Fridolin Posted April 7, 2018 Share Posted April 7, 2018 28 minutes ago, Norton said: No I'm guessing you are fine with the iFi. Audirvana should be doing the first "unfold" to 96KHz (which is what your midi setting is telling you) then your DAC is doing a second "unfold" to 192kHz. Your Audio Midi doesn't tell you this because it doesn't know what happens subsequently in the DAC. Slowly it gets clearer to me. Thanks! Link to comment
crenca Posted April 7, 2018 Share Posted April 7, 2018 29 minutes ago, rickca said: This is what I suspect is the issue: https://www.computeraudiophile.com/forums/topic/30381-mqa-is-vaporware/?do=findComment&comment=767945 Perhaps @AMR/iFi audio can confirm if iFi's new firmware relies on MQA code for upsampling PCM. If so, would that explain the DSD limitations? Hey MQA, if it is not all $voodoo$, show us the math! Link to comment
Popular Post tmtomh Posted April 7, 2018 Popular Post Share Posted April 7, 2018 @mansr and/or anyone else: Sorry if this is too OT, but what exactly happens with MQA at the final render stage? In other words, if a 24/48 MQA file is unfolded to 24/96 in software, does the final render simply upsample and then apply a minimum-phase filter? If so, then it would seem to me, based on @Archimago's tests and other info we have, that software unfolding and then final playback on a non-MQA DAC, with no final render, might actually be preferable (or at least equivalent) sonically for the MQA files, and also for other files since they wouldn't be messed with by any MQA processing at all. And if the above is true (or at least makes sense), then the question would be, is there a way to capture the software-unfolded output digitally - for example with Handbrake or something? If that were possible, then one could effectively digitally "rip" unfolded MQA files into conventional PCM files and after that initial rip could just play the 24/96 PCM files instead. crenca and MikeyFresh 1 1 Link to comment
Popular Post mansr Posted April 7, 2018 Popular Post Share Posted April 7, 2018 51 minutes ago, crenca said: I wonder if that is not a "layman's" explanation, as it were. In other words MQA is just fat code and working it in is inelegant, and iFi only had limited man hours, etc. IF it is really truly always processing, then we are back to the important question as to what, exactly, is it doing and why, and what it means for our PCM/DSD. I intend to find out exactly what it does. Hooking up the logic analyser right now. tmtomh, 4est and crenca 1 2 Link to comment
Popular Post Miska Posted April 7, 2018 Popular Post Share Posted April 7, 2018 6 hours ago, tmtomh said: And if the above is true (or at least makes sense), then the question would be, is there a way to capture the software-unfolded output digitally - for example with Handbrake or something? If that were possible, then one could effectively digitally "rip" unfolded MQA files into conventional PCM files and after that initial rip could just play the 24/96 PCM files instead. Yes it is. That is sort of what I'm doing with HQPlayer Embedded, because it's an "upsampler box" between playback computer and a DAC. The unfolded output is not recorded anywhere, but it is instead just processed in realtime by HQPlayer like any other source (Spotify, etc) would be and then output to DAC. tmtomh and MikeyFresh 1 1 Signalyst - Developer of HQPlayer Pulse & Fidelity - Software Defined Amplifiers Link to comment
AMR/iFi audio Posted April 8, 2018 Author Share Posted April 8, 2018 On 6.04.2018 at 9:42 PM, mansr said: Would it not be possible to disable all MQA related processing at the highest rates? No, this is not how MQA works on a firmware level in XMOS. Not at all. tmtomh 1 Our PowerStation is here: click me! Check out our Tidal MQA Set-up Guides below. Android (Renderer) MobileDesktop (Decoder) via USBDesktop (Decoder) via SPDIF Link to comment
Popular Post mansr Posted April 8, 2018 Popular Post Share Posted April 8, 2018 10 minutes ago, AMR/iFi audio said: No, this is not how MQA works on a firmware level in XMOS. Not at all. That's terrible. Either they're plain incompetent, or they've intentionally made it as difficult as possible to avoid their meddling. tmtomh, MikeyFresh and crenca 1 1 1 Link to comment
Popular Post mansr Posted April 8, 2018 Popular Post Share Posted April 8, 2018 On 4/7/2018 at 6:19 PM, mansr said: I intend to find out exactly what it does. Hooking up the logic analyser right now. Here's what it looks like: Yeah, no warranty here. Miska, MikeyFresh, asdf1000 and 1 other 3 1 Link to comment
tmtomh Posted April 8, 2018 Share Posted April 8, 2018 6 hours ago, AMR/iFi audio said: No, this is not how MQA works on a firmware level in XMOS. Not at all. 6 hours ago, mansr said: That's terrible. Either they're plain incompetent, or they've intentionally made it as difficult as possible to avoid their meddling. Yikes. No MQA DAC for me, then. (And appreciate iFi offering the non-MQA firmware option.) MikeyFresh 1 Link to comment
