On the PCM up sampling side of things, I’ve really settled into closed-form-M with my Allo Revolution:
To my ear it is more clear and natural sounding than sync-M. Sync-M can be fatiguing for me, but with closed-form-M I feel like I could listen all day. At first I was concerned that this might be because closed-form-M was overly smoothing things and loosing detail. But I don’t think this is the case. In most cases I hear subtle details coming through more clearly on closed-form-M.
Has anyone else had this experience with closed-form-M over sync-M? I see sync-M praised much more often on this forum than closed-from-M. Not that I need to go with the crowd or anything. Just curious what other people are experiencing.
It’s the DAC. PCM up sampling sounds better than DSD conversion on the Allo Revolution. Not sure what the Revolution is doing with DSD inputs. But 384 PCM (max for the Revolution) sounds killer with closed-form-M.
Back on my E30 DAC, the DSD (DSD256-5EC, gauss_long) is the only way to go!
interesting. I’ve not put much thought into the differences between apodizing and non-apodizing filters, other than a quick read through of a few Wikipedia pages. My cursory impression was that apodizing was somewhere between benign and helpful. Not something that would potentially smooth over details.
Any recommendations on material to read (or other posts on the forum here) regarding the impact of apodizing filters?
Apodizing filters don’t smooth over details. Instead they fix errors (distortion) in the high frequency area that makes many recordings sound harsh. Some may mistake the extra harshness/brightness to detail. These errors are from recording ADC digital filter, or from production chain software tools.
Keep eye on the “Apodizing” counter in HQPlayer. If it increases to figures >10 during playback, you should certainly be using an apodizing filter. If you use apodizing filter when such is not absolutely needed, there’s no harm either. But if you listen with non-apodizing filter material where the counter keeps increasing, you know you are listening to high frequency errors.
Is there a good example of a track that has lots of apodization errors in it that can be seen with the counter? because I haven’t seen the apod counter go above zero for a long time and would like to verify it working. Wondering if there are some conditions when streaming via Roon that it doesn’t work.
I’ve been running poly-sinc-gauss-xla, with ASDM5EC for about a week now, and enjoying it. But I notice in some cases the music sounds highly analytical to my ear, maybe even “over sharpened”. Not necessarily bad, very clear, separated, focused sounding. But for some tracks it takes on an unnatural sound.
After some experimenting I’ve started running just the basic poly-sinc-lp filter, with DSD5v2 256+fs for DSD256 output. I’m finding this to have a more neutral / less bright sound in general. Maybe just all in my head, maybe very DAC specific, etc. Whatever it is, I’m definitely having fun - HQPlayer is really a tweakers paradise!
Thanks for the recommendation. In the end the xtr-lp-short wasn’t my favorite, but your recommendation did help me realize two things:
I can run DSD512 if I pick an appropriate filter, which enables the use of the AMSDM7 DSD512+fs modulator, which I do think sounds very nice (so far, give it more time…)
The thing I’m finding is that I really like the sound of linear-phase filters. Prior to the Gauss filters I was a fan of the ext2 filter, which is linear-phase.
Yes, this is also a good recommendation. A less extreme Gauss filters gets back closer to the linear-phase filter sound I like.
I’m currently going back and forth between poly-sinc-ext2 and poly-sinc-gauss, using DSD512, AMSDM7 512+fs. Both sound good. I might (… might …) like the gausss better. But need more time with it…
I might actually like poly-sinc-gauss-short the best, but the integer up requirement and lack of apodizing feature make me shy away…
Have you (@dabassgoesboomboom) tried running at DSD512 with the AMSDM7 512+fs? My mac mini can run it fine with poly-sinc-ext2 and poly-sinc-gauss. And I think it sounds better than running DSD256 w/ ASDM7. Curious what you think of the sound difference…
Pulled the trigger on HQP the other day. I settled on poly-sinc-gauss-long using the ASDM7 modulator to upsample to DSD256x48 into my Gustard X16. So far impressed with the overall increase in SQ versus Roon’s upsampling.
