Jump to content

bhobba

Members
  • Content count

    6,024
  • Joined

  • Last visited

  • Days Won

    4

About bhobba

  • Rank
    Seriously DAC Addicted
  • Birthday 17/11/1955

Profile Fields

  • Location
    Redland Bay
  • Country
    Australia
  1. Well they do but it's inaudible - in fact its the basis of the following criticism of the whole program: https://www.xivero.com/de/hypothesis-paper-to-support-a-deeper-technical-analysis-of-mqa-by-mqa-limited/ Simply locate where music extends and noise begins and only transmit the music. I have their program that does that and it often produces smaller files than MQA. But - technology is advancing and the frequency where recorded music ends is getting higher and higher. BYW musical instruments can reach quite high in frequency before its drowned by noise: https://www.cco.caltech.edu/~boyk/spectra/spectra.htm But whether reproducing them is of any value is debatable - its the ringing that such high frequencies hitting a steep filter that's at issue. And we simply have people saying it is - we have diagrams and what not showing the blurring - but no blind listening tests etc to back it up. Personally I don't care - I like MQA - but objective evidence is necessary to back up claims. Thanks Bill
  2. Now that's the 64 million dollar question isn't it. Thanks Bill
  3. I did a google search and found at least one that went to 100K: https://www.prosoundnetwork.com/gear-and-technology/review-sanken-co-100k-100-khz-super-wide-range-condenser-microphone I think its part of my theory MQA will only show its colors as technology progresses. At the moment much less than 1% of material is more than 48K - microphones may be one reason why - I do not know. But even if the mike doesn't go that high introducing filters causes smearing, ringing etc as the paper explained. Thanks Bill
  4. You cant hear those frequencies - 20k is the limit. What is going on is when you limit your original recording to whatever frequency you are releasing it at by the usual means which is a linear phase response brick-wall filter it creates ringing over many cycles - meaning its in the audible range. This is particularly objectionable because some of it occurs before the actual spike or whatever caused it and the ear finds that particularly bad - you never hear sounds before they occur in nature. Its the whole reason for MQA. They use various methods to prevent the ringing - see the paper posted many times: http://www.aes.org/e-lib/browse.cfm?elib=17501 See figure 14 and how many cycles the ringing is even at 192k and how short MQA is. Ringing for this long is likely audicle even at 192k (and combined with some of it being before the sound that initiated it is particularly objectionable to the ear) - and it is no where near as good as MQA. That's why the MQA guys claim it BETTER than 192k - but Paul from PS Audio disagrees. Now if the MQA trick does not work properly and some actual musical content over 44.1k gets through then some ringing may get though, it is claimed by MQA they have ways to handle this - we will see how effective it is. It may be as claimed and MQA is all we need. Another way of handling it is by means of minimum phase filters like the Ayre. But it must be done both while creating the album and playing back - and is not as good at it as MQA. But if there is only noise above 48k just doing it at replay is fine - no real need to go for the better MQA filters on replay. Thanks Bill
  5. Yes that is exactly what I am saying. I have a special program that tells me how high material really goes before its just noise. I have run it over my extensive HD library and very few tracks have material above 48k just as MQA claims. But I have one genuine DXD album that when you run it through the program has musical material all the way out to 176k - likely if they recorded it at higher than DXD it would go further - how high it actually goes - who knows - but I can tell you technically its the best sounding album I have: http://www.carmengomes.com/Discografie/CarmenGomessingstheBlues.aspx Now what I did is used a program to convert it to 44.1. Since it has musical content and not just noise out to 176k music will bang against the filter and create ringing. Ringing extends over a number of cycles so even though you can hear only to 20k this is audible. And it was - when I compered it along with others you could easily hear it. Now MQA is not like the normal brick-wall filter of usual down-sampling, but does show limiting it to 44.1k is likely audible - to stop it you probably need to increase how high it goes a lot further. MQA has a secret method to handle such situations. As more modern recordings appear that actually has content that high we will see how effective it really is. I posted before Paul from PS Audio found with 192k sampling it was close - but the originals were still better - they clain it better than the original. We will see what eventuates. MQA's answer may be simply releasing it as 88.2 or 176 and using their usual tricks - who knows - but they do claim some method of resolving it. Thanks Bill
  6. MQA is a technique for restoring the information at any rate. 12Mhz will cover any actual musical content easily and its use should advance the number of recordings that are more than just noise above 48k bringing into play the secret method they have to reproduce it. At the moment pretty much everything is captured by the first unfold which anyone can do - no licence or special decoder required. You can use any reasonable method to upsample it - it makes no difference - although you would probably be bonkers using a linear phase brick wall at 48k - theoretically it would make no difference - since its just noise banging against the filter and who cares if noise pre-rings it will just sound like noise to our ears. However they went to the trouble of using methods that greatly reduce or eliminate ringing so it's reasonable to use such a filter - an example would be Ayre's minimum phase filter: https://www.ayre.com/pdf/Ayre_MP_White_Paper.pdf This is now - but if you have material with musical content above 48k this could be different. I have a DXD recording that genuinely has musical material up to 176k. It can be done. I did a comparison with that down-sampled to 88.2 and it was blurry and all over the place compared to the original - the stuff above 44.1k was banging against the filter - it wasn't just noise and you could hear it. Will MQA reproduce the full musical content of such recordings when they become more common - only time will tell. It will test MQA's secret method of extending frequency. My guess is MQA will then be found out as just a stop gap until other methods are used to convey the full spectrum of the recording. How would you do that - here is one way - locate where it becomes just noise and the following: https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/Malvar_DCC07.pdf You include in metadata for very high frequencies the value it must exceed to be considered music after analysing it. It will be larger than MQA but I don't think by much because the high frequency will be such small values easily compressed. Thanks Bill
