Topping E50 Review (Balanced DAC)

[Nonick]

Lowering digital volume in some lower range (say to -12 dB) does not lead to audio quality loss considering that resolution of current DACs is about 20 bits. But one has to know that every 6dB of lowering leads to 1 bit of digital resolution loss and thus also dynamic range and S/N ratio loss.

If you lower by 12 dB with 16bit recording, you lose 2 lowest bits of those 20 which were not part of 16bit recording. With 24bit recording you already lose valid bits of recording, but since 20 - 2 = 18bits still mean 108 dB of dynamic range, on usual listening volume levels it does not mean real loss.

Other situation comes when you lower digital volume significantly, for example to -48 dB. Then you lost 8 bits of resolution and that can be considered as audible loss on dynamics and low level detail. A setup requiring strong digital attenuation because of too strong amplifier is not suitable from S/N ratio view. Power of your amplifier should better fit your listening environment.

So in this case, when E50 is connected to PA5 - it's better to keep the E50 volume at (for example) -4.0 dB while turning the PA5 volume up/down?

 

[Nonick]

As for other, E50 requires 5V 1A power. If i use 5V 2A power supply (huawei quick charge or similar) - can this amperage damage E50 (or E50 will "pull" the amperage it needs, no more)?

 

[bogi]

So in this case, when E50 is connected to PA5 - it's better to keep the E50 volume at (for example) -4.0 dB while turning the PA5 volume up/down?

Yes if PA5 volume control is analogue. I don't know how it is implemented since PA5 is Class D.

If amplifier volume control is digital and if you need significant digital attenuation then you are really loosing quality. Such situation means you are using too powerful amplifier for your needs. Using an analog preamp could better the situation in such a case, because by lowering signal level in analog domain you are lowering also noise level, so SNR remains the same (I didn't count possible noise added by the preamp including analog pot).

 

[bogi]

As for other, E50 requires 5V 1A power. If i use 5V 2A power supply (huawei quick charge or similar) - can this amperage damage E50 (or E50 will "pull" the amperage it needs, no more)?

5V 2A means only that the power source is able to provide more current. It depends on the device how much current it needs. Under normal conditions (when the original power source was not underdimensed) the device current consumption should not raise. So the ability of power source to provide more amperage cannot kill your device. Only too high voltage could kill it.

 

[pma]

Yes if PA5 volume control is analogue. I don't know how it is implemented since PA5 is Class D.

Class D does not stand for digital. Class D amplifiers are analogue circuits, though they operate under sampling theorem - input voltage is converted to PWM duty cycle modulation. Duty cycle is not quantized, it is a (linear) function of input voltage. Sampled however not quantized.
There is a pot in PA5 to control volume.

 

[bogi]

@yanm Your analog setup seems to show that digital volume control is suitable for it in the extent you need it. It does not mean that more optimal analog setup than yours does not exist.

The original question from Mat was: does it degrade or affect sound quality in any way to push it down rather low only to reamplify it? Or is it completely fine at any decibel level?

At any level it is not suitable, as you showed in your 2nd example - you could count a case of a yet less suitable setup with higher gain amp. The simple general rule still is - the less digital attenuation you need to get usual listening levels the better for the result (being it audible or not). So my suggestion was rather general and was oriented into digital domain - it allows you to get more from your analog setup if it has suitable parameters.

Max. dynamic range (and thus SNR) of power amplifiers is usually lower than of DACs and high quality preamps/headamps. People listening on high quality headphone setup can easier benefit from high quality DAC output.

Some small level of digital attenuation is of course always suitable to avoid intersample overflows in DAC chip.

 

[yanm]

@yanm Your analog setup seems to show that digital volume control is suitable for it in the extent you need it. It does not mean that more optimal analog setup than yours does not exist.

The original question from Mat was: does it degrade or affect sound quality in any way to push it down rather low only to reamplify it? Or is it completely fine at any decibel level?

At any level it is not suitable, as you showed in your 2nd example - you could count a case of a yet less suitable setup with higher gain amp. The simple general rule still is - the less digital attenuation you need to get usual listening levels the better for the result (being it audible or not). So my suggestion was rather general and was oriented into digital domain - it allows you to get more from your analog setup if it has suitable parameters.

