Home of the Squeezebox™ & Transporter® network music players.
Page 1 of 2 12 LastLast
Results 1 to 10 of 14
  1. #1
    Senior Member Archimago's Avatar
    Join Date
    Nov 2005
    Posts
    1,087

    INTERNET BLIND TEST: MQA Core Decoding vs. Standard Hi-Res Audio

    Hey boys and girls...

    Been meaning to do this for awhile but now that there's software decoding and we know better about the digital filter, I think it's time to let the world have a listen to what MQA Decoding sounds like compared to standard 24/88 or 24/96 resampling in a controlled fashion. If you're curious about MQA (especially if you have not listened for yourself), come check out a little blind test I'm conducting.

    Download the files, have a listen, complete my survey!

    Will be interested in what you, the listener perceives...

    INTERNET BLIND TEST: MQA Core Decoding vs. Standard Hi-Res Audio

    BTW: Test will close for survey entries on September 8th, 2017.
    Last edited by Archimago; 2017-07-15 at 09:08.
    Archimago's Musings: (archimago.blogspot.com) A 'more objective' audiophile blog.

  2. #2
    Senior Member
    Join Date
    Feb 2017
    Posts
    327

    Your survey & some thoughts on the 24bit/96kHz limitation of the Transporter

    Quote Originally Posted by Archimago View Post
    Download the files, have a listen, complete my survey!

    INTERNET BLIND TEST: MQA Core Decoding vs. Standard Hi-Res Audio

    BTW: Test will close for survey entries on September 8th, 2017.
    Hi Archimago!

    I shall certainly give your test a go.

    I note your insistence in your accompanying notes that the digital input to the DAC should not be "capped" at 24/96. Although I normally use a Transporter in my system, this requirement will not present a problem since I have a (or 4!) Squeezebox Touch which I can connect to my 24/192 capable external DAC via S/PDIF with an Ethernet link to my NAS running LMS version 7.7.6-113 (simply because it's the version which Synology make readily available, & I haven't yet found the need for one of Mr Herger's more recent updates).

    However, this did set me thinking. I know that the question of whether the Transporter could be upgraded to handle 24/192 material has been raised on this forum before, & the issue has been dismissed because its architecture & processing power will definitely not support material of higher resolution than 24/96: however, I am using my Transporter simply to supply a digital data stream to my DAC, & do not even need the Transporter's internal clock function (because I have a word clock link from my DAC) let alone its DAC & analogue stage functions. It occurs to me that this might yet be capable of changing things.

    I found an interesting series of postings on appropriate digital output connections by Sean Adams himself, where he was very patiently trying to explain to someone (who had a clear excess of wealth to grey matter ratio ) that feeding the Transporter word clock input from his very expensive dCs outboard Master Clock before linking to his equally expensive dCs DAC was a BAD IDEA because of the increase in jitter that would inevitably occur: http://forums.slimdevices.com/showthread.php?39770-Setting-Transporter-to-Slave-for-World-Clock-Input/page3 & the independently authored digital audio theory exposition that Sean referenced (in post #30) to support his contention: http://www.tnt-audio.com/clinica/diginterf2_e.html. Although the latter discussion is framed in the context of using CD/SACD transports with external DACs, the same principles will clearly apply to any "2 box" source + DAC set-up. It is the approved (also in Sean's post #30 above) "Clock Backwards" configuration as described in the technical article which I am using. Of course, back in 2008, 24/96 itself was still considered to be "hi res", being the DVD-A standard, & I don't think that higher sampling rates were seriously envisaged back then, even though all 3 of the digital connections types used on the Transporter (AES/EBU, S/PDIF & TOSLINK) will in fact support data streams up to a maximum of 24/192.

    The in-built DAC (& subsequent analogue stage) of my Transporter continues to function when the external word clock is selected via LMS settings, although I do not use its output. I cannot find any LMS setting option to turn off the in-built DAC hardware if you are only using the digital outputs to feed an external DAC & this fact might present an obstacle to my idea. But bear with me...

    My understanding is that LMS automatically down-samples 24/192 to 24/96 (& 24/176.4 to 24/88.2) before passing the data stream onwards when its target is identified as a Transporter (as opposed to a later Squeezebox Touch which is 24/192 capable, but lacks the handy word clock in necessary to establish a low jitter connection to an external DAC). I am wondering what would happen in my specific configuration if that down-sampling were not performed by LMS: after all, the data being passed on from my Transporter to my DAC is just a stream of bits, & I have to set the sampling rate to be used to convert it to an analogue signal on the DAC itself.

