Sunday, 14 February 2016

MEASUREMENTS / IMPRESSIONS: Meridian Explorer2 Analogue Output - 24/192 PCM vs. Decoded MQA.

Image from here.
Well folks, the new Meridian firmware has been released, allowing for MQA decoding. And thanks to a friend who has an Explorer2 with the recent update, he was able to record the analogue output of the DAC playing a standard 24/192 PCM track and the same song as a decoded MQA (24/44)! This is the obvious next step in evaluating MQA after my previous post listening to and evaluating the undecoded MQA file.

The song selected for testing/listening was from 2L (of course) - the track "Blågutten" from the Hoff Ensemble's album Quiet Winter Nights (2011, MQA downloadable here). The recording ADC device is the excellent professional quality RME Fireface 802.

Of course re-digitizing degrades the signal very slightly but the Fireface is an excellent ADC so whatever slight change is likely very minimal though arguably it'll add some "blurring" to the signal if we accept Meridian's claims.

Based on my friend's tech notes, the Explorer2 puts out a peak 2.1Vrms and he recorded this signal with the Fireface 802 set at +4dBu allowing for -5dBFS headroom. No further volume adjustment was made between the PCM playback and MQA. We can see this when we analyze the amplitude statistics for the track:


We see that both the 24/192 PCM and MQA decode are almost exactly the same peak and average amplitudes (<<1dB difference). This is good, it means the MQA process isn't changing the output volume making comparisons like A-B testing possible without needing to do anything extra (ie. volume compensation).

What is most easy to see and as I had wondered about is the loss of resolution in the dynamic range due to the compression / "encapsulation" / "origami" process. In this sample, if we look at the noise level at very close to the same (within milliseconds) place in the song near the end as it fades out, this is what we see:


This is the Meridian Explorer2 playing the 24/192 PCM version (ie. the non-MQA version) of the track. Notice the excellent noise floor up to ~45kHz (BTW this is just with Adobe Audition 3, FFT size 8192). After this, there is a gradual rise in the noise floor which I believe is modulator noise primarily from the original recording although the Fireface ADC might be adding to it. The excellence of the noise level clearly speaks to the superb recording quality from 2L and their ability to capture >16-bits of digital resolution.

What does this look like out of the Explorer2 with MQA decoding?

What you see above is clearly a rise in the absolute noise floor as expected because the lowest bits of the 24-bit data is being used for the "encapsulation". That noise floor in fact doesn't get much better than what you see above suggesting that indeed the original 24-bit signal has been reduce and probably a fair amount of dithering noise applied (looks to me at best to be 16-bits with a fair amount, maybe 2 bits, of dithering).

Out of curiosity, I ran another ABX test to see if I could hear the difference, the foobar ABX 2.0 tool was used and here's the result:

Compared to last time when I tried this with the 16/44 source file, it looks like I was better able to identify the difference this time! As before, I used my desktop firmware upgraded ASUS Essence One with Sennheiser HD800 headphones in a very quiet room using a silent fanless computer setup. I was honestly a bit surprised at the result because I certainly did not feel that confident at the end of the ABX session. However, I did try to be more deliberate and "serious" than last time and took a number of minutes at the start to listen between the two files and pick a place to start where I thought I could familiarize myself with the slightly different sound. Like before, the trick for me was to listen to the tonality of the two files with instantaneous switching. Again, the telling component was that the MQA decoded recording did sound a little "clearer" and more "sparkly" - perhaps more accented with the percussion hits and cymbal trails. As previously mentioned, it was similar to a TV where the rendering seemed "sharper" in terms of local contrast enhancement with the MQA file. I think one could say there seemed to be more "air", maybe like adding a little 10kHz boost kind of effect. As I had expressed before, I'm not sure whether this subjective effect would be universally considered to be "better". The PCM 24/192 sounded very smooth and natural already.

Note that even though I was aware that the MQA decoded file had a higher noise floor, I did not purposely take advantage of this by listening to a quiet portion of the file with volume pumped up to listen to the hiss. That would be cheating :-).

Another way to look at the difference between the two files is using the Audio DiffMaker program to create a "difference" file. So, running the decoded MQA and subtracting the PCM playback through it, we get a pretty good ~74dB average correlated null depth using the first 15 seconds of the audio (with compensation for sample rate drift from the ADC):
This level of correlated null depth basically tells us that the difference is not going to be a "slam dunk", or obvious. I concur given my subjective lack of confidence going through the ABX.

