HQPlayer -vs- Chord WTA1

Ive been doing a lot of nightly listening sessions comparing the sound of Chord Electronics Hugo MScaler with HQPlayer filters - both setup as PCM. I have a well tuned DAC-direct loudspeaker setup in a mid-field listening room. I run Audiowise, so I used all my RF isolation kit to maximize transparency.

There is sentiment that the extreme length of the sinc filter used by MScaler is extreme overkill and introduces some false euphonics. Not everyone agrees what sounds best. Still, there is good market success of Chord DACs and credible arguments from Rob Watts that it takes a long sinc filter (and high order noise shaper) to properly recreate the original analog samples and preserve all the depth and staging of the recording. Myself, i can listen to MScaled Chord DACs all day and find great enjoyment in how they sound.

Its been great to have a system that allows me to conveniently switch MScaler/HQP with extreme RF isolation to ensure that all I am listening to is the filtering technology behind each - no conducted or radiated noise to obfuscate transparency.The MScaler is a digital-to-digital device and the internals are just mathematics for a single filter implemented in an FPGA. Jussi’s excellent HQPlayer filter engine has a broad number of variations - all implemented at the utmost precision possible. Moreso, HQPlayer’s technical implementation is so optimized that i’m incredulous how much PCM performance is extracted from ordinary hardware.

Many posts on this thread relate to a choice of HQPlayer PCM filter based on personal preference - and all probably well considered, too. But when it comes down to finding a HQP filter combination that comes closest to the sonic signature of MScaler, it is, IMO, ‘Sinc-M’ filter and ‘LNS15’ noise shaper. Sinc-M’s ultra-long filter may slightly accentuate the ‘recording venue echo’ versus ‘poly-sinc-long-lp’ or ‘poly-sinc-ext2’ (Jussi’s fav) - but i find that Sinc-M energizes the room better and ‘feels’ more effortless. When combined with LNS15 (15th order shaper - thanks Jussi) the depth increases and placement of artists is better. My A/B switch between MScaler and HQP with these settings actually indicates a preference for HQP - and I am being as objective as possible.

Many thanks to Jussi for his great work. I’d be very interested in others who have Chord kit and extremely resolving headphone setups who can corroborate my impressions.

Thanks
Dan

7 Likes

I ended ut buying Chord M-Scaler to pair with my Chord DAVE just before the holidays. I haven’t recently listened to HQPlayer but decided against using it some 1,5 years ago after liking the DAVE better on its own. For me the benefits of M-Scaler are clear, not mind blowing, but definitely there and nice. With HQPlayer it always felt like a bit give-and-take. Probably been some more development since I last experimented with it.

So you are the perfect candidate for feedback! If you are already ROON->USB->MSCALER->DBNC->DAVE then its easy to reconfigure your playback to support a path through HQPlayer (MSCALER will pass thru native 705.6/768 thats been upsampled by HQP). I’d appreciate a comparative listen with the settings i described. Thanks.

Maybe if I find some time I will try. Right now I’m more in the “enjoying the music” mindset than the “tinkering”. Also I’m not sure I have a computer at home that would be able to do run HQPlayer to satisfaction since all my GPU:s are AMD (except the integrated ones).

For PCM upsampling you don’t need anything extraordinary, or necessarily GPU…

It is different thing to run filters to 23 MHz than to 768 kHz…

HQP has allowed me to move my HugoTT back to the primary listening position. I feel that with more than 500k taps (sinc-m) at 384k using the new LSN15 shaper it has taken the musicality beyond my H2 and M Scaler. The TT’s super caps also drive the Utopia better than the H2 imho. I am using a Nuc7i3 as the endpoint through usb to the TT.