    I imagine that in the Transporter itself the conversion of 24/192 input stream would either not occur at all (with silence in the analogue stage), or that it would be done at the in-built DAC's maximum of 96kHz resulting in a reconstructed analogue signal running at half-speed. Neither of these outcomes would be of any significance to me since I'm not using the Transporter's analogue outputs at all. Nor should they trouble the Transporter's analogue stage.

    OTOH, if passing a native 24/192 signal to my Transporter would cause it to malfunction in any way or even cause damage, this would clearly be another BAD IDEA! Does anyone have sufficient understanding of the Transporter's circuitry to provide clarification on how it would handle a 24/192 data stream?

    If there would not be any hardware issues, it would seem to be a straightforward software revision for LMS to disable the automatic 24/192 & 24/176.4 down-sampling for the Transporter if the word clock input is selected in LMS settings, since that would appear to imply exclusively that my type of connection to an external DAC is being used.

    So maybe, just maybe, the Transporter could be made to handle 24/192 signals, even if only in conjunction with a word clock out equipped external DAC, with only a small mod to LMS...

    If no-one knows how the Transporter would respond to a native 24/192 or 24/176.4 signal, I'll pass the idea over to Mr Herger for his consideration.

    Anyway, I look forward to taking part in your test, Archimago, & returning my findings together with the other information you have requested!

    Dave
    Last edited by Golden Earring; 2017-07-16 at 07:13.

  3. #3
    Senior Member Archimago's Avatar
    Join Date
    Nov 2005
    Posts
    1,087
    Thanks for the comments GE...

    Well, here in 2017, I still consider 24/96 "hi-res" of course. Last I checked, my kids' ears apparently have not evolved enough to render 24/96 inadequate :-).

    Yes, the digital interfaces AES/EBU, TosLink, coaxial can be used for >24/96 but the Transporter does not support this. My understanding is that the limitation here is with the Ubicom IP3K CPU, 28Mbits RAM, and the embedded operating system. It is an interesting idea nonetheless and even if the hardware is capable of the hack, remember that you'll need to stream lossless PCM over the ethernet as FLAC decompression is likely out of the question especially over the 802.11g WiFi.

    These days, for <$100 you could just stream with a Raspberry Pi 3 or equivalent to your heart's content even beyond 192kHz over USB of course...

    Looking forward to your blind test entry! Good to see you have the Touch for 176.4/192 playback. Remember the only reason I went with the 4x samplerate is to capture a bit of the digital filter effect from MQA to get closer to "the sound" given how much they tout the effect of their variety of digital filtering. Unofficially, of course one could just play back over 96kHz; so long as the downsampling is done the same way, one should still be able to hear any significant differences between the MQA Core decode and a standard hi-res downsample.

    Enjoy...
    Archimago's Musings: (archimago.blogspot.com) A 'more objective' audiophile blog.

  4. #4
    Senior Member
    Join Date
    Apr 2013
    Location
    UK
    Posts
    944
    What are the origins of 2L's mqa files? I remember looking at some of their sample material before. Not my type of music I have to say, hard for me to get into it..


    Transcoded from Matt's brain by Tapatalk
    --
    Hardware: 3x Touch, 1x Radio, 2x Receivers, 1 HP Microserver NAS with Debian+LMS 7.9.0
    Music: ~1300 CDs, as 450 GB of 16/44k FLACs. No less than 3x 24/44k albums..

  5. #5
    Senior Member Archimago's Avatar
    Join Date
    Nov 2005
    Posts
    1,087
    Quote Originally Posted by drmatt View Post
    What are the origins of 2L's mqa files? I remember looking at some of their sample material before. Not my type of music I have to say, hard for me to get into it..


    Transcoded from Matt's brain by Tapatalk
    Hi Matt:
    The original MQA files can be found here (along with originals up to DXD):
    http://www.2l.no/hires/

    As you can see, I picked 3 of them based on variation on the type of sound (vocals, piano, violin). Yeah, all classical, not exactly the type of thing I listen to daily. However, considering the fact that much of MQA on Tidal is just poorly mastered or old non-hi-res recordings, this is as good as I think we can get when trying to try listening for differences... Especially considering they're freely available and meant for evaluation.
    Archimago's Musings: (archimago.blogspot.com) A 'more objective' audiophile blog.

  6. #6
    Senior Member
    Join Date
    Apr 2013
    Location
    UK
    Posts
    944
    Yep it's probably one of the best sources there is to test with. I have downloaded your samples and appreciate your concise tagging efforts! Had visions of them disappearing all over my library and having to browse by file location..

