I suppose I would of thought this amount of upscaling was obscene a couple of years ago, but reading about and auditioning Chord’s upscalers has changed my views on this. It is a low computational overhead for PCM and results in a signal that is more faithful to the original analogue signal.
For me it is not about more detail, but realism. Before lockdown myself and my wife would try to go to live music performances as often as possible and this resets the brain as to what a good hifi should sound like. In my opinion upscaling with HQP gets me a sound that is closer to a live performance. It is not about the amount of upscaling applied, but the quality of the filter used. I would agree with Rob Watts’ argument that the timing of transients is critical for creating a realistic sounding signal, and to get accurate timing you need the best possible reconstruction filter to reproduce the signal between the digitised samples.
Antipodes has a similar design philosphy in that they do not overdo the noise filtering as this can remove high frequency signal that is critical for transient perception. This is why Innuos servers sound lush and warm but not as realistic as Antipodes (my opinion, based on owning both). The innuos is great for laid back genres but is not a good all rounder. If I want to tune my setup for different genres I would rather do that by choosing different headphones and/or amplifiers instead of adversely damaging the HF signal in the server.
I think the mathematical theory that underlies the filter design in Chord DACS and HQP software is sound. You can test by taking a native signal with a high sample rate, delete intermediate samples and then try to reconstruct the known missing signal with different filters. Note that all DACS do upscaling, its just a question of how well the reconstruction filter represents the analogue signal. Even the so called Non Oversampling Dacs do a rudimentary upscaling using a ’sample and hold’ approach. It can be mathematically proven that this does a poor job of reconstructing the intermediate samples.
In summary, if you have not heard upscaling done well, I would recommend trying again with either HQP software or a Chord M scaler.
Indeed I use HQP and Antipodes Solution, however I differ on the DAC front. I have an Aqua La Scala MK2 Optologic DAC. No it doesn't upsample as high as Chord but still sounds fantastic. I did however find that can converting to DSD, DSD7 and Sinc - M has better realism
than PCM Sinc- M, LNS15. I guess this all depends on our DACs