If we look at the "difference" FFT, at any one instance, it looks like this:

I can hear the musical residue but shifted to a higher pitch when playing the "difference" file. If I pump up the volume, the noise tonally is higher with a rising noise level above 5kHz as shown in the FFT. Perhaps this accounts for the tonality change I heard when ABX'ing. If we show the "spectral frequency display" using Audition 3:
Spectral frequency display. Ignore the signal in the first 200ms - likely just DiffMaker not quite achieving the time compensation yet as it aligns the samples.
We see that the MQA version has "added" a band of noise to the signal especially between 15-40kHz. There's a second band of noise higher up from 75-96kHz which could be a combination of noise added by the ADC from the 2L studio and/or the Fireface ADC used to record the analogue output.

Amplitude statistics of the "difference" file (first 250ms discounted from analysis)... Although on average the difference is low, down at -70dB RMS, there were differences up to -30dB peaks.

Conclusions:

I think what is interesting about the listening experiment is that the difference was noticeable even with the ADC (Fireface) / DAC (Essence One) steps in between. This tells me that the "de-blur" DSP processing implemented by Meridian/MQA can in fact be captured through the analogue output though presumably the effect may not be as strong as listening directly (this is what my friend has noticed). Nonetheless, I can reasonably prove to myself that I was able to hear the difference in ABX testing.

Bottom line is this... So far, what I have seen of decoded MQA remains consistent with suspicions I had previously.

1. First, there is a "de-blur" DSP being applied which supposedly improves time-domain accuracy (presumably based on measured parameters of the equipment / processing steps used in the studio). I think this is the primary audible component. This is why they claim even an undecoded MQA file played through a standard DAC sounds "better"; the DSP effect is already "baked into" the sound even without the decoding process to expand the ultrasonics.

2. Second, the "encapsulation" process takes over the lower bits in the 24-bit data for data "compression". Hence we see a higher noise floor in the MQA decoded file. Based on the sample in this post, it still looks like MQA is not capable of >16-bit dynamic range when decoded.

3. Thirdly, a decoding algorithm reconstitutes the ultrasonic frequencies for final playback on the DAC to 192kHz or higher if the DAC is capable. I assume for the Explorer2, Meridian would be trying for their typical apodizing minimum phase digital filter. There is of course no way to know for sure unless we had access to tools to encode test signals. I doubt if this makes any audible difference in any case. The example here basically shows a bunch of noise added to the ultrasonic frequencies even in places where there wasn't any noise in the original 24/192 PCM file. Clearly the decoded MQA file is short of a true "lossless" reconstruction!

Given that I believe Step 1 above is the most likely to change sonic quality, I really wonder what standard 24/96 or 24/192 "digital masters" sound like with the DSP applied to "correct" time-domain parameters in the studio before "encapsulation" takes out the lowest bits.

Realizing that Meridian/MQA has provided essentially no technical details or objective results, if I am correct about what is going on as described above, I am personally not interested in MQA as a format I feel I would want. There’s no “magic” here and there are evident compromises when trying to be everything to everyone as MQA seems to be aiming for. It's aimed at data compression, being “compatible” with regular DACs, able to deliver “high resolution”, and “better sounding” even with regular 16/44. I wish Meridian/MQA success in presumably the target niche (high-resolution streaming like TIDAL) even though I think there are better ways to do this (ie. simply stream 24/48, or something like FLAC compressed 18/96 as Miska and others have discussed).

I honestly hope no studio implements this as some kind of archival alternative nor does this supplant digital downloads where true 24-bits and high samplerate are already within the grasp of consumers through places like HDTracks, Pono, etc... A "flat", standard PCM 24/88+ is compatible with essentially any DAC and capable of full sonic quality for one's music library already if one believes there is a need to even go beyond 16/44. As is often the case, when something tries to be the "jack of all trades", generally "master of none" is a typical outcome.

A big thank you to my friend with the Explorer2 and studio quality ADC for taking the time to do this and being so meticulous with the whole recording process!

[One last thing: Meridian/MQA - please stop using the description "revolutionary" for this product. I'm quite sure there are many of us who cringe every time we hear this...]
-------------------------

Have a great week ahead everyone! Enjoy the music...

Addendum:
Well, well, well... It looks like maybe the accolades for MQA isn't universal after all! Here's PS Audio's Paul McGowan's impression.