mansr Posted April 8, 2018 Share Posted April 8, 2018 Ugh, the MQA impulse response clips for several of the filters. Solstice380 1 Link to comment
Popular Post mansr Posted April 8, 2018 Popular Post Share Posted April 8, 2018 Well, the good news is that non-MQA playback uses the built-in filters of the DSD1793 chip as before, at least on the Nano (original silver). There are, however, some slight differences, probably unrelated to MQA, in the configuration sequence sent to the DAC chip. Watch out for pops when starting DSD playback. listas_fede, rickca, asdf1000 and 1 other 1 1 2 Link to comment
rwdvis Posted April 9, 2018 Share Posted April 9, 2018 iFi, Have you not been paying attention to the MQA article on the CA Home page? Seems your MQA implementation decision was not well thought out in light of recent findings. https://www.computeraudiophile.com/ca/reviews/mqa-a-review-of-controversies-concerns-and-cautions-r701/ On 3/31/2018 at 12:53 AM, AMR/iFi audio said: Enjoy the MQA magic like never before. “MQA magic”? Could you please explain? Feel free to post your response to this question in the MQA article comments section linked above. Your definition of magic would be a helpful start. On 3/31/2018 at 12:53 AM, AMR/iFi audio said: iFi first introduced MQA into its product mix with the nano iDSD Black Label in November 2017. Since then, iFi customers across the globe have been asking when this option would be available in other iFi products. Are you certain that these “iFi customers” were not just MQA shills/affiliates making such requests? It’s well known that there are a number of active MQA shills infiltrating various forums and pressuring manufacturers. On 3/31/2018 at 2:52 AM, AMR/iFi audio said: The MQA filters are in use ONLY run when MQA files are passed through and vice versa. Are you absolutely certain about this? Link to comment
AMR/iFi audio Posted April 9, 2018 Author Share Posted April 9, 2018 On 7.04.2018 at 3:57 PM, rickca said: That doesn't answer my question. Please review MQA's website for more details. On 7.04.2018 at 8:15 AM, Norton said: It's good that you are providing the option of MQA, but as a potential customer, I'd feel more reassured if the engineering department had deemed measurements necessary, prior to release. R&D does not divulge all their measurements to the Marketing Dept. and even if they did, not all measurements are for public consumption. So rest assured we do conduct measurements - the Audio Precision 2 is not cheap so we do like to amortise the heck out of it! But the fact that we are one of very few audio companies that write our own XMOS code should suggest we have knowledge of what we are doing rather than just implementing a turnkey solution. Few offer a steady stream of multiple firmware versions for a range of products spanning several years. Last and not least, one cannot measure pleasure. Hence AMR/iFi always audition, internally and externally extensively prior to release. For us, measurements are useful but not the end all and be all. It's a little akin to one buying a car based upon the fastest 0-60 and not taking it for a test drive. Our PowerStation is here: click me! Check out our Tidal MQA Set-up Guides below. Android (Renderer) MobileDesktop (Decoder) via USBDesktop (Decoder) via SPDIF Link to comment
AMR/iFi audio Posted April 9, 2018 Author Share Posted April 9, 2018 On 8.04.2018 at 3:03 PM, mansr said: That's terrible. Either they're plain incompetent, or they've intentionally made it as difficult as possible to avoid their meddling. Dear mansr, before judging competence or incompetence of anyone, it would probably be a good idea for you to learn how to write software for Xcore processors and how audio and MQA integration works on these. As a software developer you can sign up with MQA to integrate this into your sample rate conversion libraries. Past that MQA are quite open about how their system works and what it does. There are multiple patents filed and extensive material available on the MQA website. It is true that specific details are not made publicly available, however these affect neither the understanding of the operation principles nor the relative impact on software including MQA, it's merely an insurance that it is not possible to build an MQA decoder/renderer or MQA renderer without obtaining these details from MQA. Past that, given all the information available freely about MQA, the amount of disinformation presented as fact is surprising to us. We do wonder if there is truth after all in the 'those who know the least talk the most' truism. Our PowerStation is here: click me! Check out our Tidal MQA Set-up Guides below. Android (Renderer) MobileDesktop (Decoder) via USBDesktop (Decoder) via SPDIF Link to comment
