Neutron HiFi DAC V1 Review

[OHtaru]

What about High-Speed mode?

Sorry, I jumped to conclusions too early.

I just successfully reproduced the pop sound in High-Speed mode as well and it can be reproduced on Windows, iOS, and Android. It's quite random and sometimes does not appear several times in a row, so it may need to be tried several more times.

This seems to be digital rubbish coming from the MCU, as the pop sound is volume-controlled (the pop sound is most noticeable at maximum volume).

I recorded the pop sound (audio interface 96k, DAC V1 48k, attached below), and the waveform indicates that it's a single-sample pulse:

 

[neutroncode]

This seems to be digital rubbish coming from the MCU, as the pop sound is volume-controlled (the pop sound is most noticeable at maximum volume).

Nope, this pop & click is produced exclusively by ESS chip when it is reinitialized with a new mode, on power off/on+reinit and etc. So the problem comes only from there. It can be suppressed by playing with chip's modes switching and it is what DAC V1's firmware is doing. Another approach, which is often taken by other developers, is to insert an analog volume IC and mute analog output before applying changes to the DAC chip.

Probably, if you switch on High Performance mode in NConfigurator those sounds will go away because firmware will not attempt to power off DAC chip completely?

Will try to reproduce what you noticed.

 

[AndyBell]

Firmware 70 and NConfigurator 1.9.3 were released today!

For DAC V1 the Firmware 70 is the most important update because it fixes nasty bugs happened due to its asynchronous core. It also improves ALC DSP behavior when Volume Limit is applied, now ALC is not affected by it.

Hi

The new Configurator works OK on Windows, and I used it to update the DAC's firmware, but on my Android phone (Samsung Galaxy A16) it crashes as soon as the configurator connects to the DAC. Android reports 'this application has a bug...' but doesn't give further details...

Andy

 

[d1m0n]

Firmware 70 and NConfigurator 1.9.3 were released today!

For DAC V1 the Firmware 70 is the most important update because it fixes nasty bugs happened due to its asynchronous core. It also improves ALC DSP behavior when Volume Limit is applied, now ALC is not affected by it.

@neutroncode I updated to FW70. It looks like the "Apple volme scale" and volume limit slider setting is somehow broken now. Volume adjustment act only between 0 to -19dB and then jumps straight to -128 according to V1s display. Volume limit slider audible works within 0 to -19db range but doesnt limit according to V1 the dB dsiplay setting. Im still abble go up to 0db

 

[neutroncode]

Hi

The new Configurator works OK on Windows, and I used it to update the DAC's firmware, but on my Android phone (Samsung Galaxy A16) it crashes as soon as the configurator connects to the DAC. Android reports 'this application has a bug...' but doesn't give further details...

Andy

Managed to reproduce too. Will fix it asap for all platforms, surprisingly crashing Android only.

Update: @AndyBell I re-uploaded fixed APK to GP and web site. Thank you for reporting this failure!

 

[neutroncode]

@neutroncode I updated to FW70. It looks like the "Apple volme scale" and volume limit slider setting is somehow broken now. Volume adjustment act only between 0 to -19dB and then jumps straight to -128 according to V1s display. Volume limit slider audible works within 0 to -19db range but doesnt limit according to V1 the dB dsiplay setting. Im still abble go up to 0db

Yes, volume you see on display is no longer affected by the Volume Limit. It is applied to DAC chip only but DSP still thinks that volume is not attenuated (and display dB reading reflects it), it allows ALC DSP to work correctly. So display reading starts from the Volume Limit, i.e.: 0 dB = Volume Limit. For me, on Windows if I adjust volume slider slowly I am not getting any jump from -19 to -128. I tried setting Volume Limit to -20 dB and still display reading was similar when it is without Volume Limit.

Update: If Apple Volume Scale is On then display reading is coarse but that is how Apple scale works actually, when Volume Limit was affecting display reading it was smoother due to the limit.

 

[OHtaru]

The Apple Scale works the same as before on my side. The volume and ALC have been decoupled from the volume limit in FW70 and the volume always shows the relative value to the max.

 

[neutroncode]

NConfigurator 1.9.4 is out with a hot fix on all platforms as the bug with not Android-only.

 

[neutroncode]

Firmware 71 is out!

Many thanks to @OHtaru for a very productive cooperation regarding the elimination of the discovered pop & click sound, and @d1m0n for raising and further testing the issue with Apple Volume Scale problem when Firmware 70 changed behavior of the Volume Limit, now Apple scale works similar to Windows/Linux for macOS case and almost similar on iOS (somehow Apple introduced discrepancy regarding how volume is treated on macOS and iOS).

