Shenzhenaudio Topping EX5 Review (DAC and Headphone Amplifier)

[MRC01]

I put the EX5 in a measurement loop using REW on my PC, same way I've measured other equipment. I played frequency sweeps at sampling rates from 44.1 through 192 kHz. The EX5 frequency response always extends to half the sampling rate (Nyquist).

For example, play a 96 kHz sweep and the EX5 produces output to 48 kHz as it should, but it shows 44.1 on the display the entire time, even when it's clearly operating at 96 kHz. Of course, changing the EX5 digital filter changes the top end of the sweep.

This shows that the problem is a bug in the EX5 display logic. The EX5 correctly decodes and plays digital streams at all sampling rates, but when using the coax or toslink input it always shows 44.1 on the display regardless of the actual sampling rate.

 

[Veri]

This shows that the problem is a bug in the EX5 display logic. The EX5 correctly decodes and plays digital streams at all sampling rates, but when using the coax or toslink input it always shows 44.1 on the display regardless of the actual sampling rate.

That's.. odd. I wonder if yours has a bug or if everyone in this thread has the same problem.

 

[Toku]

I put the EX5 in a measurement loop using REW on my PC, same way I've measured other equipment. I played frequency sweeps at sampling rates from 44.1 through 192 kHz. The EX5 frequency response always extends to half the sampling rate (Nyquist).

For example, play a 96 kHz sweep and the EX5 produces output to 48 kHz as it should, but it shows 44.1 on the display the entire time, even when it's clearly operating at 96 kHz. Of course, changing the EX5 digital filter changes the top end of the sweep.

This shows that the problem is a bug in the EX5 display logic. The EX5 correctly decodes and plays digital streams at all sampling rates, but when using the coax or toslink input it always shows 44.1 on the display regardless of the actual sampling rate.

I don't understand how to use EX5 even if I read your text. So I don't think anyone can give any advice.
Is the USB cable connected to a Windows PC? It was
What is the device to connect with the COAX cable? It was
What are you doing with BEHRINGER DEQ2496? It was
What are you doing at REW? It was
What is My DAC? It was
I need a more specific explanation.

 

[MRC01]

I don't understand how to use EX5 even if I read your text. So I don't think anyone can give any advice.
Is the USB cable connected to a Windows PC? It was
What is the device to connect with the COAX cable? It was
What are you doing with BEHRINGER DEQ2496? It was
What are you doing at REW? It was
What is My DAC? It was
I need a more specific explanation.

Simple:

• start with a music source device D having toslink or coax output
• connect the output of D to the EX5 input (toslink or coax)
• switch EX5 to the proper input to listen to D
• play music on device D, having a sample rate different from 44.1 (use 48, 88.2, 96, etc.)
• EX5 plays the music, but its display shows 44.1 instead of the correct sample rate

If I connect device D into any other device (Behringer DEQ2496, another DAC, etc.) that device shows the correct sample rate.
If I connect device D to the EX5 using USB, the EX5 shows the correct sample rate.

 

[Berwhale]

I have just connected my EX5 to my PC using TosLink. When I switch the sample rate in Windows, the sample rate display on the EX5 switches to match - I tried 44.1, 48, 96 and 192, they all worked...

 

[MRC01]

I can get it to work too.
When connected to the toslink or coax output of my Oppo BDP-83 disc player, the EX5 display shows the correct rate.
When connected to the toslink or coax output of my PC (Juli@ sound card), the EX5 display always shows 44.1 regardless of the rate.

That would make it appear to be a problem with the Juli@ card. But when I connect the Juli@ to other devices, they show the correct rate. It is only the EX5 that always shows 44.1 when connected to the Juli@.

When the EX5 is connected to the Juli@ showing 44.1 while playing a different rate, it does work properly. It is actually operating at the correct rate, it's just incorrectly displaying 44.1.

Also: if I play music into the EX5 USB input at some rate R (say, R=192), then switch to the Juli@ coax playing music at that same rate R, the EX5 correctly displays R.

 

[Berwhale]

When the EX5 is connected to the Juli@ showing 44.1 while playing a different rate, it does work properly. It is actually operating at the correct rate, it's just incorrectly displaying 44.1.

