mozes Posted July 25, 2018 Share Posted July 25, 2018 I am enjoying the technical and Math discussions on this thread even though most of it go over my head. From a listener's perspective and this is what really counts in my humble opinion, I will share my experience. I had a chance to play with a Blu2/Dave combo for an afternoon a while back. I started playing different tracks that I am familiar with on the Blu2/Dave and all sounded sublime. I was using HD800S headphones plugged straight to Dave. Then I disconnected the Blu2 and to sum it up, the music still sounded good but boring! I felt that the liveliness of music is gone. This is what I mainly felt by dropping the Blu2. It is not important to go into audiophile terms regarding resolution and soundstage etc. What really matters is the musical enjoyment that the Blu2 offers and I expect the HMS to deliver even more of that. ElviaCaprice 1 Link to comment
Popular Post ecwl Posted July 25, 2018 Popular Post Share Posted July 25, 2018 55 minutes ago, Miska said: P.S. I just tested setting one of my filters to million taps and upsampling to 16x PCM, CPU load on my oldish quad-core Xeon E5 consumes less than 5% of CPU time. CA forum is amazing. I need to give HQPlayer a try. Especially since I can’t connect an M-Scaler to my Mojo when working on my desktop. The Computer Audiophile and beautiful music 2 Link to comment
Miska Posted July 25, 2018 Share Posted July 25, 2018 48 minutes ago, rayl1234 said: Perhaps off topic for this thread but I never could figure out how to get hqp set up such that arbitrary audio via wasapi and also via windows mixer to pass through hqp and have it output 768 or 705.6 to the usb device of my choice. I think that kind of stuff is better discussed in HQPlayer thread, I don't want to hijack M-Scaler thread discussing HQPlayer when it's not strictly relevant for the topic in generic terms (computer vs dedicated device). Signalyst - Developer of HQPlayer Pulse & Fidelity - Software Defined Amplifiers Link to comment
The Computer Audiophile Posted July 25, 2018 Share Posted July 25, 2018 One thing I love is that we have options. Custom HQP or M Scaler or whatever. All very cool. barrows 1 Founder of Audiophile Style | My Audio Systems Link to comment
mav52 Posted July 26, 2018 Author Share Posted July 26, 2018 2 hours ago, The Computer Audiophile said: One thing I love is that we have options. Custom HQP or M Scaler or whatever. All very cool. Very cool. The Truth Is Out There Link to comment
ElviaCaprice Posted July 26, 2018 Share Posted July 26, 2018 I for one, would rather have a software server solution. But until someone (Contrary, all comparisons I've seen done have concluded, no contest, M Scaler rules) can show me, from listening, that there is a software solution that can come even close to M Scaler, it's a no go. Hardware solution wins. (JRiver) Jetway barebones NUC (mod 3 sCLK-EX, Cybershaft OP 14) (PH SR7) => mini pcie adapter to PCIe 1X => tXUSBexp PCIe card (mod sCLK-EX) (PH SR7) => (USPCB) Chord DAVE => Omega Super 8XRS/REL t5i (All powered thru Topaz Isolation Transformer) Link to comment
Popular Post esldude Posted July 26, 2018 Popular Post Share Posted July 26, 2018 A simple graphical illustration for those not familiar with digital filtering. From Audacity's EQ panel. Here is a brick wall filter going from flat to about -120 db between 20 khz and 24 khz (what you might do for 48 khz sample rates). Blue is the EQ target I chose. Green is the response you'll get. First up 21 taps. Next will be 427 taps which gets us much closer. Now 1353 taps which appears to get us there for this sharp steep filter. The Computer Audiophile, Jud, asdf1000 and 3 others 3 1 2 And always keep in mind: Cognitive biases, like seeing optical illusions are a sign of a normally functioning brain. We all have them, it’s nothing to be ashamed about, but it is something that affects our objective evaluation of reality. Link to comment
