For sure that it is
!
A lot of people love Chord, but to me they are inferior to Sony. The TA-ZH1ES does upscaling to 384Khz and that is mightily impressive already, Even though they are building around FPGA too, but Chord is stuck after the FPGA where as Sony is carrying onward with digital amplifications.
As far as people enjoying Chord Upscaling features, I have no clues why they don’t even look at Sony products.
Yeah, may be using Neutron player in AndroidOS, other players can also upscale stuff upto 384Khz or 768Khz, but once again, Sony is the pioneer of DSP. So, beside the rating number as an objective to compare, they are very different.
If DSP is a thing then the new Walkman advances over the last Gen was about DSD remastering. It took Sony several years to develop this for the portable S-Master chip, which they were not able to squeeze into the DMP Z1
I agree totally on this bold part.
Chord's future audio performance is going to be limited by the FPGA processor performance development and the creative ability of Rob Watts to formulate better performing filter algorithms.
Also Chord's upsampling is a blanket approach, applying the same upsampling algorithm across the entire sound specturm as compared to Sony's DSEE Ultimate more nuanced sound source separation approach which utilizes AI to do real time seperation of the individual instrument types and vocals and applies targeted differential upsampling to individual separated sources before recombining them in real time back into a single upsampled sound.
Sony's AI is also a self-training algorithm that gets better as more data is fed into the AI(Meaning the AI is forever learning to get better with time by means of Machine Learning). AI don't require expensive high performance FPGA to run, they can function out of Lower end smartphone processor and GPU like the Quad core 1.8GHz NXP Processor found inside the Android M2 Walkmans.
This comment post on the Goldsound's youtube review of the M-scaler perfectly describes the issue with Chord's inferior approach vs Sony's AI approach to Upsampling
11 months ago
Upsampling has only one purpose : To enhance anti-aliasing in such a manner that it can preserve higher frequency detail. Upsampling using filter taps is being superseded in the imaging domain by Machine Learning methods. You can bet your bottom quid that in audio it will be coming fast. Computationally-speaking, ML to this level of detail can be done on a cheap smartphone processor. Companies like DCS and Chord fixated on classic DSP techniques are in for a beating, much to the benefit of consumers, professional or amateur alike. D-to-A conversion HAS to be a commodity, because while critical it is only a step in a broader logistics chain. The fact that it currently harbours a niche market of tech vendors is a temporary thing in the grand scheme of things, and it isn't true that commoditisation will result in a lessening of quality ; that would be like saying your WiFi modem needs a class A triode valve circuit so the bits it delivers are smoother.
Show less
https://www.youtube.com/watch?v=NSRMEIEY2oE
TLDR Summary:
Sony's AI Upsampling is more adaptive, will continuously improve(via firmware updates) in performance and sound quality over time.
Chord's Upsampling approach is fixed function, more expensive and is limited by semiconductor development.
More on Machine Learning: