"High-Resolution on a Budget: An initial look into the topping e30 ii lite dac, is this a true, budget dsd direct path dac as claimed?"

I was thinking one day I need a super cheap portable DAC for another baseline reference device in my reviews.  Not necessarily baseline measurements; it isn't difficult to make a DAC measure well these days.  I was thinking actual sound quality and how cheap could a device be before it was no longer enjoyable.  

I also wanted something with an actual Direct DSD path.  So ESS was out.  That really meant AKM, or Burr-Brown, and I already have plenty of iFi products around with the Burr-Brown DSD1793, so I chose an AKM product because I previously had a great experience with the AKM4493 in the RME ADI-2 PRO.    The AKM chip isn't quite as DIRECT DSD ala Signalyst or similar, that keep the DSD signal at 1-bit all the way to the FIR filter that converts DSD to analog.  In the Signalyst DAC, the filter itself becomes the digital to analog converter with shift registers, resistors and switches.  (What COULD have been the truest, most direct DSD DAC ever brought to market was the PSAudio Directstream because its filter is purely analog, not a digital filter implemented by analog components or some combo thereof. Unfortunately, like the ESS chipset, there is no way to bypass the quite massive DSP applied to both PCM and DSD formats as they enter the Directstream.)

The AKM chips with Switched Capacitor Filters are really, really good chips.  Then AKM had their terrible factory fire, and the newest chips are now outsourced and have moved away from SCF's to resistor based elements like the Signalyst, Burr-Brown, ESS, well, like a LOT.  It changes a LOT of things and I have seen lots of confusion in otherwise professional reviews on how DSD works in AKM based devices.  

Here is a quick rundown on how it works with the SCF chips like the 4493 (and presumably still kind of the same with their new resistor-based chips, but not quite the same as the other resistor-based chips from other brands.)  

​

​

From the block diagram of the AK4493 DAC, it is evident how the DSD data is processed in bypass mode and normal mode. Here's a detailed explanation of the volume control and delta-sigma modulation in the AK4493 DAC, based on the provided information and the datasheet.

Bypass Mode for DSD (DSDD1)

1. DSD Data Interface & Filter:
   • The DSD data interface and filter block handles the incoming DSD data. When the DSDD bit is set to '1', it indicates the bypass mode.
2. FIR Filtering:
   • In bypass mode, using AKM logic code DSDD1, the DSD signal is filtered using a FULLY DIGITAL Finite Impulse Response (FIR) filter. As mentioned earlier, many Direct DSD DACs don't use a fully digital FIR filter.  They use digital principles but are implemented with analog components, meaning the TAPS and the analog switches are the same thing.  In the AKM 4493 however, this digital filtering stage is implemented much earlier in the logic process, meaning the logic is fully digital, with digital TAPS all equally weighted at a value of 1.  Slightly different means which ultimately lead to the same end, ensuring that the high-frequency noise inherent in DSD signals is attenuated.  The fully digital filter outputs a multi-bit signal at the same sample rate as received.  There is NO further noise modulation, as step three will elaborate further.  This signal will be in unary code, and it could stay that way during its transmission to the Switched Capacitor Filters, or it could be immediately converted by digital logic to binary and then re-converted to unary code at a later stage before the Switched Capacitor Filters, which I believe is the most likely scenario.
   • Skipping ΔΣ Modulation:
     • The key aspect of bypass mode is that it skips the delta-sigma (ΔΣ) modulator. Normally, the ΔΣ modulator would convert the filtered DSD signal into another high-frequency pulse-density modulated signal with different characteristics.  However, in bypass mode, this step is omitted to preserve the original DSD signal characteristics as much as possible.  This is ,after all, pure DSD.  Or as about as pure as it gets.  (It has to be filtered somewhere, that cannot be avoided.  As long as it goes through no more DSP, it doesn't really matter if the filter is at the beginning of the chain or the end.)
3. Unary Code Output:
   • The filtered DSD signal is either in unary code or converted to unary code before it reaches the switched capacitor filters (SCFs). Unary coding is beneficial for reducing digital switching noise and improving linearity in the final conversion stages by allowing scramble code/dynamic element matching.  (It is THE standard for Delta Sigma DACS of any N-bit design.) 
4. Switched Capacitor Filters (SCFs):
   • The unary coded signal is fed directly into the SCFs, which perform the final digital-to-analog conversion. The SCFs average the high-frequency pulses to produce a smooth analog signal, effectively filtering out high-frequency noise and yielding a clean analog output.  This combined with the earlier non-decimating FIR filter with equally weighted taps create a very powerful tool for shaping the DSD signal.  Perhaps the most powerful on chip you will find.  

