I didn’t feel offended.
I just found it unneccessary, especially after you yourself said no one threw that argument at you 
You don’t have to apologize to me but I appreciate the gesture 
Thank you. I didn’t go to the bottom of the list!. I’ll try closed form to see whether I like it better than xtr.
1 Like
Thanks for the closed form recommendation. I just tried the closed form (not the fast one) with my mch files - upconverting DSD64 files to DSD256 (6 channels) and it worked great. I cannot use -xtr for mch files, it needs too much computing power for my quite powerful computer. However, closed form works great. Now I have to see whether I like the closed form better than the polysinc -2s that I have been using for mch files.
One of the quirks of closed form is that it isn’t poly-sinc meaning that it will only output integer multiples of the input sampling frequency. Sometimes I’ve been puzzled why I’m not getting any HQP output from an album when using closed-form and it turns out to be 48kHz album that I’m trying to upsample to a multiple of 44.1 kHz. If you check the Auto rate family box in settings then HQP will automatically output an integer multiple of the input sampling rate.
Thanks, good to know. So far I have only tested DSD64 mch files upconverted to DSD256 mch. I did do a 192/24 mch to DSD256 mch and it worked. Comparing the sound of poly-sinc-2s to closed form, I like closed form better, a slightly richer sound to my ear. I haven’t switched to stereo, so haven’t compared closed form to -xtr.
Larry
As I understand it the filters aren’t involved when doing DSD to DSD or at least not the ones in the main setup screen.
I thought I would post this quote by @jussi_laako from the monster HQPlayer thread, since modulators have been discussed upthread and remain to me one of the most arcane of HQPlayer’s settings. I recall a back and forth (which I couldn’t dig up) between Jussi and iFi, where whoever posts for iFi on CA expressed a preference for 5th order modulators. I’ve been listening to ASDM5 on my lowly DSD256 iDAC2 
and to my surprise I find it sounds much more natural in the treble. Like @andybob once remarked, I do also tend to like whatever I’m listening to.
Anyway the quote makes it clear modulators are very dac-dependent which explains the range of sonic qualities qualities ascribed to them upthread. Here’s the quote from Miska, p235 of the CA HQPlayer thread
The differences are more technical by nature, rather than aiming for any specific behavior. There is no clear way to describe the sonic differences. Main difference driving decision for a certain DACs is noise shaper order for 5th and 7th order modulators, 7th order modulator pushes more noise out of audio band and thus gives better SNR, but it also means that the noise outside of audio band increases on steeper slope and puts more demands on DAC’s low-pass filter. So DACs using very simple low-pass filters are better sticking to 5th order modulators.
DSD* modulators are fixed configuration ones while ASDM* modulators are adaptive in various ways based on source signal. So the DSD* ones are simpler and the ASDM* ones are more advanced algorithms. But you can just listen and decide which way you prefer, there is no clear way to say how something sounds and what should be used, apart from the modulator order as discussed above. Now in the latest beta there’s a new special beast called AMSDM7 tailored for >= DSD512 speeds (it is also allowed for DSD256, but then your mileage may vary depending on DAC - i need to do more measurements to see how it behaves in such cases).
Happy to hear what people are listening to now, or any musings on modulators etc! Myself I’m back to my old standby poly-sinc-short-mp and just enjoying without tinkering…for now.
3 Likes
I’ve also come back to DSD 512, poly-sinc-short-mp, but I’m enjoying DSD7 256+fs into a Holo Audio Spring (NOS). Interestingly that seems to be one of the shortest filters and simplest modulators.
God, what is that thing “Xtr Filter” - whether it is possible to hear the effects?
In which frequency band it affects?
Unless, of course, we don’t talk about cats - which are listening to our equipment along with us… 
The xtr filters are an available option when upsampling in HQ Player.
HQ Player is separate software from Roon. Roon can send audio to an HQ Player Zone.
The xtr filters are the “heaviest” filters in HQ Player and quite computationally intensive. They are steep low pass filters in the frequency domain and have very good bandwidth rejection. I’m not sure where the frequency cut off is but suspect it might be around 30 kHz.
Most cats prefer closed-form.
1 Like
Hi lorin,
By more natural in the treble, do you mean it’s warmer with ASDM5 with your iDAC2 (great DAC btw)?
I wouldn’t say warm. When I start messing with filters etc, 9/10 I’m listening to jazz and chasing a sound that evokes the brightness and impact of trumpet or alto sax etc without a sense of fatigue. Comparing ASDM7 to ASDM5 with the iDAC2, the latter sounds a little less strident, less piercing. The trade off …to my ears anyway… is a slightly less “hi fi” sound, maybe a diminished soundstage. With modern, dynamically compromised material, not sure I would notice a big difference.
I run the iDAC2 with the Standard DSD filter, the bottom switch position. I believe that was Jussi’s recommendation as well, I know he has one in his stable. I do like the iDAC2. The bass and especially sub bass are exceptional compared to other stuff in its price range.
1 Like
Frequency Cut off is around the Nyquist Limit of the source file. So if you upsample 44.1kHz to whatever higher target rate the sample rate conversion filter ideally has to filter out anything above 22kHz (simplified speaking).
“Steep” means the roll off starts pretty close to the actual nyquist limit.
With regard to 44.1kHz source files it’s pretty much like this:
Poly Sync roll off starts at around 20kHz (average amount of ringing).
Poly Syc short roll off starts at around 18.5kHz (low amount of ringing).
Poly Sinc xtr roll off starts at around 21kHz (very high amount of ringing).