Popular Post esldude Posted July 26, 2018 Popular Post Share Posted July 26, 2018 Now, here is a single sample impulse. Top line is one with brick wall filtering. And bottom is the impulse with no filtering. First up is the 21 tap filter. Now I'm using the db waveform view which exaggerates the size of the samples. But notice this is rounded with what amounts to a cycle of negative ringing. Notice it effects only 21 samples around the single sample impulse. Next is the same view only with the 1353 tap filter. Both of these were set to be as steep as possible 20k to 24 khz. I'm not showing the entire thing, but notice the ringing going off the image on both sides as this filter rings over 1353 samples. Again this is exaggerated for the illustration. Here is the simple waveform view without the exaggeration. Also note I am showing this at 192 khz for illustration purposes. The ringing appears to involve several samples per cycle, but at 48 khz it would not. All the energy in the ringing portion would be between 20-24 khz. Well actually on the 21 tap filter it might leak some out the upper side of 24 khz. Jud, Currawong, mav52 and 2 others 2 1 2 And always keep in mind: Cognitive biases, like seeing optical illusions are a sign of a normally functioning brain. We all have them, it’s nothing to be ashamed about, but it is something that affects our objective evaluation of reality. Link to comment
rayl1234 Posted July 26, 2018 Share Posted July 26, 2018 19 minutes ago, esldude said: Now, here is a single sample impulse. Top line is one with brick wall filtering. And bottom is the impulse with no filtering. First up is the 21 tap filter. Now I'm using the db waveform view which exaggerates the size of the samples. But notice this is rounded with what amounts to a cycle of negative ringing. Notice it effects only 21 samples around the single sample impulse. Next is the same view only with the 1353 tap filter. Both of these were set to be as steep as possible 20k to 24 khz. I'm not showing the entire thing, but notice the ringing going off the image on both sides as this filter rings over 1353 samples. Again this is exaggerated for the illustration. Here is the simple waveform view without the exaggeration. Also note I am showing this at 192 khz for illustration purposes. The ringing appears to involve several samples per cycle, but at 48 khz it would not. All the energy in the ringing portion would be between 20-24 khz. Well actually on the 21 tap filter it might leak some out the upper side of 24 khz. Just to clarify on the ringing discussion, that an impulse or a square wave or anything with such a sharp corner is not bandwidth limited. An impulse has infinite frequency bandwidth... so if the filtering before the ADC clipped things to 22 kHz, an impulse will never be sampled. (Not to mention, a true impulse also has zero probability of being sampled bec any point has zero measure [probability] within a real interval— but of course, a true impulse is a mathematical construct anyway and not a real signal....). I find that sometimes the bw limited assumption in digital sampling is forgotten in DAC discussions. Link to comment
esldude Posted July 26, 2018 Share Posted July 26, 2018 2 minutes ago, rayl1234 said: Just to clarify on the ringing discussion, that an impulse or a square wave or anything with such a sharp corner is not bandwidth limited. An impulse has infinite frequency bandwidth... so if the filtering before the ADC clipped things to 22 kHz, an impulse will never be sampled. (Not to mention, a true impulse also has zero probability of being sampled bec any point has zero measure [probability] within a real interval— but of course, a true impulse is a mathematical construct anyway and not a real signal....). I find that sometimes the bw limited assumption in digital sampling is forgotten in DAC discussions. Yes, I agree. The only purpose in an illegal single sample impulse is to investigate the filter. And always keep in mind: Cognitive biases, like seeing optical illusions are a sign of a normally functioning brain. We all have them, it’s nothing to be ashamed about, but it is something that affects our objective evaluation of reality. Link to comment
mansr Posted July 26, 2018 Share Posted July 26, 2018 15 hours ago, ecwl said: The issue is that to compute each digital filtered sample, you would need to store in memory the 62500 original digital samples, do the 62500 multiplications and add up the sum, and then do it again and again using 16 different sets of coefficients a million times to get to 1 million samples from the original 62500. Actually, you don't have to do all those calculations. Link to comment