(Quick note for below... we are now describing a different process, how DSD is converted when the Bypass mode in NOT used, just in case there is any confusion.)

DSD Processing in Normal Mode (DSDD0):

1. DATT with NO Attenuation 
   • LOCKING the DATT volume control to 100% ensures that the signal's amplitude is not altered. This setting is equivalent to bypassing the volume control but allows the signal to pass through the volume control logic unaltered.  THE SAME WILL APPLY FOR PCM IN THIS MODE.  THE AKM DIGITAL LOGIC IS USED IN THIS WAY TO ACHEIVE FIXED OUTPUT MODE FOR BOTH SIGNALS.  ALSO NOTE THAT SIMPLY LOCKING THE VOLUME AT 100 PERCENT DOES NOT SWITCH THE SYSTEM TO LOGIC DSDD1 FOR BYPASS MODE!!!  THIS IS THE MISTAKE I HAVE SEEN MANY PROFESSIONAL REVIEWERS MAKE WITH AKM CHIPSETS! 
2. DATT WITH VOLUME CONTROL/ DSD Attenuation
   • The incoming DSD signal is received and initially processed by the DSD Data Interface and Filter. This block includes the necessary aforementioned FIR filtering.  The filter, either 1 or 2, in this mode is to manage high-frequency noise inherent in DSD signals, AND just as importantly in this case, to create a multi-bit signal that will be manipulable by a volume/gain control if and when needed.
   • The DSD signal then passes through the Digital Attenuation (DATT) block. Here, the volume control is applied to the DSD signal. This block allows for precise digital volume control, attenuating the signal as required. This step is crucial when volume control is desired for DSD playback.  The output of the 'Normal' path DSD filter was almost certainly converted into a binary code exactly equivalent in value to the unary code produced by the FIR filter with equally weighted taps.  This is because the binary code will allow for much more precise volume control and actually will require less overhead to work.  

------------------------------------------------------------------------------------------------
 ​FURTHER PROCESSING OF DSD NORMAL MODE

1. DSD Data Processing:
   • When the DSDD bit is set to '0' (normal mode), the DSD signal does not bypass the ΔΣ modulator. Instead, the DSD signal undergoes the standard processing path, which includes noise shaping and modulation by the ΔΣ modulator.
2. Delta-Sigma Modulator (ΔΣ Modulator):
   • The ΔΣ modulator remodulates the filtered DSD signal into a multi-bit Delta Sigma signal.  This process helps shape the noise, pushing it out of the audible frequency range and improving the overall signal quality, potentially more-so than a single bit DSM system, as a multi-bit DSM system has much less quantization error to shape.  
3. Switched Capacitor Filter (SCF):
   • The modulated signal, if not already in unary code is converted to unary code, undergoes Dynamic Element Matching/code scrambling, and is then fed into the SCF, which performs the final digital-to-analog conversion. The SCF averages the high-frequency pulses to produce a smooth analog output, effectively filtering out the high-frequency noise.

FURTHER PROCESSING OF DSD BYPASS MODE

1. DSD Data Processing:
   • When the DSDD bit is set to '1' (bypass mode), the DSD signal bypasses the ΔΣ modulator.  It has already been a delta-sigma signal once before, nor has it been touched by any digital volume control or a redundant remodulation as has the DSD bitstream in normal mode described above.  Also, it has already been filtered by a digital FIR filter without decimation, so it is simply ready to be converted directly into analog.  
2. Switched Capacitor Filter (SCF): 

The filtered oversampled multi-level DSD signal is then fed into the SCF, which performs the final digital-to-analog conversion. The SCF further filters the high-frequency pulses to produce and even smoother final DSD signal, and it the process concerts the signal from digital into analog.  


SOME CONCLUSIONS

Benefits of Using the Normal Path for DSD with Volume Control:

• Consistent Volume Control: Applying digital volume control to DSD signals allows for consistent attenuation across both PCM and DSD formats.
• Enhanced Noise Shaping: By passing the pre-filtered DSD signal through the ΔΣ modulator, the DAC can effectively reshape the quantization noise, pushing it further out of the audible range and improving audio quality.  Since the remodulation is done with a multi-bit quantizer, this allows for greater consistency between DSD speed formats and is much better at handling the ultra-sonic quantization noise.  
• Flexibility: This setup provides flexibility in managing volume levels digitally while maintaining high fidelity in the analog output.