    EDO on the SBT works fine, I don't see "192k" written on my DAC very often, but there it is..!


    Transcoded from Matt's brain by Tapatalk
    --
    Hardware: 3x Touch, 1x Radio, 2x Receivers, 1 HP Microserver NAS with Debian+LMS 7.9.0
    Music: ~1300 CDs, as 450 GB of 16/44k FLACs. No less than 3x 24/44k albums..

  7. #7
    Senior Member
    Join Date
    Feb 2017
    Posts
    327

    Techie stuff...

    Quote Originally Posted by Archimago View Post
    Well, here in 2017, I still consider 24/96 "hi-res" of course. Last I checked, my kids' ears apparently have not evolved enough to render 24/96 inadequate :-).

    Yes, the digital interfaces AES/EBU, TosLink, coaxial can be used for >24/96 but the Transporter does not support this. My understanding is that the limitation here is with the Ubicom IP3K CPU, 28Mbits RAM, and the embedded operating system. It is an interesting idea nonetheless and even if the hardware is capable of the hack, remember that you'll need to stream lossless PCM over the ethernet as FLAC decompression is likely out of the question especially over the 802.11g WiFi.

    These days, for <$100 you could just stream with a Raspberry Pi 3 or equivalent to your heart's content even beyond 192kHz over USB of course...

    Unofficially, of course one could just play back over 96kHz; so long as the downsampling is done the same way, one should still be able to hear any significant differences between the MQA Core decode and a standard hi-res downsample.
    Hi Archimago!

    Firstly, apologies for "cherry-picking" your earlier comment in the quotation above: I'm just trying to focus on the bits of particular relevance to the ensuing comments - if you feel that I've removed the context, I apologise & stand to be corrected!

    Obviously your remarks about the internal processing capabilities of the Transporter are pretty technical, but if I've understood correctly you are saying that it is incapable of outputting a decoded data stream above 24/96 through any of its digital outputs even though in theory the connections themselves could support them. I think that this puts my idea to rest in practical terms. In any event, both my pet bat & I agree with you that anything above 24/96 should have no audible effect! So for all practical purposes (other than your test) I'm quite happy with it as is.

    I take your point about the exceptional price/performance ratio of the Raspberry Pi, but I've had a spot of good fortune in that I've managed to acquire a mid-2010 i5-based 27" iMac with Bluetooth remote keyboard, barely used, for £400. I think that's one of my better bargains given that they were £1649 when new! With its high resolution screen (2560 x 1440) I can attach it to the wall on the other side of "my" corner of the room & operate it without getting up from my sofa (I am a bit lazy these days). I picked up the matching Apple "Magic Track Pad" (also Bluetooth) s/hand & also a frame that holds it to the right of the keyboard (both convenient since I am right-handed & making the combined unit wide enough to sit snugly on my lap - the keyboard on its own is so small as to be awkward to balance) from eBay, so that should enable me to connect into the USB 2.0 input on my Mytek DAC which goes up to a mind-bending 32/384 (or DSDx256). Can't really see the point but at least I'm a bit future proof if things carry on getting sillier...

    Anyway, I've always fancied a Mac but have been put off by the cost in the past. It'll be interesting getting my Windows 10 PC, OS Sierra iMac & Linux Synology NAS talking to each other! I have the NAS, the Transporter, my Panasonic Freeview HD BD-R + HDD & my 2010 Sony TV on Ethernet at the moment. In fact there are now no useful apps still supported on the TV, so I'll simply disconnect that from my router altogether (leaving its HDMI connection to the Panasonic gizmo, of course) & use that 4th final Ethernet port on the router for the static iMac. Other stuff (laptop PC, printer, 'phone, couple of Touches synced to Transporter displaying album art & current track info) all connect via Wi-Fi & operate faultlessly, at least so far.

    But apart from running experimental stuff via USB on the iMac, I'll leave LMS running on my NAS & accept the 24/96 limitation of my Transporter even as a purely digital source.

    Thanks for your technical insight!

    Dave
    Last edited by Golden Earring; 2017-07-18 at 23:53.

  8. #8
    Will have to take the test as i own a Meridian Explorer 2 only problem is i live in the world of budget HiFi so running on a PiCorePlayer with a Yamaha A-S500 and a set of JPW Sonatas but will give it a go

    PS i normally listen to electronic music but can appreciate classical music as playing a instrument is sadly a talent i will never have plus Kim André Arnesen's MAGNIFICAT was a demo track i got with my Explorer 2 with is stunning

  9. #9
    Senior Member Archimago's Avatar
    Join Date
    Nov 2005
    Posts
    1,087
    Quote Originally Posted by Golden Earring View Post
    Hi Archimago!