Are you switching the sample rate in Windows or just playing a file with a different sample rate? It sounds to me like Windows is resampling to 44.1 and the EX5 is displaying this correctly.

 

[MRC01]

I don't use Windows, this is on Ubuntu. I configured Pulseaudio to play audio at its native rate without resampling. At first I thought it was being resampled to 44.1 and the EX5 display was correct. I disproved this in 2 ways:
1. I plugged that same coax output, while it was playing, into other devices which showed the correct sample rate.
2. From REW, I played frequency sweeps at 96 kHz into the EX5 via the coax input, while measuring the EX5 analog output. The EX5 showed 44.1 the entire time, but its analog output was clean to 48 kHz, showing that it was operating at 96 kHZ despite its display incorrectly showing 44.1.

In short, the EX5 display is buggy/glitchy with the toslink/coax input. Sometimes the display shows the correct rate, sometimes it gets stuck at 44.1 even while the EX5 is actually operating at different/higher rates.

 

[MRC01]

Question about these digital filters... Comparing Amir's graph to the EX5 manual, they don't match, and the names the manual gives filters don't match their graphs.

Filter #1 starts attenuating early and is -12 dB @ 20k. This is normally called "slow" but the manual calls it "fast apodizing". Apodizing, sure, as it is fully attenuated by Nyquist. But it's not fast, it's slow.
Filter #2 is flat at 20 kHz and fully attenuated by Nyquist. This is normally called "fast apodizing" but the manual calls it "slow minimum". It may be minimum phase, but it sure is not slow.
Filter #3 attenuates early and is -6 dB @ 20 kHz, and it doesn't fully attenuate until 28k. Very slow. Yet the manual calls it "fast minimum phase".
Filter #4 doesn't jive. Amir shows its stopband at 24.1k, but the manual shows filter 4 stopband just at or below 24k. What Amir calls #4 looks like #7 in the manual, which is called "fast corrected minimum". This name seems to fit.
Filter #5 attenuates early and is -4 dB @ 20k, and stopband is not until 28.3k. This is very slow, like filter 3. Yet the manual calls it "fast linear".
Filter #6 is called "brickwall" and its graph matches this description
Filter #7 has a stopband at or just below 24k, so what Amir calls #7 looks like #4 in the manual. It's called "slow linear", but does not attenuate early, and its stopband is 24 kHz, so it's not "slow".

Could this be errors in the printed manual?

I measured my EX5 with REW sweeps to 22.05 at 44.1 kHz with each of the 7 filters. The names and graphs in the EX5 documentation are indeed incorrect as they do not match their measured responses. Measured responses:

Filter	19k	20k	22.05k	phase
1	-3.3	-10	-40	+15, -150
2	0 / flat	-1.5	-24	0 / flat
3	-3	-5	-14	+24, -84
4	0 / flat	-0.2	-7	+210, 0
5	-2	-3.6	-9	0 / flat
6	0 / flat	-0.3	-10	0 / flat
7	0 / flat	-0.3	-10	0 / flat

Filter #2 is most correct. The manual calls it "slow minimum" but it is neither slow, nor minimum phase. Filters 6 and 7 were identical in my testing. They are a bit flatter and smoother than #2 in the passband, but they do not fully attenuate by Nyquist. The impulse responses were consistent with the phase; that is, the ones with flat phase had symmetric impulses, the others had asymmetric impulses.

For those who like pictures, here's what I measured:

 

[Berwhale]

In short, the EX5 display is buggy/glitchy with the toslink/coax input. Sometimes the display shows the correct rate, sometimes it gets stuck at 44.1 even while the EX5 is actually operating at different/higher rates.

Strange; I am able to switch the Windows sample rate randomly and the EX5 always reflects the Windows setting. I can do this whilst audio is playing without interrupting playback. I wonder if the issue is specific to your Juli@ sound card.

 

[MRC01]

Strange; I am able to switch the Windows sample rate randomly and the EX5 always reflects the Windows setting. I can do this whilst audio is playing without interrupting playback. I wonder if the issue is specific to your Juli@ sound card.