mansr Posted April 9, 2018 Share Posted April 9, 2018 1 hour ago, rwdvis said: On 3/31/2018 at 10:52 AM, AMR/iFi audio said: The MQA filters are in use ONLY run when MQA files are passed through and vice versa. Are you absolutely certain about this? I checked the behaviour of the Nano (silver). The DAC chip is configured exactly the same as with firmware 5.20 for non-MQA playback. I don't have any other iFi products, but I have no reason to believe they'd be different. Link to comment
mansr Posted April 9, 2018 Share Posted April 9, 2018 11 minutes ago, AMR/iFi audio said: Please review MQA's website for more details then. Very funny. Link to comment
AMR/iFi audio Posted April 9, 2018 Author Share Posted April 9, 2018 13 hours ago, mansr said: Ugh, the MQA impulse response clips for several of the filters. This has nothing to do with MQA. 99% of DAC's out there using digital filters will clip when presented with a full scale signal that does not conform to the strictures of the Nyquist–Shannon sampling theorem. This holds specifically true of square-wave signals and one sample long impulses used as test signals. This has been known for many decades and isn,t actually a problem, except for test-signals that are not in compliance with the strictures of the Nyquist–Shannon sampling theorem. This sort of thing is usually covered in year one or two EE Courses. Real musical recoding's are low-pass filtered before the analogue to digital recording process in order to comply with the strictures of the Nyquist–Shannon sampling theorem. Thus, unless severely clipped in the digital domain to start with (in which case what sort of fidelity are we talking about in the first place), they do not contain any signal components that can trigger these "digital filter" overs. So just to recap, this has nothing to do with MQA and is not an issue outside the lab. 13 hours ago, mansr said: There are, however, some slight differences, probably unrelated to MQA, in the configuration sequence sent to the DAC chip. The end-result of how the chip is configured is the same. 13 hours ago, mansr said: Watch out for pops when starting DSD playback. We did. Our PowerStation is here: click me! Check out our Tidal MQA Set-up Guides below. Android (Renderer) MobileDesktop (Decoder) via USBDesktop (Decoder) via SPDIF Link to comment
mansr Posted April 9, 2018 Share Posted April 9, 2018 7 minutes ago, AMR/iFi audio said: This has nothing to do with MQA. The MQA resampler clipping has everything to do with MQA. I have never before encountered a resampler that clips on a single-sample full-scale impulse. It may not be likely to occur with music, but why take the chance? Link to comment
mansr Posted April 9, 2018 Share Posted April 9, 2018 11 minutes ago, AMR/iFi audio said: 13 hours ago, mansr said: There are, however, some slight differences, probably unrelated to MQA, in the configuration sequence sent to the DAC chip. The end-result of how the chip is configured is the same. Yes, the end result is the same. However, with the new firmware, the headphone amp is unmuted before the DAC is configured as opposed to after with firmware 5.20. This sometimes results in a pop or click. MikeyFresh 1 Link to comment
mansr Posted April 9, 2018 Share Posted April 9, 2018 Switching between MQA filters also causes some nasties: As does switching from non-MQA to MQA: MikeyFresh 1 Link to comment
firedog Posted April 11, 2018 Share Posted April 11, 2018 I was looking at buying one of your DACs that is MQA compliant and has an SPDIF out, for use as a USB/SPDIF converter and for MQA playback. But now it appears the MQA update also disables the SPDIF out on your DACs. Is this true for all your MQA compliant DACs with an SPDIF out? Main listening (small home office): Main setup: Surge protector +>Isol-8 Mini sub Axis Power Strip/Isolation>QuietPC Low Noise Server>Roon (Audiolense DRC)>Stack Audio Link II>Kii Control>Kii Three (on their own electric circuit) >GIK Room Treatments. Secondary Path: Server with Audiolense RC>RPi4 or analog>Cayin iDAC6 MKII (tube mode) (XLR)>Kii Three . Bedroom: SBTouch to Cambridge Soundworks Desktop Setup. Living Room/Kitchen: Ropieee (RPi3b+ with touchscreen) + Schiit Modi3E to a pair of Morel Hogtalare. All absolute statements about audio are false Link to comment