Addendum 2:
For the purpose of further clarification of what I'm seeing in the noise floor, here's the composite image of the various releases of the same track analyzed at the same place (3:49.000 into the track) with DSD64 and undecoded MQA played back using an ESS9018 DAC, compared to the Meridian Explorer2 playing back the "studio master" 24/192 PCM and the MQA track but this time decoded. Each sample recorded with the same RME Fireface ADC.


Though we can perhaps look at other samples, it's quite clear from this example that by 10kHz, the decoded MQA signal is unable to achieve the same resolution as the 24-bit PCM; in fact, not surprisingly, it follows the undecoded MQA which itself is subject to digital filter parameters for 44kHz playback. In comparison, we can look at DSD64 (SACD resolution) which follows 24/192 until about 24kHz then deviates with the large amount of ultrasonic grunge from noise shaping.

BTW: The reference was set at -20dB so as to show the relative noise floors a bit better (as I circled at the bottom of the graph).

56 comments:

  1. It would be nice to see a non MQA recording (or file) with no noise in the ultra high frequencies and then see if the MQA process adds some. I remember looking at the L2 site and some files have noise rising at 40kHz. Not sure why but this is unwanted.

    ReplyDelete
    Replies
    1. Hi Ran. Yeah, alas until we have other studios providing MQA samples to listen to and test out, the 2L high-res samples all have a tendency to demonstrate an increasing noise floor from ~40kHz on up. (Easily seen with any of their 24/192+ samples.)

      In any event, clearly the MQA decode has added noise and increased the noise floor in the audible spectrum already.

      Delete
    2. I can understand the increased noise floor but the the noisy high frequencies should not be there. If that is the outcome of some fancy DSP then there is a big question mark on the fidelity when compared to the original. It will be very interesting when studios will provide the "before and after" files. Until then, I am on the fence.

      Delete
    3. Try the 2L-111 track. It's practically free of modulator noise.

      Delete
    4. The Fireface uses an AK4620 ADC. With one exception (the AK5394A) pretty much all the ADC's used for audio (including the AK4620) have a noise floor that rises around 30 KHz. This makes a bit of a mess looking for ultrasonic components in recordings if the ADC provides them.

      A handful of recording interfaces do use the AK5394A (Protools and the EMU "M" class I know) but a lot do not.

      Delete
  2. "but the the noisy high frequencies should not be there"

    Whenever you're filtering a digital bitstream to get analog music, you're balancing frequency domain and time domain performance. MQA makes much of its time domain performance ("de-blurring"), so obviously the balance is shifted away from the best performance at eliminating high frequency noise.

    ReplyDelete
  3. It would be interesting to see a capture of the I2S stream going into the DAC chip. Of course, obtaining one might entail voiding some warranties.

    ReplyDelete
  4. Noise Differences

    Hi Mans Rullgard. I can confirm, that in the above mentioned track, the difference in the playback noise (of the noise floor in the recording), between the PCM 192 kHz and the MQA 192 kHz, both with the Meridian Explorer 2 begins to differ apart at around 10 kHz and maxes at around 30 kHz, with the MQA encoded (to 192 kHz) is about 22 dB higher, than that of the equivalent plain PCM version.

    Compared with the DSD64 version of that file, the noise floor of the recording compared to the PCM 192 kHz version is the same until 25 kHz (so the DSD64 has lower noise up to 25 kHz than the MQA 192 encoded file) and the DSD64 noise crosses the MQA 192 noise at around 30 kHz.

    Juergen

    ReplyDelete
  5. Archimago,
    After reading this and previous posts about MQA, I am led to a slightly different conclusion. I would say: Of course, aside from marketing talk, there is no way to 'improve' original recording. But hats down to MQA for managing to sound almost indistinguishable from the original 24/192 recording while maintaining several times lower bit rate, which is great news for streaming services.

    By the way, can you compare 16/44 with the sme file passed through MQA?

    ReplyDelete
    Replies
    1. Hi Gok,
      But what's the audible difference between 24/48 and 24/192? Or even blind testing results from 16/44 compared to 24/192 to prove the value of hi-res in the first place?! We've discussed this many times here and as far as I'm concerned, the vast majority of music out there isn't recorded with any need for hi-res anyway.

