If you have a really dense (or compressed/clipped/loudness war) recording, it is even better to stick to higher rate DSD output. At least DSD256 or even better DSD512, using the EC modulators.
PCM conversion, even at 1.5M, can become congested and lack “depth and air”.
If it’s 44.1/48k source, use of an apodizing filter is typically must. Also good to stick to some shorter filter, such as poly-sinc-short. And possibly minimum-phase.
Another approach is to use longer one, but one that is especially transient optimized, such as poly-sinc-gauss family.
I’m nowadays very happy with
1x=poly-sinc-gauss-long
Nx=poly-sinc-gauss-hires-lp
So much that I have made this the new default in latest release.
I normally would but the specific filters I’m thinking of adding are $200 (LCD-5 FIR Filters from Mitch Barnett). More details in the link below, I’d like to be relatively sure before placing the order though (scroll to the bottom for details on the filters)
What you just said confirms my understanding though. I’ll only be applying the convolution to PCM (not any native DSD which would need to convert to PCM first at 1/16 DSD rate, which could be CPU intensive). No multichannel either.
He offers the filter up to 384k and will split the L&R mono as well.
Yes, it’s not great on the LCD-5, although I did use it on my LCD-X and enjoyed it. Crinnacle was even worse. I don’t think these are meant to be tuned to Harman
Resolve v2 is what I’m using now and it’s pretty nice. Most owners are using his PEQ or a modified version of it tailored to personal preferences. The handful of people who have bought the convolution FIR filters from Mitch Barnett have been impressed with the results.
I go back and forth on if I prefer PCM over DSD and usually PCM wins. But listening now to your suggested settings and damn if it isn’t sounding very good right now.
If your output is DSD, then DSD sources wouldn’t get converted to 1/16th PCM, but instead convolution is performed at the source DSD rate. This can get heavy.
If your output is PCM, then the DSD-to-PCM is usually not very heavy, but depends on your conversion settings though. In that case, convolution would happen 1/16th rate in PCM which is usually not particularly heavy.
Psycho Killer (Live LP version) and for any similar circumstance where all I want is every last drop of transient slam I can get… poly-sinc-short-mp
long-gauss, which sounds great but by comparison sounds like someone pushed the loudness button, cannot compete…
Sorry for the delayed acknowledgement as I had been busy trying to find a place to move and settling in.
Finally, relaxing with new convolution files I was sent and wanted to find out if someone could direct me on the difference between using 192 vs 384 mono wav files. My X26Pro I believe is set to 32 in the settings as well as Roon.
Also, if adding a bass shelf of 4db the already add convolution matrix, is it ok to still add the shelf to the matrix pipeline with a -4db gain adjustment?
From my understanding you should use the highest possible available. I went with the 384k (even if the highest PCM source you are playing is 192k there is no harm in this)
@jussi_laako :
I was reading this: https://www.pttweb.cc/bbs/Headphone/M.1637516070.A.EA0
And I was wondering how to choose the best setting on HQPlayer for my Pegasus (Pontus like for signal processing).
I’m going through an iFi Zenstream, so limited to DSD256x48.
I’m still a fan of poly-sinc-gauss_long or hires but for the shaper,? “…Denafrips products use FPGA to convert 1-bit DSD to 6-bit (7-level) DSD…”
Still ASDM7ECv2… or AMSDM7 512fs which seems more " fluid ".
Given the architecture described by Denafrips, the “NOS” setting does not come into play for DSD (?)
NOS setting on the Pontus II is only used for incoming PCM.
The manual says:
"3.5 NOS/OS
The PONTUS allow the user to change the sampling mode on the fly.
NOS, as the name suggested, does not over-sampling to digital input data.
In OS mode, the PCM 44.1kHz or 48kHz based audio data are up-sampled to the maximum rate of PCM1411.2 or PCM1536. There is no up-sampling of DSD audio signal.
3.6 PROPRIETARY R-2R AND DSD DECODING ARCHITECTURE
The PONTUS is equipped with 24Bit R-2R DAC to decode PCM data stream and 32 steps FIR analogue filters hardware decoder to decode DSD data stream. These designs guaranteed the PCM format can be perfectly decoded, at the same time, the DSD format can be perfectly decoded as well. It is rare in the currently market that a R-2R DAC can hardware
decode both the PCM and DSD formats."
I don’t fully understand the 32 step DSD part, anyone?
Its a 32 step FIR (Finite Impulse Response) filter, which in theory is how @jussi_laako DSC-1 works for DSD conversion as well. (Direct 1-Bit DSD, 32 tap FIR converter) but Denafrips R2R DACs are a bit of a black box so who knows what the secret sauce is
I like the content on a low-bitrate stream PlanetPootwaddle (128k MP3) and continue to be impressed with poly-sinc-mqa/mp3! It makes it more than listenable
With other filters the stream can sound quite gritty and crunchy in the upper frequencies. Would I like a higher bitrate, of course, but this is all they stream…
Thank you Jussi for all your talent, work and persistence improving my listening experience so much.
Could I get some recommendations on what filters to use for a DAC that is using TI PCM1794A chip?
I have a DAC with this chip made by a local hi-fi components maker, so very few people have a DAC like that and probably none of them use HQPlayer
It can do 24/192 PCM maximum and I am feeding it with RPi4/Allo Digione Signature acting as NAA endpoint to the DAC’s AES/EBU input. I haven’t tried all the filter combinations yet, but I keep using and coming back to 1x/Nx poly-sinc-ext2 and NS4 combo.