Sloop John B Posted July 26, 2018 Share Posted July 26, 2018 12 hours ago, Miska said: P.S. I just tested setting one of my filters to million taps and upsampling to 16x PCM, CPU load on my oldish quad-core Xeon E5 consumes less than 5% of CPU time. Can I suggest you take the opportunity created by all this publicity and create a new filter or alter a current one and call it xxxxxxx-2MT. (or 1MT). I realise you have previously questioned the usefulness of the TAPs numeration but it seems to have caught the imagination. I was always a bit amazed at how many DAVE owners were using HQP and maybe just maybe this is why the upscaler has come into existence? It’s a great opportunity to sell more HQP as if there was a 1MT filter happy users like me could spread the word out there in other Hifi fora where HQP is barely known. .sjb jhwalker and Sloop John B 1 1 Link to comment
misterspense Posted July 26, 2018 Share Posted July 26, 2018 5 hours ago, esldude said: Yes, I agree. The only purpose in an illegal single sample impulse is to investigate the filter. The response to an 'illegal' signal is often used in discussions to ground a preference for a minimum phase filter, because of the ringing the filter shows for an illegal signal that does not exist in normal digital music. The impulse cannot be present in the music because it was bandwith delimited in the ADC during the recording. Currawong 1 Link to comment
jhwalker Posted July 26, 2018 Share Posted July 26, 2018 1 hour ago, Sloop John B said: Can I suggest you take the opportunity created by all this publicity and create a new filter or alter a current one and call it xxxxxxx-2MT. (or 1MT). I realise you have previously questioned the usefulness of the TAPs numeration but it seems to have caught the imagination. I was always a bit amazed at how many DAVE owners were using HQP and maybe just maybe this is why the upscaler has come into existence? It’s a great opportunity to sell more HQP as if there was a 1MT filter happy users like me could spread the word out there in other Hifi fora where HQP is barely known. .sjb I am thinking the same thing - would love to see more HQP filters that emulate / parallel other popular solutions, along with comments / documentation on which are more / less "accurate", etc. FWIW, I've recently settled on poly-sinc-short-mp, ASDM7 for conversion to DSD. Given I don't have the tools to model, would love to see graphs for this and the other filter choices. John Walker - IT Executive Headphone - SonicTransporter i9 running Roon Server > Netgear Orbi > Blue Jeans Cable Ethernet > mRendu Roon endpoint > Topping D90 > Topping A90d > Dan Clark Expanse / HiFiMan H6SE v2 / HiFiman Arya Stealth Home Theater / Music -SonicTransporter i9 running Roon Server > Netgear Orbi > Blue Jeans Cable HDMI > Denon X3700h > Anthem Amp for front channels > Revel F208-based 5.2.4 Atmos speaker system Link to comment
audio.bill Posted July 26, 2018 Share Posted July 26, 2018 Here's a link to a presentation by M Scaler designer Rob Watts detailing its objectives and technology. He explains how his WTA filter implemented with 1M taps achieves better than 16 bit accuracy while claiming that even the best other DACs today are only 2 or 3 bits accurate with reconstruction. The M Scaler uses 528 DSP cores running at 4096 FS with a total bit depth of 56 bits. Many more details are in the linked presentation. I expect this to generate some further discussion which I look forward to... ? The Computer Audiophile 1 Link to comment
mansr Posted July 26, 2018 Share Posted July 26, 2018 40 minutes ago, audio.bill said: Here's a link to a presentation by M Scaler designer Rob Watts detailing its objectives and technology. He explains how his WTA filter implemented with 1M taps achieves better than 16 bit accuracy while claiming that even the best other DACs today are only 2 or 3 bits accurate with reconstruction. The M Scaler uses 528 DSP cores running at 4096 FS with a total bit depth of 56 bits. Many more details are in the linked presentation. I expect this to generate some further discussion which I look forward to... ? 2-3 bits? That's crazy talk. The whole presentation looks rather hand-wavy to me. Link to comment