      You see, at the end of the day, the difference is minimal if not indistinguishable already. The fact that I can ABX the MQA compared to the PCM 24/192 is because MQA *added* something to the sound. They put in some kind of DSP process that made it sound different (supposedly better). It's not about accuracy to the source 24/192 but basically a form of "remastering" of the original PCM audio to a sound quality which Meridian claims is more accurate in the time-domain to the original performance (before the ADC added supposed distortions to the signal).

      Although the "encapsulation" / "origami" technique seems smart, IMO it's more of a curiosity and unnecessary from a sound quality perspective. If there is to be a legacy to this technique, it's whether the DSP "deblur" portion has any merit as a valuable studio production technique.

      Delete
    2. I just added Addendum 2 which contains the undecoded 24/44 MQA file to compare with the decoded MQA file. A 16/44 file could be derived from the 24/192 PCM of course but not necessarily useful to show because noise characteristics would be more a function of the type and strength of dithering used.

      Delete
    3. In the original 2L recording, there is clearly rising noise from around 45 kHz upwards, but is there actually any *signal* up there: in other words, if you compared the 192 kHz file with a 96 kHz version, would the only difference be noise at frequencies above the 96 kHz version's Nyquist? If so, fully half of the 192's size/bandwidth is wasted on noise.

      Where is the "deblurring" actually stored? If it's part of the "main" 16 bits and accessible to ordinary DACs then couldn't we just have that and ignore questionable content that is either inaudible or mostly noise anyway? It still looks to me that MQA is not using size/bandwidth in the most efficient way.



      Delete
    4. Signal content up to?

      Hi "Otto Nikolaus". With the 192 kHz PCM file, I can clearly see (not shown on this page), that the signal content does go up to around 50 kHz (leaving aside, if this frequency is audible or not, but it has an influence on the timing behavior).

      As the MQA decoded (to 192 kHz) signal, does fade into the MQA noise at around 25 kHz (because the MQA noise is much higher at the 30 kHz area, than the plain 192 kHz PCM), but may also influence the timing.

      PS: The rise of the noise floor from 45 kHz upwards is still very low in level and in the above shown graph, the primary focus is the difference between PCM 192 and MQA 192, so the relative level and not the absolute one.

      Juergen

      Delete
    5. Thanks Juergen. So there is signal up to around 50 kHz, but this is very close to the 48 kHz cutoff for 96 kHz so doesn't my point about 192 kHz being a waste of space still stand, regardless of the audibility (or not) of such high frequencies?

      As a general point, is there *any* material available with content (not just noise) above 48-50 kHz? Are there transducers that go that high, at each end of the recording/listening chain?

      Delete
    6. My statement of the signal content in the above mentioned song is valid, when integrate the content of this song over the complete length. They are some small signals during the song, that do go over 50 kHz. Even this is in no way audible in the frequency range, but can influence the timing behavior and so the cues for localization etc. and identifying.

      For me personally, 96 kHz sample rate would be sufficient (and would only "waste" 1.5 times of the MQA Flac), if the recording and the playback would not be done with sharp brick wall filters, but with soft (analog type minimum phase) filters, that would start slowly at 30 kHz and would be infinite at half sample rate (so no alias at all).

      And this all only with moderate dynamic range compression. For me it would mean nothing, if the producer, mastering engineer would compress a song to "hell" (for example the actual CD from Adele has DR05) and would sign it as "authorize" = "this is the sound, that the artist, producer, ... intended". So even with the MQA blue light on, wouldn't change this DR05.

      Juergen

      Delete
    7. is there any documentation on the microphones used for the recording? Most recording microphones have an abrupt rolloff around 16-18 KHz. A few are extended to maybe 22 KHz. The smaller diaphragm microphones would only be useful for closer miking since the noise floor is proportional to the diaphragm size. This limitation really is a big question mark in the value of high resolution recordings. Even when a 1/2" microphone is specified to have response to 30 KHz (B&K 4133 for example http://www.bksv.com/doc/Bp0100.pdf ) that is for on axis only. If you look at the on axis to off axis response its pretty dramatic (see page 7 of the document). The screens around most recording microphones will have even more impact on the response.

      I think this is partly why the skill and sensitivity of the recording engineer using the available tools matters more than any hirez (real or pseudo) recording system. Personally I will continue to use high sample rate systems where possible but more to avoid artifacts from the brick wall filters than anything else. And maybe its placebo effect.