    Firstly, apologies for "cherry-picking" your earlier comment in the quotation above: I'm just trying to focus on the bits of particular relevance to the ensuing comments - if you feel that I've removed the context, I apologise & stand to be corrected!

    Obviously your remarks about the internal processing capabilities of the Transporter are pretty technical, but if I've understood correctly you are saying that it is incapable of outputting a decoded data stream above 24/96 through any of its digital outputs even though in theory the connections themselves could support them. I think that this puts my idea to rest in practical terms. In any event, both my pet bat & I agree with you that anything above 24/96 should have no audible effect! So for all practical purposes (other than your test) I'm quite happy with it as is.

    I take your point about the exceptional price/performance ratio of the Raspberry Pi, but I've had a spot of good fortune in that I've managed to acquire a mid-2010 i5-based 27" iMac with Bluetooth remote keyboard, barely used, for £400. I think that's one of my better bargains given that they were £1649 when new! With its high resolution screen (2560 x 1440) I can attach it to the wall on the other side of "my" corner of the room & operate it without getting up from my sofa (I am a bit lazy these days). I picked up the matching Apple "Magic Track Pad" (also Bluetooth) s/hand & also a frame that holds it to the right of the keyboard (both convenient since I am right-handed & making the combined unit wide enough to sit snugly on my lap - the keyboard on its own is so small as to be awkward to balance) from eBay, so that should enable me to connect into the USB 2.0 input on my Mytek DAC which goes up to a mind-bending 32/384 (or DSDx256). Can't really see the point but at least I'm a bit future proof if things carry on getting sillier...

    Anyway, I've always fancied a Mac but have been put off by the cost in the past. It'll be interesting getting my Windows 10 PC, OS Sierra iMac & Linux Synology NAS talking to each other! I have the NAS, the Transporter, my Panasonic Freeview HD BD-R + HDD & my 2010 Sony TV on Ethernet at the moment. In fact there are now no useful apps still supported on the TV, so I'll simply disconnect that from my router altogether (leaving its HDMI connection to the Panasonic gizmo, of course) & use that 4th final Ethernet port on the router for the static iMac. Other stuff (laptop PC, printer, 'phone, couple of Touches synced to Transporter displaying album art & current track info) all connect via Wi-Fi & operate faultlessly, at least so far.

    But apart from running experimental stuff via USB on the iMac, I'll leave LMS running on my NAS & accept the 24/96 limitation of my Transporter even as a purely digital source.

    Thanks for your technical insight!

    Dave
    That's a nice selection of devices Dave :-).

    About a decade ago, I remember feeling "disappointed" that my Transporter was "only" capable of 24/96. Over the years with experience, recognition that many 176.4/192 recordings have no content in the highest frequencies, and of course recognition that my ears/brain complex is just not able to appreciate the difference, when it comes to listening to music, it ends up being the Transporter that I fire up with room correction done by LMS and BruteFIR.

    Of course the Mytek DAC like what you have or my Oppo Sonica DAC with higher frequency rates and DSD128+ would be great for upsampler experimentation, trying different filters, etc...

    In any event, enjoy the sound! And have fun with experimenting :-).
    Archimago's Musings: (archimago.blogspot.com) A 'more objective' audiophile blog.

  10. #10
    Senior Member Archimago's Avatar
    Join Date
    Nov 2005
    Posts
    1,087
    Quote Originally Posted by Eyerex View Post
    Will have to take the test as i own a Meridian Explorer 2 only problem is i live in the world of budget HiFi so running on a PiCorePlayer with a Yamaha A-S500 and a set of JPW Sonatas but will give it a go

    PS i normally listen to electronic music but can appreciate classical music as playing a instrument is sadly a talent i will never have plus Kim André Arnesen's MAGNIFICAT was a demo track i got with my Explorer 2 with is stunning
    Your Yamaha / Meridian Explorer 2 / JPW Sonata would be a great system to test with. As I mentioned in the blog, there's certainly no need for megabuck systems! After all if MQA is a significant upgrade in sound quality that's worthwhile enough for TIDAL streaming to everything from megabuck systems to headphone users, a system such as yours and your impression could/should/would be a very fine test subject along the wide price-points audiophiles use.

    Enjoy!
    Archimago's Musings: (archimago.blogspot.com) A 'more objective' audiophile blog.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •