Mediocre measurements for new Sonos Amp

Looks like the line-in tests were done without modifying the line-in level in the Sonos app in any way, resulting in poor test results with external sources, but (obviously) excellent amp results when using AirPlay streaming. Tester apparently didn't even took the time to add a music service to the Sonos app at all, so there's no test result for app-internal streaming sources either. Using a turntable as line-in source without adding a pre amp in-between doesn't sound too well-prepared either.

Apart from that it's an excellent breakdown of the AMP's built quality and overall look and feel.

I don't get the point about phono preamp; the reviewer just used signals from test equipment into the line-in.

Is someone from Sonos able to speak to the higher than expected noise levels that were measured (~93 dB SNR), which is worse than Sonos has been claiming.

I enjoyed the pretty graphics, but the review has quite a few errors.

I would expect the performance of the Line-In signal to be no worse than that of AirPlay or directly streamed sources, assuming that the Line-In gain ('Line-In Source Level') is adjusted correctly. However, nowhere is it mentioned what adjustments, if any, were made.

As for sampling at 48KHz, capping signals at ~20Khz: so what? It's above the frequency required for any human's hearing.

I too suspect the line levels were set at the default, which for mine was I think level 2. At level 2 with the volume turned all the way up, the music didn’t get very loud at all. I have since experimented and it is now set to level 7 which gives me volume more than I’m comfortable with.

Looks like some very iffy "science" going on over at that site. Doubt the admin has any real science background. Looks like he simply doesn't understand how to set the input level.

I think he's pretty accomplished:

https://www.audiosciencereview.com/forum/index.php?threads/a-bit-about-your-host.1906/

OK, great engineering (not science) background, so he should be able to set the input level appropriately. Wonder why he didn't, apparently?

chicks wrote:

OK, great engineering (not science) background, so he should be able to set the input level appropriately. Wonder why he didn't, apparently?

It's a more than just adequate test model for conventional audio amplifiers.
I'd love to see such an in-depth breakdown of my (very conventional and indeed very analogue) Denon home stereo setup with integrated amps. A test like the one above would very much cover any important parts of this product.

I simply don't feel this test model to be sufficient enough to cover the range of modern digital streaming amplifiers.
The lack of covering tests of HDMI sources due to lack of compatible test hardware says a lot...

According to this article, Sonos specs the SNR at 116db, which puts the noise WELL below audibility. They certainly can't lie about that, so this guy is simply not measuring correctly. Perhaps Sonos' engineers should show him how to do it correctly.

https://hometheaterhifi.com/blogs/sonos-amp-review/

Lots of experience testing. Not Amir’s first rodeo.

https://www.audiosciencereview.com/forum/index.php?threads/master-index-for-audio-hardware-reviews.2079/

chicks wrote:

According to this article, Sonos specs the SNR at 116db, which puts the noise WELL below audibility. They certainly can't lie about that, so this guy is simply not measuring correctly. Perhaps Sonos' engineers should show him how to do it correctly.

https://hometheaterhifi.com/blogs/sonos-amp-review/

But Sonos does not include that 116dB spec in their own literature. Someone from Sonos should reply here and clear this up.

I see his measurements of the Bluesound Node are pretty mediocre, as well. Something just doesn't seem right with his measurements.

https://www.audiosciencereview.com/forum/index.php?threads/review-and-measurements-of-bluesound-node-2i-streamer.6631/

I just started reading it, but it reads like someone who knows nothing whatsoever about Sonos - not familiar with daisy chaining Ethernet or using it as a bridge, unaware of several previous Sonos products with inputs. And what’s that crack about the custom power cord? I guess I’ve never tried other cords but I always assumed they were standard figure 8 size. Is the Amp not standard size?

If sonos has data that proves these measurements wrong then lets see them. I have more invested in sonos than most - 3 Connect Amps, 1 Connect, 1 Play 5 Gen 2, Pair of Play 1s with sub and currently have a pair of play1s and a spare connect amp and connect in storage but I would like to know that technically the units measure well especially if you are a company that speaks of engineering excellence.

Sonos is likely measuring SNR under different conditions. In the ASR review, SNR was measured at 5W output. SNR may be much better when the amp is putting out more power. Also, A-weighting may a factor.