      Delete
    8. Thanks for this comment Demian,
      It hits the nail on the head... Whether one cares for hi-res or not, certainly by the time we get to 176 and 192kHz samplerates, we are way beyond the capabilities for microphones to pick up useful signal such that there is just *no content* up there anyone. It's essentially all noise!

      Delete
    9. It would be interesting to compare waveforms between the PCM originals and the various captures on a transient. I would really like to get a handle of the "time domain" errors they have found. I just don't see anything that would not be expected from a band limited system.

      Delete
    10. It would be interesting to compare waveforms between the PCM originals and the various captures on a transient. I would really like to get a handle of the "time domain" errors they have found. I just don't see anything that would not be expected from a band limited system.

      Delete
    11. Demian: Yup. Would be interesting... Alas unless MQA releases this or someone with the actual MQA mastering tools could run a test signal, it's all a mystery.

      As you suggest, I agree, I don't see anything unusual and the hub-bub around pre-/post-ringing is just the result of passing the "impulse" through a bandwidth-limited system; not reflective of actual music as if we could be perturbed by sound around 22+kHz!

      Delete
  6. This comment has been removed by the author.

    ReplyDelete
  7. NB A score of 8/19 has a p=0.055, which is just shy of meeting the standard threshold for 'significance'.

    ReplyDelete
    Replies
    1. typo there...meant 8/10 of course

      Delete
    2. IOW it indicates more testing is needed to confirm that this was audible.

      Delete
    3. Yes, of course. Remember, the DiffMaker difference file suggests the sound difference is subtle. But to be honest, that's a better score than I've generally achieved with blind testing 24/96 vs. 16/44!

      It's not meant to be a formal test of statistical significance for me, rather a shot at seeing if there is merit to the idea that MQA makes the sound a little different. And I think it does; though subtle enough that I'm not going to get overly excited.

      Having said this, I'd be impressed if any of the audiophile press reviewers out there will try to take a test like this for themselves before declaring that MQA makes an obvious or clear difference!

      Delete
    4. It's a night and day difference. If you can't hear it, your equipment, your ears, and your dog are crap. Besides, any kind of blind or ABX testing is flawed.

      There, saved you the time reading the audiophile press.

      Delete
    5. Thank Mans. You're without doubt a prophet :-).

      Delete
    6. If you think there's any hope of getting anything sensible from the audiophool press, try this (not about hi-res). http://www.enjoythemusic.com/hificritic/vol5_no3/listening_to_storage.htm
      Good job I've got a Synology NAS already because I can improve the SQ just by converting to a striped array! Now, it could be that this was dated 1 April but I couldn't see any date on the page.

      Delete
    7. Congrats Otto on your choice of NAS that is clearly upgradable to better sound quality!

      Alas, I am sadly just backing up with mirroring but I do use Windows Server 2012 R2 which I believe has been said to sound better than the typical Windows 8 or 10... :-)

      Delete
    8. For those wary of jumping down the rabbit hole, I quote 2 wonderful extracts:
      ...QNAP1 promoted the leading edge piano transients, following through with a lighter, brighter instrument tone – possibly Steinway-like? The same piano had more lower mid body on QNAP2 and slightly softer hammer impact, perhaps more like a Bosendörfer.
      (QNAP1 and QNAP2 are 2 different QNAP 4-bay NASes.)
      ...If anything, this track highlighted a fundamental shift in timbre between the storage sources. This wasn't the gentle tweak of a DAC's digital filter option; we felt it was more akin to changing loudspeakers. System sound was improved as if the DAC itself had been upgraded, say from a £500 to a £2000 model.
      So, by choice of NAS you can not only change the sound of a grand piano to what seems like another manufacturer, you can get as big a change as by upgrading your DAC or even speakers!

      Delete
    9. “Alice laughed. 'There's no use trying,' she said. 'One can't believe impossible things.'

      I daresay you haven't had much practice,' said the Queen. 'When I was your age, I always did it for half-an-hour a day. Why, sometimes I've believed as many as six impossible things before breakfast. There goes the shawl again!” --- Lewis Carroll

      You've mentioned a few of those "impossible things" in the above post and I just had breakfast. :-)

      Thanks for the humor!

