messages
291
Likes
359
Location
India
hi all,
help me interpret these two unique charts. For a moment lets forget authenticity of measurements, circuit topology etc.

Link : (1) Audio Reveal (2) Atol
Audio Reveal : The noise figure is 87 (not bad i guess) but distortion is high. and a rising distotion with power. But can not interpret /decipher/chaff out noise characteristics.
Atol : almost flat distortion graph (both harmonics at high level) BUT with rising power it does not taper down. what does it mean ?
thanks
 
Last edited:
Top Bottom