It's possible. But as I mentioned above, the Juli@ sound card works perfectly with other devices. So whatever the problem is, it's specific to the combination Juli@ + EX5, not either device alone.

 

[MRC01]

Having mentioned some EX5 anomalies, I'll add what I like about it:

• Clean measurements: low distortion, flat phase & frequency response
• Configurability: 6 different digital filters to choose from (it says 7, but 2 measure identically)
• Very low noise: quieter (lower noise) than my JDS Atom
• Volume control: perfect channel balance at all settings
• Outputs
  • Both SE and balanced, both headphone & line
  • Switch headphone/line without having to unplug headphones
• Look & feel: connectors, knobs, etc. feel good not cheap
• Internal power supply: no wall-wart

It will be a tough decision whether to keep it due to the things I like, or return it due to the glitchy sample rate display. While I'm deciding, maybe Topping will read this, find out what the problem is and fix it with updated firmware.

 

[MRC01]

Lovely, mine is now stuck to 192 even though the actual content is actually 44.1 (I don't think I own any 192 files).

@JohnYang1997 , @SHENZHENAUDIO looks like another bug. Is there any chance you can have an option to display volume level instead of sample rate? It's close to useless even when it's correct, but really unhelpful to have the display stuck on an incorrect sample rate.

@dorirod you're not the only one. This looks like the issue I encountered too: the displayed sample rate is incorrect, does not match what the EX5 is actually playing.

 

[Carlo2AC]

Having mentioned some EX5 anomalies, I'll add what I like about it:

• Clean measurements: low distortion, flat phase & frequency response
• Configurability: 6 different digital filters to choose from (it says 7, but 2 measure identically)
• Very low noise: quieter (lower noise) than my JDS Atom
• Volume control: perfect channel balance at all settings
• Outputs
  • Both SE and balanced, both headphone & line
  • Switch headphone/line without having to unplug headphones
• Look & feel: connectors, knobs, etc. feel good not cheap
• Internal power supply: no wall-wart

It will be a tough decision whether to keep it due to the things I like, or return it due to the glitchy sample rate display. While I'm deciding, maybe Topping will read this, find out what the problem is and fix it with updated firmware.

Is the display fully configurable? as in after i set the filter i prefer, volume and everything i can just turn the display off completely?

 

[MRC01]

From what I can tell, the display is always on when you are using it (music is playing). You can adjust the brightness but I haven't found a way to disable/blank it when music is playing.

 

[Carlo2AC]

From what I can tell, the display is always on when you are using it (music is playing). You can adjust the brightness but I haven't found a way to disable/blank it when music is playing.

Since this is staying in a shared bedroom with a silent media sharing PC always turned on, how dim can i make the display?

I had to remove the led from the Atom AMP and desolder the 9000 lumen flashlight from the Khadas Board as an example

 

[MRC01]

It's so subjective it's hard to convey, but if you removed the LED from your Atom, the EX5 display is probably too bright for you even on its dimmest setting.

 

[Berwhale]

When the screen is on Auto, it dimms after about 60 seconds, hides the sample rate and mode and only displays the input. It's not as bright as it looks in the picture...

I think you could cover this part of the screen with some black PVC tape and still leave the IR sensor exposed.

Just don't play any MQA files because the MQA Status light is quite bright.

 

[MRC01]

When the screen is on Auto, it dimms after about 60 seconds, hides the sample rate and mode and only displays the input. It's not as bright as it looks in the picture..

Thanks for the tip, I just confirmed this works. It blanks within 25-30 seconds. And it makes the incorrect displayed sample rate less annoying: out of sight, out of mind.

 

[Carlo2AC]

When the screen is on Auto, it dimms after about 60 seconds, hides the sample rate and mode and only displays the input. It's not as bright as it looks in the picture...

View attachment 169144

I think you could cover this part of the screen with some black PVC tape and still leave the IR sensor exposed.

Just don't play any MQA files because the MQA Status light is quite bright.

It is decently big, black PVC tape maybe would look quite ugly if put it on the screen? there is nothing in the firmware that i could change?

After dimming it won't turn back on independently of what i do on my computer a part from playing MQA files?