No errors loading the files in the logs as far as I can tell. Not sure this is useful but see screen grab below of a (hopefully) relevant section of the logs initialising the filters that appear not to work:
Please check that the “process” item doesn’t have same filter listed more than once. You can also clear it up, save, and redo the filter upload process so that you are sure the filter is there only once. For some reason, same filter is being loaded multiple times.
That log is incomplete (copy-paste preformatted text quote would be better than a picture). But I can see that the level ends up disappearing to -86 dB range.
Sorry for delayed response.
I have tried reloading the filters from scratch as suggested. The filters from Thierry are loaded per below - showing that they are all different:
From my previous post when I showed the plot for each of the 4 filters, only the first pipeline (_LL.wav) showed a plot for magnitude. As an experiment I loaded the _RR.wav Filter in Pipeline 1 and the _LL.wav filter in Pipeline 3. Results for this below showing that I only get a plot in Pipeline 1 and that it is different for _RR.wav than for _LL.wav. This seems to imply the filters are OK, but are only getting picked up when applied to the first Pipeline:
Plot when LL in Pipeline 1
Plot when LL in Pipeline 3
Plot when RR in Pipeline 1
Plot when RR in Pipeline 3
Could it be that there is something preventing the the magnitude of the Filters in Pipelines 2, 3, 4 from loading? This only started happening after I updated to HQPe v4.32.5 on filter sets that previously worked fine for over 2 years.
Only change in this respect from previous release is that there was a bug that caused HQPlayer to crash if one of the convolution filters failed to load.
When this error happens, log file says “loading of the convolution filters failed”. Other than that, there have been no changes in this area in a while.
Could you send me those original filter files over email so I could check? You could also perform backup from HQPlayer Embedded and send that too so I can check if the settings file has got messed up or something. You can use for example Google Drive link or similar instead of attaching those files.
I just emailed you on you Signalyst contact address with share link with the Filters and a backup of the Settings.
Thanks for your help and support!
Fanless PC’s possibly going to jump up in performance with 13th gen Intel?
Of course this is multi-thread but possibly single thread will see an efficiency improvement?
11th Gen Intel can already do DSD512 using ASDM7ECv2 at 65W TDP, but it is certainly getting better in that respect!
Now the interesting thing I’m waiting is when the Atom cores that are currently used as efficiency cores on 12th and 13th Gen will begin to appear on Atom SoC used in UP-Boards and such. These would be very nice, since they have AVX2 that has been earlier missing from Atom cores.
I have to report that I get exactly the same thing with my HQP setup. Have been using Thierry’s filters with HQP successfully for more than a year but once I upgraded to 4.32.5 I got the strange channel behaviour. For me when music starts to play it is only on the left channel but then, about 5 seconds later, it starts up on the right channel - but here it seems to start playing from the beginning of the track, so I get two versions of the music playing simultaneously but a few seconds offset in time. Its quite hard to listen to so I haven’t done extensive testing but it also sounds like this delayed R channel may be a bit distorted also.
I have reloaded the files from Thierry but with no fix to the problem. For now I have reverted to doing the room correction in Roon before the signal is passed to HQP.
Hello everyone, I’m wanting to update my harware where I use Roon Core with HQplayer, I currently use an older All In One HP where I updated to an i7, but older generation, with DDR3 and I’m only able to use ASDM7ECV2 with upsampler to DSD 128, if I change it to 256 the sound is interrupted (even if the CPU is about 40% in use). 256 is the limit of my DAC (DS Directstream) I have no plans to upgrade it.
I’m thinking of buying a NUC, model NUC11TNHI70000 that comes with a 2.8GHz i7 1165G7 and I’m thinking of putting 16GB DDR4 memory (I can put more).
I ask my colleagues, will I be able to run with this configuration with the heaviest filters in upsampler for DSD 256?
I accept other suggestions, thanks!
Which OS and/or HQPlayer build (package) are you using? I’d just like to know if this is specific to some specific build of HQPlayer.
It sounds like synchronization of the parallel processing is going wrong.
Btw, does Roon even know how to do channel multi-path mixing?
I’m using Sonicorbiter OS version 2.8 which is the latest version. I usually update the Sonicorbiter OS a few days after I see you have released a new version of HQPe - this will then incorporate the latest version of HQPe.
I too am using Sonicorbiter 2.8 (and have checked that I have the very latest version of this by running the software update routine).
Regarding Roon and room correction, this uses a different set of filters to the ones uploaded to HQP for the Matrix settings. Thierry at Home Audio Fidelity provides two sets of filters, one for Roon and one for HQP.
Just wondering, because this setup is more complex than a typical room correction, since it has left/right channel mixing through special filters in addition to direct left and right channel convolution corrections. Normal convolution engines cannot do this (not the simple HQPlayer convolution setup either). It requires the matrix processing that provides possibility to free channel re-routing and mixing.
I guessThierry knows this since he provides a filter set specifically labelled as “HQP” and, when I first asked him about using his filters in HQP, he gave detailed instructions on how to install the four files in the Matrix page.
No doubts about that…
I’m looking into this issue.