• Welcome to ASR. 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!

Bluesound Node Icon Streamer Review

Rate this streamer/DAC/Preamp:

  • 1. Poor (headless panther)

    Votes: 43 19.0%
  • 2. Not terrible (postman panther)

    Votes: 98 43.4%
  • 3. Fine (happy panther)

    Votes: 73 32.3%
  • 4. Great (golfing panther)

    Votes: 12 5.3%

  • Total voters
    226
So, with a test track that is capable of generating imaging artifacts in the ultrasonic range, if you were looking how close the recording of the DAC matches the digital file from the phase standpoint, the Bluesound Node Icon with QRONO d2a actually beats the Fosi ZD3.
Appreciate your time doing these tests.
Are these phase changes audible? I suspect not.
 
you speak filter of dac ....but present on the recording in 44.1?
There is nothing ultrasonic in the recording. But the DAC process causes ultrasonic images (audio band reflected at the nyquist frequency into the ultrasonic band). This is why a DAC needs a filter, and is why (at 44.1) that filter needs to be sharp since there is only the band from 20kHz to 22.05kHz to attenuate the reflected images.

See figure 2 on page 3:

 
Are these phase changes audible? I suspect not.
I suspect this increased phase accuracy in the audible spectrum, is more audible than the apparent side effect of ultrasonic IMD - which is in the out of band spectrum and some -80dB down.

In reality it's probably tit for tat and neither are particularly audible, but these are interesting tests nonetheless (thanks GXAlan for taking the time to run then), as it likely provides a reason for QRONO being implement as it is.

Everything in audio seems to be a trade-off in some way, if everything is as appears then for me the increased phase accuracy through the audible band looks to be a benefit vs the ultrasonics trade-off. In an ideal world we would have the phase accuracy without the ultrasonic IMD, but I guess we don't know how intrinsically linked the two are, and if a steeper filter would then compromise the accuracy. At worst this is all just a technical benefit, and best it might be audible - but that'll need a large pool of testing. And I guess the possibility exists that maybe as QRONO is further developed a newer filter marrying the best of both could be developed.
 
I suspect this increased phase accuracy in the audible spectrum, is more audible than the apparent side effect of ultrasonic IMD - which is in the out of band spectrum and some -80dB down.

A slight correction to this. IMD occurs in the speaker/tweeter. Even when it is ultrasonic tones that are inter-modulating, It can create in-band (below 20kHz) distortion products, and we don't know how far down they are because it depends on how bad the distortion in the speaker is. Speakers are orders of magnitude worse for distortion than electronics are.

Even so, I expect that in most cases they are inaudible. But perhaps not in all.
 
In defense of QRONO d2a?

To be clear, this is just a scientific approach. I'm not sure it's even audible, but seeing these results was worth sharing.

I'm having trouble with doing a blind test. When I volume match in Audacity and save a FLAC, it increases in size tremendously even though I have dithering off, which prevents me from attaching the files to the post.

Test CD: IASCA 2023 Reference CD

I chose this since it's a bunch of high quality 16/44.1 kHz content. I chose Track 7 on the IASCA disc, which is Anvil of Crom from Telarc's Great Fantasy Album, a very percussive track. I have that disc somewhere as well, so I could check the liner notes, but this is DDD and done by Telarc, which was one of the best audiophile labels in the 90's. I chose this because I saw the imaging artifacts.

I then recorded:
1) Node Icon from network share --> E1DA Cosmos ADC at 24/96
2) Fosi ZD3 from USB --> E1DA Cosmos ADC at 24/96

Fosi ZD3
View attachment 434379

Node Icon
View attachment 434378

There is a huge shelf above 16 kHz which is sort of weird, since I see that on the original FLAC too. Notice the spurious tones at 30 kHz? That's the problem of the QRONO d2a.

So then, it's worth looking at the original source file:

Source File
View attachment 434380

That's my fault. There isn't a lot of content above 16 kHz to begin with.

So, you'd ASSUME that I chose a poor test track. And you'd assume that all of the content above 16 kHz is buried to -70 dB or lower, and be completely masked.

I then brought things into DeltaWave. Pretty obvious that white is the Fosi ZD3 and blue is the Node Icon.
View attachment 434381
View attachment 434382

Nothing surprising yet.

But QRONO talks about phase, so I compared the Fosi ZD3 and Node Icon. So now we are seeing a difference.
Phase: Fosi ZD3 vs Node Icon
View attachment 434383


This is interesting because the difference occurs in the audible range of 6kHz and up.

