AK4490 and AK4493 are back in production. AK4499 also has this functionality. In addition Cirrus Logic and TI DACs offer this. In fact TI/BB DAC chips don't even offer "DSP path" for DSD, it is always direct feed to the D/A stage. There are of course many AKM based DACs on the market, not just RME. I have also for example TEAC NT-503.
And then you of course have number of DACs with a discrete DSD stage on the market. (T+A, Holo, Denafrips, TEAC, Esoteric, Marantz...)
I too saw that the 4493 is back in production, and I’m glad to hear AKM is slowly rebuilding what they had—I suppose we need a little friction with ESS to encourage competition for producing some new DAC chips that exceed the specs of the ES9038pro and the AK4499. But I can’t say I’m excited about ongoing access to the 4490 or 4493, or the DACs that utilize them.
I’d much prefer AKM to get the AK4499 up and running again, or a worthwhile equivalent (the last I read AKM had no plans to reboot the 4499). The reason they’ve prioritized their inferior 4490 and 4493 chips instead is pure economics—there are far more mid-tier, affordable devices on the market with these two chips than the 4499-based flagships, so demand for them is far greater. Either way, Topping replaced the 4499 with the 9038pro in the D90se and it’s outperforming its predecessor.
Going back to the original topic, my question is what there is to be gained on a practical level by these 48-base or direct bypass DSD units. A lot of the devotees of these technologies seem to be more focused on the engineering involved than on the sound coming out of the speakers. The DACs you mention all perform with considerable inferiority compared to the modern delta-sigma DACs, and many of them aren’t just DACs anyway—they’re combination pre-amp, amplifier or signal processor devices. Any arguments for them seem to be based in a bygone era technologically
The measurement differences among the current top performing DACs are unlikely to be audible among anyone over the age of 10 years old. And Amir has consistently debunked the myth of the R2R DACs since we passed into the modern delta-sigma era over the last couple of years.
Well, to be exact, it is not a pure DAC. It has bunch of DSP processing and is not able to do bit-perfect D/A conversion where the actual input data ends up untouched in the D/A conversion stage.
It has at least following DSP:
- MQA decoder and digital filters
- Digital oversampling filters
- Digital volume control
- Asynchronous sample rate converter
- Delta-Sigma modulator
Sure, it has less DSP functionality than the RME. But on the other hand, the RME is able to also operate in bit-perfect D/A conversion mode with DSD inputs (but only the AKM chip versions).
Well, pretty much all delta sigma DACs have those filters (that’s the foundational basis of how they work), so I wouldn’t necessarily say they make the D90se anything other than a DAC—it’s a pure delta sigma DAC. I think that this kind of thinking is yet another example of focusing more on what’s under the hood than what quality of sound is coming out of those analog outputs.
I have a ton of DSD/DSF files I’ve purchased from sites like nativedsd and others, along with a few hundred I’ve used JRiver to convert from SACD iso into quad DSF files. And the Topping decodes these files as DSD 256 with extraordinary results. I have a hard time believing that at this sampling rate, with quantization noise being pushed up above what moths can hear, that “DSD direct bypass” would sound superior in any appreciable way.
We can keep debating these granular differences in processing among DACs, but the bottom line is that all the arcane methods to decode digital files you’re referring to, along with any didacticism about the “purity” of R2R ladders, have been rendered obsolete in the last year or two—in what Amir calls the “modern era” of DACs. The eight channel ES9038pro implementations that have been arguably perfected over the past year have consistently demonstrated superiority over R2R DACs and the “boutique” examples you’ve mentioned, by every accepted criterion.
So, say what you want, but when evaluating the pure functions a DAC performs, including the fairly universal filters inherent to the delta sigma DACs that sit at the top of Amir’s SINAD charts, measurements are the only objective methods we have to determine the sound quality of the digital to analog conversion process—and when performed with rigor, proper controls and standardization, they are ALL that matter. With the exception of the RME, R2R DACs don’t even approach the performance of the prevailing delta sigma DACs like the D90se.
And if you read Amir’s review of the RME, he gave it a strong recommendation based on its combined strengths as an amplifier, signal processor and DAC in one—not on its measurements as a DAC alone per se. It’s a great unit for what it is—but it’s in no way a better DAC than the D90se. And if all you’re looking for is a DAC, the Topping is a way better deal from a cost perspective.