How important is SNR sand THD?

[mrjayviper]

I've been reading various reviews and it's seems the above are given very high importance. Looking at the SINAD "ranking" in the measurements, mid priced amps/DACs can rival expensive ones.

But if SINAD/THD is the only measure, it seems to me Chinese manufacturers have no reason to release more expensive options but they also have expensive models.

Your thoughts? Thanks

 

[Yasuo]

Very...

 

[DVDdoug]

Frequency response is also important but that's about it for determining "sound quality" of electronics. If these three things are better than human hearing there is nothing to "improve". See AudioPhoolery.

Of course there are other considerations such as features and amplifier power and surround sound, and when you put speakers in a room there are acoustical effects (reverb & reflections) etc.

 

[tomchr]

I would argue that any individual spec in isolation is irrelevant. I encourage you to look at many measurements when drawing conclusions about equipment performance. I would also argue that good IMD (and/or multi-tone IMD) is more important than THD. Here's a long-winded explanation of why (with sound bite examples):

In addition to the specs, one should also consider reliability, usability, etc. when selecting gear.

Tom

 

[Katji]

But if SINAD/THD is the only measure, it seems to me Chi-Fi manufacturers have no reason to release more expensive options but they also have expensive models.

It is not the only measure, not the only measure used by manufacturers, in design and production cost model, and not simple correlation to place of manufacture.

 

[sergeauckland]

Not very important, as long as 'adequate'. The only real thing current stellar figures indicate is good engineering. Because getting very good figures is now trivially easy, less than good figures indicate poor design, or, with expensive 'boutique' products, a deliberate, almost inverse snobbery, in demonstrating that they're 'above' mere measurements.

Distortion levels of better than -60dB 0.1% (at all levels and frequencies) are audibly completely transparent, so anything better than that is unnecessary for transparency. Similarly, noise better than -80dB is completely inaudible unless one has a significant mismatch between loudspeaker sensitivity and output level.

In my view, we've been at the point for some years now where the audio specs, in terms of dBs and kHz have become irrelevant, as anything of sensible quality is quite Good Enough. Even before I retired several years ago, Professional Audio tender documents stopped specifying equipment in terms of dBs and kHz, and concentrated on a functional specification, what it had to do, as the audio was taken for granted as being Good Enough.

S.

 

[Plcamp]

Distortion levels of better than -60dB 0.1% (at all levels and frequencies) are audibly completely transparent, so anything better than that is unnecessary for transparency.

I intuitively expect that noticeable distortion levels would be higher at the extremes of the hearable spectrum, and lower in the sensitive midband. Isn’t the distortion detectable by humans a widely varying level vs frequency, am I wrong, or do we not know?

 

[sergeauckland]

I intuitively expect that noticeable distortion levels would be higher at the extremes of the hearable spectrum, and lower in the sensitive midband. Isn’t the distortion detectable by humans a widely varying level vs frequency, am I wrong, or do we not know?

Yes you're right, that hearing sensitivity is higher in the midband, but, tests done a number of years ago (no, I don't have references) indicated that 1% distortion (-40dB) was hard to hear on programme material (not tones or noise) so 0.1% (-60dB) was completely inaudible at any frequency. I have never seen anything since to contradict this.
Bear in mind that tape recorders typically had 3% distortion on peaks, and there are many wonderful recordings done in the '60s and '70s when that level of distortion on peaks was the norm.

Consequently, I can't get excited about an item with 0.001% distortion compared with something of 30 years ago with 0.01% distortion as neither will be audible, nor will something from 40 years ago with 0.1% as long as it's at all levels and frequencies. One issue with early equipment that boasted 0.1% distortion was that in some cases that was only at 400Hz-1kHz, and it was only at 1W output, not at full power and across the audio band. Yes, there were amplifiers that had 0.1% distortion and 20 watts output and a 20-20kHz frequency response, just not all three.

S.

 

[solderdude]

But if SINAD/THD is the only measure,

It isn't. It is just a single number.
Distortion at 1kHz + noise distance to a 1kHz reference level.
One should not make more of this metric than it is.

S/N ratio only becomes important when noise is becoming audible.
There are many different types of distortion. One more audible than another so even percentages alone won't tell you anything about the actual type of distortion nor audibility of it.

 

[MCH]

I do think that SNR is important and I aim for high values in my purchases.

If you use digital attenuation like most people nowadays and your amp has a high gain, you will be attenuating the signal quite a lot on a regular basis (well, unless you always play very loud). And i dont know, but i don't see the reason to go for a dac with a max SNR that doesn't even clear the CD dynamic range this day and age and have hiss getting out of your speakers or headphones.

So for SNR, the higher the better!

Does it make sense or do i need to return to the school?

Ps: and it bothers me specially the people that comes with the jet taking off and the leaf falling. Dude, both things don't need to happen at the same time, and you still might want to hear the leaf falling 10 minutes before the plane takes off?