By using the normal path (DSDD = 0), the AK4493 DAC can apply volume control to DSD signals, ensuring that users have the flexibility to adjust playback levels while benefiting from the advanced noise shaping and modulation techniques integrated into the DAC.

Benefits of Using Bypass Mode of Volume Control and Modulator:

1. Preservation of DSD Characteristics:
   • Bypass mode allows the DSD signal to maintain its original 1-bit, high-frequency characteristics, dependent on the quality and parameters of the pre-filtering. This can be important for purists who prefer the unique sound quality and characteristics of DSD audio, which can be altered by further digital processing.
2. Reduced Processing Complexity:
   • By bypassing the ΔΣ modulator, the signal processing path is simplified. This reduction in processing stages can result in lower latency and fewer opportunities for digital artifacts to be introduced into the signal.
3. Lower Power Consumption:
   • Skipping the ΔΣ modulation stage can reduce the overall power consumption of the DAC. This is beneficial for battery-powered devices or applications where power efficiency is critical.
4. Direct Digital-to-Analog Conversion:
   • The DSD signal, after FIR filtering, is converted directly to analog using the Switched Capacitor Filter (SCF). This direct path can result in a cleaner and more transparent signal path, which some audiophiles may prefer.
5. Simplified Signal Path:
   • A simpler signal path with fewer stages can enhance the overall reliability and stability of the DAC operation. Fewer processing stages mean there is less chance for signal degradation or synchronization issues.
6. High-Fidelity Playback:
   • For high-resolution audio playback, preserving the integrity of the original DSD signal can yield a more accurate and high-fidelity sound. This can be particularly noticeable in high-end audio systems where every detail of the audio signal is critical.

Use Cases for Bypass Mode:

• Audiophile-Grade Audio Equipment: High-end DACs used in audiophile-grade audio equipment often prioritize maintaining the purity of the original audio signal. Bypass mode is ideal in these scenarios.
• Battery-Powered Devices: Portable audio devices that rely on battery power can benefit from the reduced power consumption in bypass mode.
• Minimalist Design Approaches: Audio systems designed with a minimalist philosophy, aiming to use the least amount of processing possible, can leverage bypass mode to achieve their design goals.

Conclusion:
​Bypass mode in the AK4493 DAC offers a streamlined and purist approach to digital-to-analog conversion for DSD signals. It preserves the original characteristics of the DSD signal, reduces processing complexity, lowers power consumption, and provides a simplified signal path that can be beneficial in high-fidelity audio applications. This mode is particularly suitable for audiophile-grade equipment where maintaining signal purity is paramount.

---

---

SO THEN, ANDREW, WHAT WAS THE BIG DEAL? WHY DID YOU CALL OUT SOME ONLINE AND PAPER MAGAZINES FOR SAYING THAT A DIFFERENT TOPPING PRODUCT (THE E70V), WITH A TOTALLY DIFFERENT AND NEW AKM CHIP,  INDEED ALLOWED ACCESS TO THE PURE DSD BYPASS MODE?  (WHEN IT OBVIOUSLY DOES NOT.)

We will leave aside the fact that it's a totally different chip for later, but getting ahold of this Topping E30 II Lite has shed a bit of light on the 'controversy'.  You see, the advertising propaganda for the E30 indeed says that it offers the true DSD BYPASS mode.  It instructs its users to simply put it in FIXED OUTPUT mode, and the DSD will not have a volume control and therfore will bypass the internal modulator.  This is both stated and implied.

In the case of the Topping E30 II Lite, that MAY indeed be the case.  I have spent hours cooking up multiple tests to sniff out the truth, but the hard facts are, no matter in Fixed or Variable Output, everything measures EXACTLY the same!!!  And knowing how direct bitstream DSD interacts with analog output stages in a very different way than non-direct DSD that take full advantage of the performance gains offered by multi-bit Delta-Sigma noise shaping, my Spidey Sense is up.  But I can't find as of yet a true smoking gun with this particular AK4493 chip in this particular Topping E30 II Lite DAC.  The jury is still out, but my opinion is that NO, it doesn't use the bypass mode at all when you lock the volume control at 100 percent (no attenuation on either DSD or PCM).  