So now, the test is to compare the digital file at 16/44 versus the recordings...
Phase: Recording of Fosi ZD3 vs. Digital File
View attachment 434384

Phase: Recording of Node Icon vs. Digital File
View attachment 434385


Just before you come to your conclusion.... take a look at the Y-axis. Let's set them to be equal:

Bluesound Node Icon Recording vs. Digital File

View attachment 434387



Fosi ZD3 Recording vs. Digital File
View attachment 434388



So, with a test track that is capable of generating imaging artifacts in the ultrasonic range, if you were looking how close the recording of the DAC matches the digital file from the phase standpoint, the Bluesound Node Icon with QRONO d2a actually beats the Fosi ZD3.
The filter in the Fosi ZD3 is the default ESS minimum phase filter as is evident from the stop band ripple in Amir's review. A linear phase filter would have constant group delay in the pass band.
1741443542513.png
 
Changed mine to poor after watching the video review. Idiotic filter implementation and idiotic not to give alternatives. Poor headphone amp for the price and careless on jitter which might not be audible just shows an attitude of not striving for excellence with minimal effort.
Edit : not grounding the balanced connection or include Limited Dirac so all you need is the upgrade are other reasons.
 
Last edited:
Appreciate your time doing these tests.
Are these phase changes audible? I suspect not.
I suspect this increased phase accuracy in the audible spectrum, is more audible than the apparent side effect of ultrasonic IMD - which is in the out of band spectrum and some -80dB down.

In reality it's probably tit for tat and neither are particularly audible

So, the obvious conflict of interest is that I own a Node Icon. That said, my posting history will still show that I got the icon after trying the Fosi ZD3 and WiiM Ultra and finally settled on the Icon. Down the line, if they ever come up with a Node Icon v2, I know a recommended score from Amir boosts the resale value.

I don’t think any of this will be that audible either. This does not have the highest SINAD on the market, but it sure gets up there for SNR, especially when factoring the output voltage that’s under 4V. Residual noise is the hiss you get from the tweeter when you are right next to it, and the Node Icon is impressively quiet.

So, when a company clearly has demonstrated engineering discipline, it’s hard to reconcile the use of QRONO based upon the silly high IMD. On the other hand, when MQA/Bluesound say that we think the imaging artifacts are so rare in real music despite test tone weaknesses, but that the benefit in real music is present, you get an engineering/marketing explanation.

It could be marketing to help drive MQA licensing. It could be misguided audiophoolery which can affect anyone. It could have a modicum of truth.

The filter in the Fosi ZD3 is the default ESS minimum phase filter as is evident from the stop band ripple in Amir's review. A linear phase filter would have constant group delay in the pass band.
That’s an example of a product with a non-changeable filter (even though it has an ESS DAC) and my current standalone DAC du jour for comparison. The only other DACs I have are the Topping D50s and Korg DS-DAC-10.

After the added phase measurements from GXAlan I've changed my vote to GREAT.
Changed mine to poor after watching the video review. Idiotic filter implementation and idiotic not to give alternatives. Poor headphone amp for the price and careless on jitter which might not be audible just shows an attitude of not striving for excellence with minimal effort.

I think the headphone output is a valid point. We know clipping is an issue. It’s sort of sad that they paid for the THX AAA license but didn’t give us flagship level amplification.

If you have lower impedance headphones though, it does beat the WiiM Ultra though.

Not including Dirac LE is definitely a bit of “nickel and diming” but the whole XLR thing is blown out of proportion. You might run into ground loops or you might not. While it occurs MORE than other products, if it doesn’t occur for you, it really does have some of the very best noise performance. In theory, there is no ground reference for XLR since it should be the differential.

The voltages are also sort of weird for sure. I wonder if we have some sort of intersample over tweaking.

That said, the real weakness? It’s made in China and the tariffs have bumped the price up to $1149 for those of us in the USA.

It was already pretty pricey, and that extra $150 would have been the cost of Dirac.
 
That’s an example of a product with a non-changeable filter (even though it has an ESS DAC) and my current standalone DAC du jour for comparison. The only other DACs I have are the Topping D50s and Korg DS-DAC-10.
That reduces the argument to the QRONO filter having less phase distortion than a steep minimum phase filter. This says nothing about its benefits compared to a steep linear phase filter which is what is closest to the theoretically ideal filter. In reference to your title, your comparison is a quite weak defense.

The D50s has linear phase filters. It should be superior to the QRONO filter both in image rejection and amplitude and phase accuracy. Filter 5 should be representative of typical steep linear phase filters.
 
That reduces the argument to the QRONO filter having less phase distortion than a steep minimum phase filter.
This says nothing about its benefits compared to a steep linear phase filter which is what is closest to the theoretically ideal filter.
In reference to your title, your comparison is a quite weak defense.

300% agree. But would still stand by the opinion that it’s inaudible.

The bigger story now is that Bluesound has upped their price 15%. WiiM Ultra is still available at $300 after coupon.
 
  • Like
Reactions: VQR
A slight correction to this. IMD occurs in the speaker/tweeter. Even when it is ultrasonic tones that are inter-modulating, It can create in-band (below 20kHz) distortion products, and we don't know how far down they are because it depends on how bad the distortion in the speaker is. Speakers are orders of magnitude worse for distortion than electronics are.

Even so, I expect that in most cases they are inaudible. But perhaps not in all.
Ah ok, thank you. So the presence of ultrasonics (well, in this case introduced due to the filter) - is what then creates IMD at the speaker. The presence of ultrasonics is not itself IMD.

I think I have understood that correctly? Cheers
 
Back
Top Bottom