It would be helpful to hear from them.

MSNWatch wrote:

If sonos has data that proves these measurements wrong then lets see them. I have more invested in sonos than most - 3 Connect Amps, 1 Connect, 1 Play 5 Gen 2, Pair of Play 1s with sub and currently have a pair of play1s and a spare connect amp and connect in storage but I would like to know that technically the units measure well especially if you are a company that speaks of engineering excellence.

Why are these measurements so important to you? If you are not happy with what your ears tell you, why do you keep buying Sonos products then?

Here is an interesting white paper on the NAD M32.

https://nadelectronics.com/wp-content/uploads/2017/03/DirectDigital_WhitePaper.pdf

This is the same technology that Sonos uses in the new amp. The 2nd page talks about SNR at full power and at normal listening levels. Considering this, the Sonos amp streaming SNR number, measured in the ASR review to be 92dB @5W, are good. Not as good as the much more expensive NAD M32 but the Sonos amp is not measuring poorly. The amp's SNR (again 5W reference) with analog input is a little worse at ~87 dB, but still not bad.

Assuming this French site is correct, Sonos is using Qualcomm DDFA chips:
https://translate.google.com/translate?hl=en&sl=fr&u=https://www.lesnumeriques.com/ampli-audio-video/sonos-amp-p45781/ifa-2018-sonos-presente-amp-son-amplificateur-stereo-connecte-n77667.html&prev=search

From the Qualcomm DDFA literature:
A unique and patented digital architecture delivers market leading audio performance across the entire audio frequency band delivering:

SNR and Dynamic range > 113dB (A-weighted, 20Hz to 20KHz)
THD+N < 0.005% (20Hz to 20KHz)
Closed-loop operation and high-resolution DSP helps maintain performances at all times:

Distortion does not increase at maximum volume like many other amps
No audible noise or hissing sound
Performances are maintained across a wide range of audio input formats


-----
This guy apparently doesn't have the ability to measure via digital inputs.

So, either the ADC in the Amp has issues, or he's measuring something wrong. My money's on the latter...

Here's an amplifier that appears to use the same Qualcomm chipset, with measurements using an Audio Precision analyzer.

http://www.elytone.com.tw/odm_oem/index.php?route=product/product&path=57&product_id=153

Note that THD+N is 0.005% across the audible band.

Yes but Sonos specs THD+N at 0.1%. I assume that is 0.1% at 1W, which agrees with the measurement in the ASR review.

So, reading through the review a bit more carefully this time. He does measure a digital input, "streaming from Android", and gets excellent THD+N of 0.002% @5W, 1kHz, which is within the spec for the Qualcomm chipset.

I apologize for not reading more carefully the first time through. The question still remains, why is the analog input adding so much noise?

How many customers will actually be using the analog input in 2019? Very few, I'm guessing. I do, as I still love to spin LPs, but I'm probably the exception (I own the Connect:Amp, not the Amp). The measured analog performance probably won't much matter for LP playback, but for CD's it might, though hopefully most folks have ripped their CDs to a NAS long ago, and won't much care about the analog input...

chicks wrote:

He does measure a digital input, "streaming from Android", and gets excellent THD+N of 0.002% @5W, 1kHz, which is within the spec for the Qualcomm chipset.

i think that's the bottom line. It's a great amp for streaming (and I assume for optical/hdmi input as well) but the analog input is flawed. I wonder if there is any chance that it was a bad unit. Maybe someone from Sonos will say something?

lahey wrote:

i think that's the bottom line. It's a great amp for streaming (and I assume for optical/hdmi input as well) but the analog input is flawed.

I remain a little skeptical of this, because the test methodology does not specify what Line-In gain was used, and I see no reason why the Amp's ADC should introduce below-par noise performance. I'll feed mine with a good analog signal at the weekend and see if I can detect any differences between that and the same content in FLAC pulled from NAS.

I listen primarily from the line in and it sounds great to me but what do I know. :8

pwt wrote:

I'll feed mine with a good analog signal at the weekend and see if I can detect any differences between that and the same content in FLAC pulled from NAS.

The analog input SNR measurement after he changed the software setting was 87 dB at 5W. So probably you will not hear the noise.