ecwl Posted July 26, 2018 Share Posted July 26, 2018 4 hours ago, mansr said: Actually, you don't have to do all those calculations. Can you elaborate so I can understand better? I understand that if you're using a million taps to do 16fs, since you only start with 62,500 samples, you'll end up with with 62,500 computations and then sum them up. The only way I can see not doing all those calculations would be to set more coefficients to zero. Except isn't that cheating a little? Then I can claim a 100-tap filter is 1 million taps because the other 999,900 taps have coefficient of zero. Am I missing something here? 11 minutes ago, mansr said: 2-3 bits? That's crazy talk. The whole presentation looks rather hand-wavy to me. I also found the presentation a little too hand-wavy. But that's why I find @rayl1234 diagram (2nd post of this thread) to be helpful, showing how tap length affects amplitude accuracy in the time domain. Can you elaborate more how this is wrong? I feel like I'm hearing lots of people say this whole Chord approach is just marketing or plain wrong (which is a fair enough opinion). And I don't mind finding out that I'm wrong. I am just wondering about the why so that I can learn more about this. Link to comment
rayl1234 Posted July 26, 2018 Share Posted July 26, 2018 24 minutes ago, mansr said: 2-3 bits? That's crazy talk. The whole presentation looks rather hand-wavy to me. I am not going to play RW defender beyond saying this bit. While I sometimes find his statements imprecise, he has almost always clarified when asked at a show or even on his thread at headfi. He is an earnest artisan and enthusiast more than a commercial marketer. But he does have certain philosophies so to speak. Like against dsd... Currawong 1 Link to comment
mansr Posted July 26, 2018 Share Posted July 26, 2018 3 minutes ago, ecwl said: Can you elaborate so I can understand better? I understand that if you're using a million taps to do 16fs, since you only start with 62,500 samples, you'll end up with with 62,500 computations and then sum them up. The only way I can see not doing all those calculations would be to set more coefficients to zero. Except isn't that cheating a little? Then I can claim a 100-tap filter is 1 million taps because the other 999,900 taps have coefficient of zero. Am I missing something here? With a large number of taps, the calculation can be done more efficiently through the use of FFTs. The end result is the same even though fewer multiplications are performed. 3 minutes ago, ecwl said: I also found the presentation a little too hand-wavy. But that's why I find @rayl1234 diagram (2nd post of this thread) to be helpful, showing how tap length affects amplitude accuracy in the time domain. Can you elaborate more how this is wrong? How what is wrong? rayl1234's graphs look correct. 3 minutes ago, ecwl said: I feel like I'm hearing lots of people say this whole Chord approach is just marketing or plain wrong (which is a fair enough opinion). And I don't mind finding out that I'm wrong. I am just wondering about the why so that I can learn more about this. The approach of using a long filter is technically sound. The marketing talk about why they need to be that long is sketchy and over the top. ecwl 1 Link to comment
mansr Posted July 26, 2018 Share Posted July 26, 2018 2 minutes ago, rayl1234 said: I am not going to play RW defender beyond saying this bit. While I sometimes find his statements imprecise, he has almost always clarified when asked at a show or even on his thread at headfi. He is an earnest artisan and enthusiast more than a commercial marketer. But he does have certain philosophies so to speak. Like against dsd... I believe him to be sincere in his desire to build the best possible DAC. Some of his engineering decisions in pursuit of that goal also make perfect sense. Some others seem, not exactly bad, but definitely influenced by what you call his philosophies. In marketing oriented material, such as the presentation slides linked above, he tends to make some rather large leaps, such as the "2-3 bits" claim. That just doesn't make sense at all. Oh, and those "transient" waveforms he shows are not band limited as he claims them to be. Link to comment
Jud Posted July 26, 2018 Share Posted July 26, 2018 8 minutes ago, mansr said: Oh, and those "transient" waveforms he shows are not band limited as he claims them to be. Yes, bandwidth-unlimited waveforms at a specific frequency are ever popular for that sort of thing. One never knows, do one? - Fats Waller The fairest thing we can experience is the mysterious. It is the fundamental emotion which stands at the cradle of true art and true science. - Einstein Computer, Audirvana -> optical Ethernet to Fitlet3 -> Fibbr Alpha Optical USB -> iFi NEO iDSD DAC -> Apollon Audio 1ET400A Mini (Purifi based) -> Vandersteen 3A Signature. Link to comment
