A SINAD of 80 or SINAD of 100 Can You Really Tell The Difference?

[Jim Creek]

I think SINAD measurements combined with Frequency Response measurements will give a pretty good idea on how that piece of equipment will sound. Room acoustics has a huge impact on SQ. But equipment that measure well will make it easier to adjust by treating the room or adding an equalizer or subwoofer.

Equipment that measure poorly combined with problematic room acoustics are impossible to deal with.

 

[solderdude]

I think noise floor, frequency response (under a correct load), and multitone paint a much better picture.
Maybe IM distortion and distortion over a frequency range (needs harmonic spectrum as well) are also needed.
Especially for amps that never will reach their max output levels in practice a SINAD near max output voltage will say very little.
For amps also various loads are essential.

 

[dasdoing]

how quiet of a noise can you guys even hear in room? https://www.audiocheck.net/blindtests_dynamic.php?dyna=78

 

[sarumbear]

I very much doubt it. All the results I've seen indicate human ears are good for 70dB max, usually much less.

What if you sit next to a monitor and hear the hiss? SINAD includes noise.

 

[fpitas]

What if you sit next to a monitor and hear the hiss? SINAD includes noise.

If there's no signal, what's the SINAD in that case?

 

[sarumbear]

If there's no signal, what's the SINAD in that case?

SINAD is a combination of distortion and noise. When there’s no signal there’s no distortion but there’s still noise, which you may hear.

Technical Library

www.ap.com

 

[fpitas]

SINAD is a combination of distortion and noise. When there’s no signal there’s no distortion but there’s still noise, which you may hear.

Technical Library

www.ap.com

Right. So that still doesn't answer the question. If there's no signal how do you define SINAD?

 

[theREALdotnet]

I think noise floor, frequency response (under a correct load), and multitone paint a much better picture.

I’m a sucker for multitone paint.

 

[sarumbear]

Right. So that still doesn't answer the question. If there's no signal how do you define SINAD?

All active devices are noise generators. Noise is hence generated by the device under test irrespective of the signal. You calculate SINAD by comparing that noise level to the nominal/reference output level.

 

[fpitas]

All active devices are noise generators. Noise is hence generated by the device under test irrespective of the signal. You calculate SINAD by comparing that noise level to the nominal/reference output level.

You've dodged my point. SINAD is inappropriate here. You short the amp input and measure the noise output, then use that with the sensitivity of the speaker. SINAD does not figure into that.

 

[sarumbear]

You've dodged my point. SINAD is inappropriate here. You short the amp input and measure the noise output, then use that with the sensitivity of the speaker. SINAD does not figure into that.

I am sorry that I couldn't explain this to you. All I can say is that SINAD and the S+N/N ratio are the same when there is no signal. However, a speaker's output is not the best way to measure an active device. It will confuse you.

 

[amirm]

You've dodged my point. SINAD is inappropriate here. You short the amp input and measure the noise output, then use that with the sensitivity of the speaker. SINAD does not figure into that.

As a practical matter, the SINAD/THD+N meter doesn't care if there is or is not a signal. It simply subtracts a tone that you give it and what remains, which in this case is all noise, is reported as THD+N relative to whatever signal level you would have played, but didn't.

 

[solderdude]

When the SINAD is noise dominated... yes.
In that case the difference between SINAD 80 and 100 may well be audible.

 

[sarumbear]

When the SINAD is noise dominated... yes.
In that case the difference between SINAD 80 and 100 may well be audible.

As SINAD figure is not descriptive of the ratio of distortion and noise, you will never know, hence the higher the SINAD safer for you. Contrary to some member's beliefs, it is not really a SINAD race but a race to a clean clean and quite amplification

 

[jomo]

As SINAD figure is not descriptive of the ratio of distortion and noise, you will never know, hence the higher the SINAD safer for you. Contrary to some member's beliefs, it is not really a SINAD race but a race to a clean clean and quite amplification

More like a race to unnecessarily clean amplification, which seems to be the name of the game these days. SINAD just confuses what is most important and overweights the value of low harmonic distortion. If there were two amplifiers equal in everything but SINAD. Amp A with a SINAD 80 dB (-80 dB THD, -125 dB noise), amp B with a SINAD of 120 dB (-200 dB THD, -120dB noise. The amps would sound identical and amp A, with the mediocre SINAD, would be quieter. Unless of course someone can prove they can detect -80 dB distortion while music is playing through a loudspeaker. This is the case where an overused engineering figure is fogging objectivity.