Max. dynamic range (and thus SNR) of power amplifiers is usually lower than of DACs and high quality preamps/headamps. People listening on high quality headphone setup can easier benefit from high quality DAC output.

Some small level of digital attenuation is of course always suitable to avoid intersample overflows in DAC chip.

Indeed, it is all depending on the system. It is also not always true that analog attenuation is better. If the noise floor of the preamplifier is higher than that of the E50 (which is quite low), then using that preamp to control volume will actually degrade the sound quality and one would be better off to use the volume control of the E50. All in all, it is why I proceeded with an example rather than a blanket statement. On the other hand, I fully agree that if you're using a headphone it is likely more appropriate to use the volume control of the (high-end) headphone amp.

Your last statement "Some small level of digital attenuation is of course always suitable to avoid intersample overflows in DAC chip" is not necessarily true. If DAC volume control is used, and if the DAC is of sigma-delta type (which, I think is the case for the ES9068AS chip used in the E50), then the volume control is often performed on the modulated stream (i.e., already oversampled signal). In that case, intersample overs still occur. If the digital volume control is done upfront of the DAC section, then intersample overs occurrences are decreased.

 

[bogi]

Indeed, it is all depending on the system. It is also not always true that analog attenuation is better. If the noise floor of the preamplifier is higher than that of the E50 (which is quite low), then using that preamp to control volume will actually degrade the sound quality and one would be better off to use the volume control of the E50. All in all, it is why I proceeded with an example rather than a blanket statement. On the other hand, I fully agree that if you're using a headphone it is likely more appropriate to use the volume control of the (high-end) headphone amp

I fuly agree. One can compute the range of suitable digital volume control for his system like you did.

Your last statement "Some small level of digital attenuation is of course always suitable to avoid intersample overflows in DAC chip" is not necessarily true. If DAC volume control is used, and if the DAC is of sigma-delta type (which, I think is the case for the ES9068AS chip used in the E50), then the volume control is often performed on the modulated stream (i.e., already oversampled signal). In that case, intersample overs still occur. If the digital volume control is done upfront of the DAC section, then intersample overs occurrences are decreased.

Sorry I was not clear enough. I meant digital attenuation in software player yet before the digital signal enters E50 or other DAC device. I already discussed the advantage of using digital volume control in software player over using it in DAC in order to avoid intersample overflows in this thread: https://www.audiosciencereview.com/...ing-e50-review-balanced-dac.26219/post-992240

E50 contains delta sigma based ESS chip, whose digital volume control is placed between the oversampling stage and the delta sigma modulator (that allows volume control of DSD input signal too - without downsampling it). Important is that volume control does not appear at beginning of digital processing. I don't know a DAC chip which would perform volume control at the beginning of digital processing chain. So yes, I agree with you, that DAC digital volume control is not the right place where you could avoid intersample overflows. And since the right place is your software player, from my point of view it is better to perform digital volume control only in software player and skip it in DAC.

 

[yanm]

I fuly agree. One can compute the range of suitable digital volume control for his system like you did.


Sorry I was not clear enough. I meant digital attenuation in software player yet before the digital signal enters E50 or other DAC device. I already discussed the advantage of using digital volume control in software player over using it in DAC in order to avoid intersample overflows in this thread: https://www.audiosciencereview.com/...ing-e50-review-balanced-dac.26219/post-992240

E50 contains delta sigma based ESS chip, whose digital volume control is placed between the oversampling stage and the delta sigma modulator (that allows volume control of DSD input signal too - without downsampling it). Important is that volume control does not appear at beginning of digital processing. I don't know a DAC chip which would perform volume control at the beginning of digital processing chain. So yes, I agree with you, that DAC digital volume control is not the right place where you could avoid intersample overflows. And since the right place is your software player, from my point of view it is better to perform digital volume control only in software player and skip it in DAC.