Popular Post barrows Posted July 26, 2018 Popular Post Share Posted July 26, 2018 15 hours ago, The Computer Audiophile said: One thing I love is that we have options. Custom HQP or M Scaler or whatever. All very cool. Yes, and we must remember that not all audiophiles are going to be interested in setting up, and learning the ins and outs of a sophisticated HQPlayer software based oversampling system. For them, the ability to just purchase a hardware based solution which is plug and play is a good thing. Ultimately, it is clear to me that a well sorted HQPlayer set up, with an appropriately stout computer, and a Renderer (NAA) feeding the audio DAC is more flexible, and also allows for virtually unlimited filter choices in the future. But again, this is not going to be for everybody. The Computer Audiophile, 4est, ElviaCaprice and 2 others 4 1 SO/ROON/HQPe: DSD 512-Sonore opticalModuleDeluxe-Signature Rendu optical with Well Tempered Clock--DIY DSC-2 DAC with SC Pure Clock--DIY Purifi Amplifier-Focus Audio FS888 speakers-JL E 112 sub-Nordost Tyr USB, DIY EventHorizon AC cables, Iconoclast XLR & speaker cables, Synergistic Purple Fuses, Spacetime system clarifiers. ISOAcoustics Oreas footers. SONORE computer audio | opticalRendu | ultraRendu | microRendu | Signature Rendu SE | Accessories | Software | Link to comment
ElviaCaprice Posted July 26, 2018 Share Posted July 26, 2018 This is where I would disagree. A renderer is less flexible than a single server due to it's confines to lacking a flexible OS. A well implemented server with clean power and clocking is far more flexible to stream to an M Scaler. And until a software solution only upon a server is shown to match the performance of a hardware solution such as M Scaler, it will not be the desired approach, let alone the complexity of setting up that software for the common audiophile. In turn this is/will bode well for Chord DAC's and future DAC/AMP designs for driving speakers/headphones direct, bypassing any need for separate pre-amp/amp. ElviaCaprice 1 (JRiver) Jetway barebones NUC (mod 3 sCLK-EX, Cybershaft OP 14) (PH SR7) => mini pcie adapter to PCIe 1X => tXUSBexp PCIe card (mod sCLK-EX) (PH SR7) => (USPCB) Chord DAVE => Omega Super 8XRS/REL t5i (All powered thru Topaz Isolation Transformer) Link to comment
look&listen Posted July 26, 2018 Share Posted July 26, 2018 1 hour ago, ecwl said: I find @rayl1234 diagram (2nd post of this thread) to be helpful, showing how tap length affects amplitude accuracy in the time domain. Can you elaborate more how this is wrong? +1 Link to comment
barrows Posted July 26, 2018 Share Posted July 26, 2018 10 minutes ago, ElviaCaprice said: This is where I would disagree. A renderer is less flexible than a single server due to it's confines to lacking a flexible OS. A well implemented server with clean power and clocking is far more flexible to stream to an M Scaler. And until a software solution only upon a server is shown to match the performance of a hardware solution such as M Scaler, it will not be the desired approach, let alone the complexity of setting up that software for the common audiophile. In turn this is/will bode well for Chord DAC's and future DAC/AMP designs for driving speakers/headphones direct, bypassing any need for separate pre-amp/amp. Say What??? My post re a Renderer was about using HQPlayer, first of all. Second of all, a good Renderer will give better performance with either an mScaler or a DAC. A full featured server will never achieve the lower noise footprint of a well designed, high end, Renderer. ElviaCaprice 1 SO/ROON/HQPe: DSD 512-Sonore opticalModuleDeluxe-Signature Rendu optical with Well Tempered Clock--DIY DSC-2 DAC with SC Pure Clock--DIY Purifi Amplifier-Focus Audio FS888 speakers-JL E 112 sub-Nordost Tyr USB, DIY EventHorizon AC cables, Iconoclast XLR & speaker cables, Synergistic Purple Fuses, Spacetime system clarifiers. ISOAcoustics Oreas footers. SONORE computer audio | opticalRendu | ultraRendu | microRendu | Signature Rendu SE | Accessories | Software | Link to comment
Recommended Posts
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now