 

[sarumbear]

SINAD just confuses what is most important and overweights the value of low harmonic distortion.

Why?

This is the formula. Where do you see the weighing?

 

[jomo]

Why?

This is the formula. Where do you see the weighing?

With the SINAD formula, Pnoise and Pdistortion are treated equally, yet the ability to detect these defects is quite different. Distortion needs a signal to be heard. It is either masked by the distortion of the loudspeaker or masked by the many harmonics that are already in the music. I'm not aware of anybody that can hear distortion beyond -70 dB in music....Singular tones are another thing. However, noise can be heard when the signal is off and the only thing to mask this is the noise floor of the listening environment. If you have a really quiet listening room or closed back headphones, you may benefit from -120 dB. These two "distortions" are not created equal.

 

[danadam]

Amp A with a SINAD 80 dB (-80 dB THD, -125 dB noise), amp B with a SINAD of 120 dB (-200 dB THD, -120dB noise. The amps would sound identical and amp A, with the mediocre SINAD, would be quieter.

Do you have examples? And how many of them?

FWIW I went through every 10th (by date) review of an amplifier by amirm and, if I didn't mistype anything, here's the data:

Spoiler: data

Amplifier	SINAD	SNR @ 5 W	SNR @ Full
Trinnov 8m 8 Channel	106	109	129
Audioengine N22	64	86	91
Audiophonics HPA-S400ET	105	109	128
boXem ARTHUR 2408/N2	96	102	116
Sonance DSP 2-150	75	96	111
Outlaw 2220	85	86	104
Sabaj A1	68	80	87
Schiit Vidar	81	82	107
March Audio P502	93	100	120
ATI AT522NC	98	102	121
IOM NCore Pro PWR	94		113
Class D Audio CDA-250C	80	81	97

So Sonance looks like an oddity and Audioengine N22 (SINAD 64) may be quieter than Schiit Vidar (SINAD 81) or Outlaw 2220 (SINAD 85) at 5W, but not at full power. Other than that it seems pretty well correlated.

But maybe I was just lucky, so anybody, feel free to go through all of them, don't be shy

 

[mdsimon2]

Do you have examples? And how many of them?

FWIW I went through every 10th (by date) review of an amplifier by amirm and, if I didn't mistype anything, here's the data:

Spoiler: data

Amplifier	SINAD	SNR @ 5 W	SNR @ Full
Trinnov 8m 8 Channel	106	109	129
Audioengine N22	64	86	91
Audiophonics HPA-S400ET	105	109	128
boXem ARTHUR 2408/N2	96	102	116
Sonance DSP 2-150	75	96	111
Outlaw 2220	85	86	104
Sabaj A1	68	80	87
Schiit Vidar	81	82	107
March Audio P502	93	100	120
ATI AT522NC	98	102	121
IOM NCore Pro PWR	94		113
Class D Audio CDA-250C	80	81	97

So Sonance looks like an oddity and Audioengine N22 (SINAD 64) may be quieter than Schiit Vidar (SINAD 81) or Outlaw 2220 (SINAD 85) at 5W, but not at full power. Other than that it seems pretty well correlated.

But maybe I was just lucky, so anybody, feel free to go through all of them, don't be shy

View attachment 300063

SNR at full power is a terrible comparison metric because full power is not consistent between amplifiers.

However, would love to see a large scale comparison between 5 W SINAD and 5 W SNR. I think they mostly correlate very well but I feel like I’ve seen a few outliers. In general I agree with @jomo that noise performance is more important than THD performance.

Michael

 

[sarumbear]

With the SINAD formula, Pnoise and Pdistortion are treated equally, yet the ability to detect these defects is quite different. Distortion needs a signal to be heard. It is either masked by the distortion of the loudspeaker or masked by the many harmonics that are already in the music. I'm not aware of anybody that can hear distortion beyond -70 dB in music....Singular tones are another thing. However, noise can be heard when the signal is off and the only thing to mask this is the noise floor of the listening environment. If you have a really quiet listening room or closed back headphones, you may benefit from -120 dB. These two "distortions" are not created equal.

That is why @amirm's chase to highest SINAD makes sense, doesn't it? If a SINAD is at levels around 120dB, you will neither have distortion nor noise. Aim to the top and be safe.