Jussi, you posted this a few years ago, but can I ask if this rule of thumb also applies to higher input rates? With DSD 256 or DXD source files, do they need to be upsampled to at least DSD 512 when using digital volume control in HQPlayer?
Yes, from technical perspective that gives you headroom over what the source content could technically have. Practically analog and acoustic noise floor of the recording source is limiting factor rather than technical capabilities of something like DSD256.
So from theoretical perspective DSD256 should be converted to at least DSD512. And DXD to at least 705.6k. From practical perspective it doesnāt make much difference for DSD. Since DXD is just 24-bit, you can alternatively convert to 352.8k 32-bit to have extra headroom. Many DACs support this. Note though that on macOS getting 32-bit output needs integer mode support.
Many thanks - especially for the 32 bit upsampling suggestion, which hadnāt occurred to me. I normally convert DXD to DSD, but Iām limited to 256 output if I keep using the EC modulators in HQPlayer. Iāll do some listening comparisons with volume control between DSD256 with EC against DSD512 (and 32 bit 352 with EC). With my power amps I need to have the volume control in the range of -32 to -24.
I have the T+A DAC 8 DSD, and have always used the default setting for bit output: would this be 24 or 32 for that DAC?
That is certainly best option.
32-bit for that one is the default. But it is not worth sending PCM to it, so it doesnāt matter anyway since DSD is always ā1-bitā.
Thanks. One more question: should use of digital volume control (-30db) in this context influence choice of particular filters and modulators?
You get technically best results if you use for example ASDM7EC and set DAC8 DSD analog filter to ācleanā (instead of āwideā). But you can try and compare yourself which way you prefer. You get good results in other ways too. Choice of filter doesnāt matter in this respect.
@jussi_laako, Iām trying to use HQPlayer for digital volume control on files which are already up-sampled to 768Khz with a different software (PGGB). Iām using a USB-to-SPDIF converter between my DAC and PC that can accept 24/768 but not 32/768.
Given that the upsampling is already done prior to HQPlayer, should I be using digital volume control in HQPlayer?
And if yes, is it better to do 24/768->24/768, or 32/768->24/768, when using volume control? Iām using the LNS15 noise shaper.
In fact, if HQPlayer supported Sony Wave64, I could even do 64/768->24/768, avoiding noise shaping in PGGB entirely. However HQPlayer does not appear to support Sony Wave64.
(downside of PGGB is the continuous ringing it introduces, but that is OT)
It doesnāt matter if you have prior upsampled material or not, volume control is independent of this. So yes, you can use volume control just fine.
More resolution the source can have, the better. So certainly 32-bit is better.
I donāt know what is Sony Wave64, but RF64 WAVs are supported already. You can also use WavPack which will save you some space.
Thanks! When converting 64/768->24/768 from a PGGB-upsampled source and using volume control in HQP, is the HQP Filter setting used at all, or should it be set to ānoneā? I have set the noise shaper to LNS15.
Yes, filter setting is used in all cases. So if you want to turn it off, you need to select ānoneā.
Is it reasonable though to use noise shaping (going 64f/768->24/768) without any filter, when using HQPs digital volume control? Of course the original data was already upsampled with a filter by PGGB, but would the use of volume control in HQP necessitate further filtering in any way?
No, these two things are separate and 768k rate is enough for noise shaping.
On this topic, upsampling to 768Khz in HQP and using LNS15 noise shaping, would you expect any loss of quality at all at -15db? How about -20db or lower? Is it DAC-dependent? Or is there always a loss of quality with attenuation, just more noticeable as the attenuation gets more severe?
No loss in quality. The only limiting factor is analog noise floor of the DACās analog output stages. Quality limitations are not at the digital side.