That brings me around to the products I reviewed with the latest AKM dual chip AK4191 + AK4499.  I had my first experience with this very different AKM chip in the Topping E70V Velvet. The controversial one. The thing about this chip or chips, is they are VERY different from the more well known and highly regarded AK4490, AK4493, etc, which were all based around switched capacitor conversion, and AKM were the MASTERS at it.  Then comes that dreadful factory fire, and things really changed.  Not only were a lot of our chips now being outsourced, AKM switched (no pun intended) from what they do best in Switched Capacitors over to Switched Resistors.  Really, this is a whole new ball game.  

And now for a little speculation.... in the past perhaps it was a common practice when using the AK449x chips to activate the DSD bypass mode when also 'deactivating' the Volume control for full fixed output across formats.  Makes total sense.  But this has to be programmed in the chip logic to happen that way.  It is two different actions.  And they absolutely do NOT have to be performed at the same time.  When I reviewed the Topping E70V Velvet, I got the same Spidey senses I mentioned with this Topping E30 II lite.  The two modes, volume control on, and volume control fixed or 'bypassed' measured exactly the same.  Once again, not a thing in the measurements to suggest this had two different paths for DSD conversion.  It certainly still could have been the case, so I messaged Topping directly and they directly got back to me and said in no uncertain terms that 'NO', the E70V does not offer the bypass mode.  

And for more confirmation, the other product I have reviewed with the AK4191 + AK4499 chipset, the SMSL D400, actually has a THIRD entry under the menu that specifically has a selection for 'DSD BYPASS MODE', along with the other two modes, that simply determine whether the DAC is used as a pre-amp with volume control, or as a DAC only with fixed volume output.  You want BYPASS MODE DIRECT DSD?  No other way to do it except to select that particular, unambiguous option.  Just selecting to use fixed volume control will not cut it.  

And remember, the technology in THIS multi-chip AKM 4191 + 4499 DAC is totally different than previous AKM DACs, and a 'deep', well not so deep dive into the SMSL version's measurements shows massive differences in the filter behavior and overall performance characteristics that I was fully expecting to see in a DAC that actually has two different DSD modes available to activate.  So, there is NO DOUBT about those two DACs.  The Topping E70V?  NO PURE DSD BYPASS.  The SMSL D400?  YES, YES, YES it has the PURE DSD BYPASS OPTION.  (And did I mention this was entirely new tech for AKM that differs pretty massively from their bread and butter?  Yeah, it needs some firmware work and let's leave it at that.)

But this little Topping E30 II Lite?  I am 90 percent sure it does NOT allow access to the DSD Bypass mode in spite of advertising it prominently as a feature.  Surely no  company has ever gotten something wrong, exaggerated, or just flat out lied? 

And as I have thought about it some more, considering this is a super small, super cheap device that costs less than most 2 meter RCA interconnects these days, why even SHOULD it have the extra logic programming to do something it doesn't need to do?

Because it measures admirably well in both PCM and DSD modes, both fixed output and variable volume output.  DSD measures identically in either output mode.  And the actual filtering they are using on DSD is EXTREMELY gentle, which allows one of the biggest strengths of DSD to shine out, and that is the transient response.  Also, it allows enough ultrasonic noise to enter the ears, and even though we cannot hear it, that ultrasonic quantization noise, unlike random PCM quantization noise, stays harmonically related to what we can hear.  Psycho-acoustic experts theorize that this plays a big part in why DSD sounds so 'good' to many people.  It goes beyond our basic hearing and how the noises are processed in our neural networks.  And that is where all the REAL work is done!  Between the ears!  And, well, with the ears too.  This blog entry has gotten way to long already so I will save more info on why DSD could sound better for another day.   

And finally, I am back to pondering the fact that this is a cheap product in which there is no way it has the ability to articulate the minute differences that might exist between a pure DSD bypass mode conversion and one that decides to not take the bypass, yet would rather taxi right on into the Modulator City.  

I will have a more proper review soon, locatable under the 'review' tab you see above.  It won't go over all this stuff again; I will just link to it where appropriate.  But now a preview of the review.. The Topping E30 II Lite is a good sounding little product for the price, and measures way too good for the price.  See you on the other side of that review!

---

---

IMPRESSIVE PERFORMANCE FOR $99 US. These measurements are achieved with the E1DA COSMOS APU AES-17 Hardware notch as a pre-amp for the E1DA COSMOS ADC. For 1khz distortion measurements with proper REW frequency response compensation, I have no problem saying it matches anything that a 20 grand Audio Precision tester can do, on this one particular test! THD is -119.6db, and SINAD is a very impressive 116.2dB. I don't want to give away too many of the measurements, but the overall dynamic range also is quite impressive as it reaches over 121dB A-weighted. Full review coming soon!