1 Like

I was always meaning to test HQPlayer against Churd Hugo 2 internal filter and this was a good post to find. I only have a copy of HQPlayer 3, not 4 but I downloaded a copy to test the new LNS15 noise shaper. I believe this is the key. With just N9 it sort of sounds better than the internal filter but wasn’t as clean. Using the LSN15 it’s very clean and liquid sounding. Real improvement over the the internal. I’ll keep testing before I shell out money for the new version. Didn’t think investing in the m scaler made sense for the Chord Hugo 2. Would rather spend on used TT 2 or Dave.

2 Likes

Poly-sinc-long lp made me upgrade. Maybe a minuscule less detail than sinc-m but a more relaxed listening for extended sessions. If I had known that HQP would sound as good as or better than mscaler I would have spent my money on TT2. Right now mscaler is in a drawer. If I don’t feel the need to pull it out my take the hit and sell it.

Sounds like something I should try then. I’ve always liked the idea of going more minimal (but somehow always ended up with more boxes).

The M-Scaler “sounds” great but has not found a visually pleasing place to sit yet. Maybe I’ll be playing with HQPlayer again next weekend to challenge the Chord.

This is good to know.

Funny thing is I went with Chord dacs to get away from HQPlayer since I liked the xtr filter. Luckily my current setup makes it pretty easy to use and can quickly switch between HQPlayer and roon direct with no oversampling for testing.

Let us know your impressions.

For some reason I couldn’t get the poly sinc long lp to work. Not sure if needed to wait longer for it to start but then if have that much delay on start not worth it. Sinc-m only has 2 second delay.

So far sinc-m is much better than Chord Hugo 2’s internal filter. It has mostly similar sound but more bass impact, big soundstage, more detail, and more liquid sound. Just feel you can hear further into the music. This is with the LSN15 shaper. When I tried before with N9, thought it changed the sound too much from the Chord sound. Not sure I’d stick with that over stock.

Make sure you have the [Adaptive Output Rate] checked. Delay is only about a second with sinc-M.

The mode of transmission (usb v dual-coax), the sample rate (384/768k) and the filter algorithm used (WTA, HQP,…) are all independent. No mystery here.

I think @Martin_Kelly is referring to Chord’s (Rob Watts) comments to various posts re: Mojo and it’s input spec’s. I think there may be a bit of a chink in the M Scaler armor as his comment on my post was something like ‘Well it’s not a WTA filter so it can’t be as good’. Also the comment that the tap length is not as important as the filter construction. Upscaling and filtering is evolving and I am sure we will see even more approaches in the future. Right now @jussi_laako 's approach is allowing for more experimentation at a much lesser cost. imho of course.

1 Like

First time there I see Rob Watts admit that number of taps is not important. Because that’s the only thing he’s been bragging about for a long time. And indeed it is not important, the overall filter design matters.

And my comment is that is is much better because it is not WTA filter… :smiley:

(not that I would have just single “it”, I have many different filter algorithms one can choose from)

3 Likes

I’m having trouble with using LSN15. I use the sinc-m filter and whenever I skip a track, a loud high pitched noise is heared from the speakers. Even when I stop playing, that high pitched noise remains in the speakers.

I’ve changed the shaper to TPDF and now, that noise is only heared when I skip tracks, but it is not heared afterwards. I would really like to solve this problem. So my questions are these:

  1. How does the LSN15 shaper work and would it cause high frequency noise?
  2. Is there any way to prevent the loud noise when changing tracks?

Thanks.

I’m using Allo USBRIDGE Signature (Dietpi OS image based on kernel 4.14.92) with Chord Qutest.

1 Like

I’m not able to reproduce this with Holo Spring DAC or Chord Mojo connected to my Linux workstation.

Can you try with the DAC directly connected to the HQPlayer computer to simplify the setup a bit?

It’s a bit hard since I don’t have a laptop and the iMac is in another room. I’ll try it tomorrow.

However, since the Qutest is a 32bit DAC, do I need a shaper at all? Would the LSN15 help with the SQ?

Oh, and I have another problem: with the higher sample rate files (PCM96) that are upsampled to PCM768 I always get static in the right speaker at some point during playback. This doesn’t happen at PCM705…