  7. Nope - anybody can stream MQA either as is or with the first unfold done. Its 'magic' is on the second unfold. However it must be said except for much less than 1% of material there is nothing above 48K, so the value of the second unfold is debatable. Its the, for now, very rare material with content above 48k that benefits - and exactly how it recovers stuff above 48 k that is more than just noise is a closely kept secret. I have said it before and will say it again - MQA's true worth will require better recording with content above 48k. As of now just get something like Audirvana or Amarra (the windows version will soon do it - the mac version already does) is enough. It really is something for the future - only the first unfold is of value now. The new Mytek ADC for example converts from an amazing 12mhz: https://mytekdigital.com/brooklyn-adc/ Once that sort of sampling becomes standard we will know MQA's true value IMHO. Thanks Bill
  8. And those economic forces include the near BS MQA people are shoving down the public. I say something very simple I have investigated its technicalities. Listen and decide for yourself and make your purchasing decisions on that. Its easy - don't be sidetracked. As for the DS and MQA I own both a DS with a bridge and a DSJ that has a built in bridge. It does full certified MQA decoding to 192k max. It was put in because DS customers demanded it - Paul, the PS Audio owner is not a fan - he did it just for the customers - not because he wants it - he doesn't - in fact he a bit of a critic - but for reasons slightly different than usual. He thinks it sounds better than 48k - but not as good as 192k - his issue is they said it would sound better - it was close but of the 192k masters converted to MQA to him the original sounded better. He doesn't like being told 'porkies' or what he considers porkies. MQA decodes to what the source was - even up to 784k and theoretically beyond - some MQA ADC's's even convert from 12mhz (I think its the Mytec). When it converts it back and the DAC does not go as high as the source the MQA was created from, the filter used in the DAC is designed to go as high as the DAC will go - in the case of the DS 192k. If its audible is unknown - but I doubt it - still one never knows. Thanks Bill
  9. Exactly. The person that did it was surprised. It was very rare - but a few people can consistently do it. Can that be extended from 23 bit dithered to true 24 bits - as I said I don't think tests have been done - but we will not know until they are. My opinion -- about as likely as a snowflake in the Sahara. Thank's Bill
  10. The corporate body pushing MQA is part of he market just like an advertisement is. Thanks Bill
  11. Of course there are issues with MQA eg it gets effective about 23 bits resolution - still below the thermal noise limit - by dithering - not by actually being 23 bits. There are others as well I have detailed elsewhere, but as I try to emphasize its all just a trade-off and you get one thing better at the expense of another. The market will decide if the MQA trade-off is the way to go - not us. I am a big fan personally - but that means nothing - it's what the majority think that matters to its success or not. All our discussion matters naught- with one exception actually understanding what its doing which I have posted quite a bit about. But that doesn't matter at all regarding if you like it or not, its final place in Hi Fi, success etc - understanding it is just an interesting exercise. In more detail the folding is inaudible because the filter they use makes anything above 48khz less than 16bits. That means any folding is less than 16 bits. They only transmit the first 16 its so the folding never even gets through. They restore nearly 24 bits by the above mentioned dithering. Is it as good as actual 24 bits - I don't think double blind tests have been done - but surprisingly tests have been done that show a few people can hear below 20 bits which is generally considered the thermal noise limit so I suspect there are a few people that can tell actual 24 bits from effective 23 bits dithered. MQA also does a statistical analysis of the music to determine exactly what number of bits they should transmit and what filter to use - 16 is about average. Blurbs about it use triangular sampling as an example - what they actually use is something more complicated and determined by analysing the music. Details of this are a very closely kept secret, as well as how they transmit frequencies above 48k for the much less than 1% of recordings that have it. I have said it before, and will say it again as recording technology gets better more and more stuff will contain musical information above 48k. That will be its real test. At the moment full unfolding is pretty useless since with the first unfold its lossless to 48k (well almost - there is a slight frequency drop above about 20k) and very very few recordings now exceed that. Thanks Bill
  12. Hi All The source will be the best source I know - a little notebook I have running HQ Player with a SSD disk all my main music is stored on. The notebook does not have the puff to do DSD conversion although it is one of HQ Players main claims to fame. Would love to have a Macbook with the latest Audirvana available - that does the lot including MQA. Have put off getting one - a Widows version is supposedly coming out. So if anyone cn bring one - great. Thanks Bill
  13. Rawl has tastes a lot more varied than me - he is also into Classical which I am not. That's a good thing - he got me onto a group called the Swingles that are simply amazing - Capella was not my thing until he played them - amazing. Thanks Bill
  14. Of course he isn't - just not my preference. Don't worry someone there will play some Cohen. I think I actually have every album he ever made because quite a few do like him. Thanks Bill
  15. Its not evil - I just do not like it and the KD exposes those bits I do not like - its gravelly and very close miked - not my cup of tea. Mike Lenehan for example loves it and Leonard Cohen in general - just a personal preference thing and believe me I do get teased about it. Thanks Bill
×