Popular Post AMR/iFi audio Posted April 11, 2018 Author Popular Post Share Posted April 11, 2018 On 8.04.2018 at 3:03 PM, mansr said: That's terrible. Either they're plain incompetent, or they've intentionally made it as difficult as possible to avoid their meddling. The following post will address what to most will be excessive detail - some of the fundamental in's and out's of the XMOS inner workings and limitations. All the following basic technical facts are covered by the extensive documentation XMOS provide in support of their products, which one may study in-depth over at http://www.xmos.com/ The U8 processor used by current and historical iFi USB audio products, excluding the original iDAC micro, has 8 logical cores and a total of up to 500 MIPS spread across the cores. Due to the hardware/software XMOS architecture, usually a single core is needed for a single function, eg. the SPDIF output requires the sole use of one core and a certain amount of MIPS. The XMOS architecture does not allow processes to be selectively terminated or instantiated at run time. They are either active or not, nor can they be ‘suspended’. That is the consequence of the revolutionary feature of XMOS – it is a software-driven multitasking system with hardware-like guaranteed response times. A further limitation imposed by XMOS is that even though the processor has 8 cores, when using USB at high speed (as required for USB Audio Class 2) only 6 cores are allowed to be active. Another basic fact is that any audio format that includes a ‘signature code’ to identify it, be it - HDCD, DoP DSD or MQA requires the decoder to analyse each audio sample to identify both its signature code and the relevant stream and then take the correct decoding action. Prior to adding MQA, our XCore100 U8 chip had all of its 6 allowed cores used and almost all available 500MIPS consumed when decoding DSD512 or 705/768kHz. In order to run the MQA engine and thus integrate MQA decoding, we needed to free up at least one core and enough MIPS. Because the S/PDIF output functionality was not widely used, it has been progressively dropped from current iFi products long before we committed to implementing MQA. Hence to remove the S/PDIF out function and to reduce the maximum sample rate to gain sufficient MIPS to allow MQA to operate was the logical solution. We still seek ways to remove these limitations, however no other solution has presented itself thus far. Last and not least - if there is someone more proficient and capable than our XMOS team (who were in no small part assisted by the MQA XMOS team), is able to better implement MQA functionality into XMOS products, we wish to talk to that person. Please PM us with your CV and samples of your work which we will forward to our Technical Director - Thorsten Loesch. He always has job openings for exceptional XMOS programmers, freelance or permanent. tmtomh, rafaelpernil, Bootzilla and 1 other 4 Our PowerStation is here: click me! Check out our Tidal MQA Set-up Guides below. Android (Renderer) MobileDesktop (Decoder) via USBDesktop (Decoder) via SPDIF Link to comment
rickca Posted April 11, 2018 Share Posted April 11, 2018 On 4/6/2018 at 8:12 PM, rickca said: What is the nature of this required digital processing? It's hard to imagine that it is so resource intensive that you had to drop DSD512. Thanks for actually answering my question @AMR/iFi audio. Much better than your original response https://www.computeraudiophile.com/forums/topic/38962-ifi-audio-mqa-all-across-the-board-its-here/?do=findComment&comment=801586 Pareto Audio AMD 7700 Server --> Berkeley Alpha USB --> Jeff Rowland Aeris --> Jeff Rowland 625 S2 --> Focal Utopia 3 Diablos with 2 x Focal Electra SW 1000 BE subs i7-6700K/Windows 10 --> EVGA Nu Audio Card --> Focal CMS50's Link to comment
firedog Posted April 11, 2018 Share Posted April 11, 2018 Thanks for the answer. Main listening (small home office): Main setup: Surge protector +>Isol-8 Mini sub Axis Power Strip/Isolation>QuietPC Low Noise Server>Roon (Audiolense DRC)>Stack Audio Link II>Kii Control>Kii Three (on their own electric circuit) >GIK Room Treatments. Secondary Path: Server with Audiolense RC>RPi4 or analog>Cayin iDAC6 MKII (tube mode) (XLR)>Kii Three . Bedroom: SBTouch to Cambridge Soundworks Desktop Setup. Living Room/Kitchen: Ropieee (RPi3b+ with touchscreen) + Schiit Modi3E to a pair of Morel Hogtalare. All absolute statements about audio are false Link to comment
AMR/iFi audio Posted April 12, 2018 Author Share Posted April 12, 2018 On 11.04.2018 at 4:25 PM, firedog said: I was looking at buying one of your DACs that is MQA compliant and has an SPDIF out, for use as a USB/SPDIF converter and for MQA playback. But now it appears the MQA update also disables the SPDIF out on your DACs. Is this true for all your MQA compliant DACs with an SPDIF out? At this point in time we cannot offer S/PDIF out and MQA together. But feel free to select between the different firmware versions. Past this, S/PDIF out on MQA does not exist. Our PowerStation is here: click me! Check out our Tidal MQA Set-up Guides below. Android (Renderer) MobileDesktop (Decoder) via USBDesktop (Decoder) via SPDIF 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