Well explained. By the way, this shows the main disadvantage (in my opinion) of MQA: even a slight software volume attenuation will kill the MQA signalling information so that the second unfold cannot be performed… I wonder if the MQA pipeline ensures that no intersample overs ever occur? Most likely we will never known as MQA is so obtuse.

Anyway, that’s only valid if MQA is used. Unfortunately, my case as I am mostly streaming off Tidal.

 

[bogi]

even a slight software volume attenuation will kill the MQA signalling information so that the second unfold cannot be performed…

Yes, only MQA capable software or hardware can do DSP with MQA content. Since not all SW and HW manufacturers are fans of MQA colonized world I expect this point will later lead to death of MQA format. The whole audio world is simply (I hope) not going to pay MQA company for ability to perform DSP.

Unfortunately, my case as I am mostly streaming off Tidal.

Use HiFi tear when possible. About a year ago it looked that Tidal will be fully MQA-ized. But later creation of Tidal HiFi tear seems to be Tidal reaction to growing dislike of MQA. I think Tidal does not want to be part of potential MQA fail and death without an alternative. I think MQA propaganda in long term cannot beat true lossless audio.

 

[MalinYamato]

This is a review and detailed measurements of the new Topping E50 balanced USB DAC. It was sent to me by the company and costs US $269. I just reviewed its companion headphone amplifier, the Topping L50.

The E50 takes my favorite orange LED display and makes it even nicer!

View attachment 150583

It is a gorgeous bright orange display. A single button turns the unit on and cycles through the inputs. You also get a remote control:

View attachment 150584

As you see, power is provided through a 5 volt barrel connector. Topping provides a USB to barrel cable adapter so you can use any USB power supply.

As with more recent Topping products, it uses TRS output for balanced output instead of XLR which takes up a lot more space.

Topping E50 Measurements
As usual we start without dashboard. Let's test the balanced output with USB input:

View attachment 150585

Another exemplary performance from Topping, landing it almost at top of all DACs ever tested:
View attachment 150587

RCA output is nearly as good:

View attachment 150588

Noise performance is excellent as well:

View attachment 150589

Linearity is as perfect as you can get:
View attachment 150590

IMD performance is excellent:
View attachment 150591

I see a hint of ESS dac chip IMD hump. I wonder if that is what powers it.

Multitone performance is superb with almost no frequency dependency:
View attachment 150593

Jitter over USB is perfect:
View attachment 150594

It degrades just a hair over Toslink/Coax:
View attachment 150595

There are three filers available:
View attachment 150597

They impact our THD+N vs frequency thusly:
View attachment 150599

Conclusions
What can I say but another star performance from Topping. Company is so focused on squeezing every bit of noise and distortion out of these products while providing them to us at bargain prices. And in this case, great looks as well. I wish I could complain about something but nothing comes to mind.

-----------
As always, questions, comments, recommendations, etc. are welcome.

Any donations are much appreciated using: https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/

did you recommend this dac? you did not say that you put it in your recommended llst!

 

[Mat]

I fuly agree. One can compute the range of suitable digital volume control for his system like you did.


Sorry I was not clear enough. I meant digital attenuation in software player yet before the digital signal enters E50 or other DAC device. I already discussed the advantage of using digital volume control in software player over using it in DAC in order to avoid intersample overflows in this thread: https://www.audiosciencereview.com/...ing-e50-review-balanced-dac.26219/post-992240

E50 contains delta sigma based ESS chip, whose digital volume control is placed between the oversampling stage and the delta sigma modulator (that allows volume control of DSD input signal too - without downsampling it). Important is that volume control does not appear at beginning of digital processing. I don't know a DAC chip which would perform volume control at the beginning of digital processing chain. So yes, I agree with you, that DAC digital volume control is not the right place where you could avoid intersample overflows. And since the right place is your software player, from my point of view it is better to perform digital volume control only in software player and skip it in DAC.

wait, what? From my understanding, using app volume is usually the worst form of volume control

 

[bogi]

wait, what? From my understanding, using app volume is usually the worst form of volume control

You didn't backup your understanding by any fact.

https://www.audiosciencereview.com/...review-balanced-dac.26219/page-35#post-989390

 

[Mat]

You didn't backup your understanding by any fact.

https://www.audiosciencereview.com/...review-balanced-dac.26219/page-35#post-989390

Neither did you, in fairness.