 

[pioterd]

so you won't get clipping

May I ask you if ES9219 can be used in 2Vrms mode as a line-level output device, as a DAC for bigger amplifier without the risk of clipping when using full 2Vrms output from ES9219?

 

[neutroncode]

May I ask you if ES9219 can be used in 2Vrms mode as a line-level output device, as a DAC for bigger amplifier without the risk of clipping when using full 2Vrms output from ES9219?

Yes, it is perfectly valid scenario for DAC V1. Normally, impedance of the amplifier is >1 kΩ (10-100k) which is no load case and gives 2 Vrms with lowest THD.

Edit: If you decide to get DAC V1 for this usage scenario, set THD Compensation -> Load Impedance to No Load (DAC/Preamp) optimized specifically for this usage scenario. It was kindly provided and measured by @jkim with the following results: THD = -117.6 dB, THD+N = -114.4 dB at 0 dBFS. Basically min possible THD+N for ES9219.

 

[OHtaru]

Recently I briefly tested the behavior of MIC AGC, but it does not seem to behave like a traditional AGC/compressor. After encountering a loud volume, it no longer attempts to restore the gain, yet at this point it appears to act like a noise gate.

I have read the manual regarding the AGC parameters in NConfigurator, but this still fails to form an intuitive model in my mind. Where is the "threshold" here? Or is it a program-dependent parameter?

@neutroncode Could you please explain how the AGC works? It would be quite clear if a gain transfer curve could be provided. Thanks.

 

[neutroncode]

Recently I briefly tested the behavior of MIC AGC, but it does not seem to behave like a traditional AGC/compressor. After encountering a loud volume, it no longer attempts to restore the gain, yet at this point it appears to act like a noise gate.

AGC is not a Compressor, it is Automatic Gain Correction. This kind of DSP can be found in XBox software for the mic btw for the voice-chat functionality and it is much simpler than Compressor (both these DSPs are in Neutron Player and Neutron Recorder apps).

There is a Target Audio Gain setting, so the captured audio is amplified/attenuated towards that Target Audio Gain. The RMS is used to get loudness but the Window size is not exposed to settings (that can be done probably) which affects the latency. So after a loud volume RMS shall calm down, probably smaller Window would be useful to adjust how fast it is done.

 

[OHtaru]

AGC is not a Compressor, it is Automatic Gain Correction. This kind of DSP can be found in XBox software for the mic btw for the voice-chat functionality and it is much simpler than Compressor (both these DSPs are in Neutron Player and Neutron Recorder apps).

There is a Target Audio Gain setting, so the captured audio is amplified/attenuated towards that Target Audio Gain. The RMS is used to get loudness but the Window size is not exposed to settings (that can be done probably) which affects the latency. So after a loud volume RMS shall calm down, probably smaller Window would be useful to adjust how fast it is done.

Thanks. That's much clearer now.

But there's one more thing: it seems that the noise gate is triggered to open and remains enabled after a continuous full-scale segment unless a new recording is restarted. Is this expected?

 

[neutroncode]

But there's one more thing: it seems that the noise gate is triggered to open and remains enabled after a continuous full-scale segment unless a new recording is restarted. Is this expected?

According to DSP algorithm there is nothing what could be stuck. I will PM you regarding more description in order to understand your question better.

Btw, NConfigurator 1.9.6 is out.

 

[migo77]

Hello @neutroncode, can you post nconfigurator version for 32bit arm, like RPI2B please? I'm using neutron dac in headless mode on RPI2 with Debian 12 configured as UPnP endpoint and have no options to control DAC's configuration. Thank you!

 

[neutroncode]

can you post nconfigurator version for 32bit arm, like RPI2B please?

Should be no problem, once ready will post an update here.

 

[migo77]

Should be no problem, once ready will post an update here.

Thank you! I'll test it and report back.

 

[neutroncode]

Published 32-bit arm (armv7l) version on the DAC V1 Details page:

Neutron HiFi DAC V1: Specs, Features And NConfigurator Downloads

Explore the Neutron HiFi DAC V1 - a portable, high-resolution USB DAC and preamp featuring ESS ES9219Q, 32-bit/384 kHz PCM, DSD256 support, and integrated DSP for studio-quality, audiophile-grade audio on the go.

neutronhifi.com

Short user manual for guys using DAC V1 as preamp/decoder with headless devices, such as RPI2B or similar:

1. Start NConfigurator as HTTP daemon: NConfigurator --httpd 8083.
2. Open https://nconf.neutronhifi.com in a browser on any device within the same LAN/WAN.
3. Then click 3-dot button -> Custom Address -> put IP of your headless device where HTTP daemon is running.
4. Press Connect.

 

[migo77]

It is working for me with RPi2B. Changing DSP settings stops playback, is it normal? Thank you!