To illustrate the above here are 3 graphs showing 44.1kHz White Noise played back through HQPlayer with 3 poly sinc filters (minimum phase variant of each) upsampling to 4x PCM rate. Recorded using a little Oppo HA-2 SE -> line out -> little Tascam Field Recorder (analog input recorded at 48kHz). While this setup doesn’t provide solid result about the filters it certainly shows how they roll off at high end…
poly sinc short mp:
poly sinc mp:
polay sinc xtr mp:
5 Likes
I’m hearing this earlier roll-off of poly-sinc-short-mp-2s more now with my iFi Pro iDSD than with my previous DACs.
I’ve left it on poly-sinc-short-mp-2s for a couple of weeks and enjoyed things. Just this weekend I tried poly-sinc-mp-2s and my ear/brain hears better transparency / more extended frequency response.
I don’t think it’s just artificially brighter?
I tried to go back to poly-sinc-short-mp-2s but now it does sound slightly too smooth / too rolled off in comparison.
I find this strange because I can’t imagine there’s that much more happening above 18kHz?
Unless it’s not only the extended frequency response that my ear/brain enjoys but it’s the longer ringing?
This is with Tidal RBCD streaming.
I am coming from Chord DACs so maybe my ear/brain system needs adjusting (ear/brain burn in).
I do find poly-sinc-mp-2s a nice fit in between short-2s and xtr-2s.
Hi @roomas,
Thank you for those most interesting graphs, unfortunately the pic hosting service seems to have expired. Are you able to upload them directly using the upload button (up pointing arrow) in the post editor ?
I won’t try to predict what you’re hearing Sean, but one of the most depressing graphs I think I’ve ever seen is this one about frequency response with age:
Ha. Fortunately Andy I’m on the right side of the mens chart (i.e. the left side 
) . But yes we all know things will only go downhill with time (for everyone).
At the moment I listen mostly with headphones, for critical listening anyway, so I don’t have any room effects at play.
Anyway, we know from what @jussi_laako has said, poly-sinc-short-mp has optimal transient response performance (that’s written in the manual) and on the other end of the scale the poly-sinc-xtr has the best frequency response of the poly-sinc family, at the cost of transient response…
Maybe my own ears are a little more sensitive to frequency response than transients with the poly-sinc family of filters? Maybe the transients are already ‘good enough’ for the entire poly-sinc family, even if there is a scale of best to worst?
Maybe others are also a little more sensitive to frequency response and this is why so many ask ‘xtr filters- why do they sound so good?’ Or is it something else. The better stopband attenuation (even better than -175dB!?)? The longer ringing?
It’s complicated stuff and there’s no way to find a simple explanation even though I’m trying myself in my own mind:grin: . There’s no free lunch with anything in life - it would be nice to have optimal transient response and optimal frequency response while maintaining -175dB stopband attenuation - but something’s gotta give I guess.
Anyway Rob Watts has written many times that the ear/brain system can need ~4 weeks to ‘burn in’ (i.e. adjust) with system changes, so maybe I need to give poly-sinc-short-mp-2s another 2-3 more weeks to see if I can adjust to it and not find it ‘too smooth/rolled off’. I was quite happy with it until I switched to poly-sinc-mp-2s and xtr-2s, which did immediately sound more transparent. As I mentioned, I don’t think I’m hearing artificially created ‘brightness’ with those filters.
Maybe poly-sinc-short-mp is actually filtering high frequency noise in RBCD recordings with it’s earlier roll-off and this is what my ear/brain system needs to adjust to? I haven’t got a clue 
Just sharing some fun observations and some questions that are in my mind - nothing too serious
1 Like
And a very tiny part of me is remotely interested in a chart of transient response sensitivity with age… next to the often talked about frequency response sensitivity chart.
I haven’t yet Googled if one exists.
I just turned 48 and funny thing doing room corrections I discovered I can’t hear tones over 16 kHz or so and that explains why I barely notice any difference when changing filters. Now question is what there is above 16 kHz, harmonics maybe?
1 Like
Well I’m 56 so it’s tempting to say nothing worth listening to 
. But yes, harmonics, overtones, some elements of cymbals etc.
Some folks have said that the preference of old men for extended high frequency sources like the Koetsu Rosewood moving coil cartridge etc. is because of the march of time so depressingly displayed above. Makes me wonder if my system sound stupidly bright to my kids (don’t get me started on ear buds).
Anyway, the easiest way to check if you can hear between 16k to 22k is a blind test like this. I’ll see how I do on an iPad, speaker may not get anywhere near that.
Edit: Over 3 x 10 tests I got 50% or less. Obviously the iPad speaker is the issue here …
4 Likes
From Rob Watts:
"It is well known that 96 kHz recordings sound better than 44.1 kHz (CD) recordings. Most people believe that this is due to the presence of ultrasonic information being audible even though the best human hearing is limited to 20kHz. What is not well known is that 768 kHz recordings sound better than 384 kHz and that the sound quality limit for sampling lies in the MHz region. 768 kHz recordings cannot sound better because of information above 200 kHz being important – simply because musical instruments, microphones, amplifiers and loudspeakers do not work at these frequencies nor can we hear them.
So if it is not the extra bandwidth that is important, why do higher sampling rates sound better?
The answer is not being able to hear inaudible supersonic information, but the ability to hear the timing of transients more clearly. It has long been known that the human ear and brain can detect differences in the phase of sound between the ears to the order of microseconds. This timing difference between the ears is used for localising high frequency sound. Since transients can be detected down to microseconds, the recording system needs to be able to resolve timing of one microsecond."
This is why I’m semi interested in a chart of transient response sensitivity with age, not just frequency response…
1 Like