Are you saying that because an app is 64bit (like all apps in the modern era), it processes audio in 64 bit depth (and is thus better than an external device)? Is this only for (any) application, and separate from the OS system audio (which is in 16 or 24 bit usually)?

 

[bogi]

Neither did you, in fairness.

Are you saying that because an app is 64bit (like all apps in the modern era), it processes audio in 64 bit depth (and is thus better than an external device)? Is this only for (any) application, and separate from the OS system audio (which is in 16 or 24 bit usually)?

You overlooked that I mentioned floating point format in the linked post. DAC chip performs DSP operations in fixed point format because it does not have resources to run more demanding operations. As I stated above, each DSP is subject of rounding errors and therefore it needs to be computed in higher precision than the input data has. The rounding error propagates into next operation in the chain and grows. With floating point computing you get smaller rounding errors allowing more complicated algorithms or longer iterations to run without affecting required data precision. All well known software player like foobar2000, Jriver and others are performing all DSP operations in floating point arithmetics.

Other significant advantage of running DSP on computer is significantly higher computational power of current computers in comparison with resource constrained $10 chips. That means you can use better algorithms. Note that it is about real time computing with target sample rate as time limiting factor.

Remember that all the oversampling, delta sigma modulation and volume control performed in DAC chip are DSP performed yet in digital domain, before the D/A conversion itself and thus they require computational power. All these operations can be performed on usual computer running higher quality algorithms. There is no magic in DAC chip digital processing what could not be performed better on your computer.

 

[Mat]

All well known software player like foobar2000, Jriver and others are performing all DSP operations in floating point arithmetics.

Okay but does something like Apple Music count? Because it's audibly worse with something like Soundcheck on. Or do you mean only DSP players and not software in general?

All these operations can be performed on usual computer running higher quality algorithms. There is no magic in DAC chip digital processing what could not be performed better on your computer.

I assume it's not about raw power but a chip being purpose-designed rather than a CPU which just computes generally, otherwise we wouldnt needs DACs at all. Audio doesn't require much.

 

[bogi]

I assume it's not about raw power but a chip being purpose-designed rather than a CPU which just computes generally

If resources are missing, a chip cannot do wonders. Limited DAC chip resources affect for example the way how oversampling is done in DAC chip. It is filtered only up to some sample rate and above it very simplified oversampling is performed without proper filtering of high frequency content due to insufficient computational power to do so on high sample rates. No problem with that on a typical PC.

otherwise we wouldnt needs DACs at all. Audio doesn't require much.

I see you don't understand how delta sigma DAC works. I cannot explain here everything on few lines. You would need to understand why we use PCM format to store recordings, what for digital processing is performed in delta-sigma DAC chip (with PCM input), what is delta-sigma modulator, how it works, what appears on its output, why that output is easier to convert to analog signal. You can google on web and self study. I think it is better to stop here.

 

[Mat]

If resources are missing, a chip cannot do wonders. Limited DAC chip resources affect for example the way how oversampling is done in DAC chip. It is filtered only up to some sample rate and above it very simplified oversampling is performed without proper filtering of high frequency content due to insufficient computational power to do so on high sample rates. No problem with that on a typical PC.


I see you don't understand how delta sigma DAC works. I cannot explain here everything on few lines. You would need to understand why we use PCM format to store recordings, what for digital processing is performed in delta-sigma DAC chip (with PCM input), what is delta-sigma modulator, how it works, what appears on its output, why that output is easier to convert to analog signal. You can google on web and self study. I think it is better to stop here.

No need to get snarky. If you can't explain something sufficiently, don't tell others to do homework.

 

[Nonick]

Any Topping E50 owners: did you experience high frequency pitch noise from E50?
When i turn it on - high pitched noise is audible from cca 20-30cm.

 

[MalinYamato]

I love how the MQA logo is a sticker.

it is a politically correct sticker ... those who hate MQA may remove it and pretend there is no MQA in the unit, to keep their peace of mind.