      Delete
  8. Archimago,
    At the risk of committing a foul, I wonder if you could briefly comment on another area of "audible difference". I have tried to find an article you have written on this but have unable to locate one...
    Is there an audible difference between using an hdmi input vs a usb input, say specifically on the Oppo 105? I know audiophiles comment on jitter rates, etc., but have you either measured an audible difference or subjectively heard an audible difference?
    Thanks very much for your patience and expertise,
    sk16

    ReplyDelete
    Replies
    1. Hi sk16,
      Honestly I have not spent much time with HDMI. The reason being the only devices I have access to are receivers and of course with those there are many layers to deal with in those complex devices beyond the DAC function.

      I've posted the (tongue in cheek) look at HDMI cables and the fact that they make no difference to jitter or anything else I could find here:
      http://archimago.blogspot.ca/2013/11/guest-review-measurements-quantum-hdmi.html

      And there's these measurements with my Onkyo TX-NR1009 from 2013 comparing it with other inputs of the machine including its TosLink and coaxial SPDIF inputs. Unfortunately quite a jittery machine and surprisingly the coaxial had higher amplitude sidebands compared to TosLink or HDMI:
      http://archimago.blogspot.ca/2013/10/measurements-onkyo-tx-nr1009-as-hdmi.html

      Ultimately, jitter with HDMI depends on the implementation of the device (just as it is with DACs and SPDIF inputs). HDMI derives the audio clock from the HDMI video clock to keep the data in sync and thus, more jittery because you are at the mercy of the source device (computer HDMI output, Blu-Ray player, etc...). This is similar to S/PDIF and unlike asynchronous USB. I have not seen much more on HDMI presumably because it's rare for "audiophile" DACs. I'm sure to look into this more if the audiophile world ever evolves down the HDMI path.

      As for the Oppo 105. I have not measured that specifically, just the USB input. It would be interesting though if anyone has measurements! My bet would be that it'd be clearly more jittery than USB.

      Delete
    2. Just to mention that the Linn DS streamers have multiple HDMI inputs, the idea being that you route your cable receiver, sat receiver, DVD/BD player, etc., all through the Linn to get the benefit of its DAC for all sources. http://www.linn.co.uk/hifi-separates/network-music-players
      So someone cares about HDMI for audio!

      Delete
    3. Thanks Arch and thanks Otto.
      Just wondering if either of you have an opinion of whether "more jittery" eliminates an HDMI DAC input from audiophile status.
      FWIW, based on reputation and past performance Oppo would seem to maybe be a candidate for fine performance.
      I personally think MCH is the way to go. No less an authority than Floyd Toole state that MCH is a superior format.
      Thanks again,sk16

      Delete
    4. No, "more jittery" doesn't exclude from audiophile status! Just look at the Audio Note CD player and DAC measurement!

      The fact is that it's just not audible unless obscenely bad and the *concept* of jitter is used more as an advertising boogeyman than a real problem. The analogue output from my receiver for example sounds just fine fed into my pre-amp despite knowing about the higher jitter measurements.

      Yes, I agree about MCH. Done right, it sounds phenomenal. Hard part of course is "doing it right" with the sound room, placements, extra speakers and cables, calibration of levels and relative time delays, etc... I personally am looking forward to getting enough time to try making multichannel room filters with Acourate.

      Otto: Interesting, I forgot about the Linn's. Do you know if anyone has measured something like the DSM's analogue outputs with HDMI data? That would be very telling. I looked around but could not find anything in the Linn literature talking about low jitter HDMI input.

      Delete
    5. No, I'm not aware of jitter measurements for HDMI on anything. I was hoping that you might be. Even if Linn were to claim low jitter, that would be meaningless unless verified by independent test and compared to other equipment using HDMI.

      Interestingly, Naim, Linn's obvious competitor in the high end streaming/DAC market here in the UK, does not include HDMI in any of their products as far as I can see, and I don't think that Meridian does either.

      Like you, I suspect that jitter is just something that some journos use to explain subjective differences when no other explanation seems to fit. No doubt something else will come along eventually to replace jitter as the prime suspect.

      Delete
    6. Thanks again Arch and Otto. Great inputs.
      I will add that I have recently downloaded, configured and designed a house curve for Dirac Live Room Correction, MCH version. Really nice results in my opinion.
      Any plans to do an analysis of Dirac?
      Thanks again dudes!

      Delete
    7. Hi SK16,
      No plans for me with Dirac at the moment since I have Acourate as my main room correction filter creation system. Glad you're liking the MCH results!

      Delete
  9. Archimago, nice work as usual. With respect to your speculation on a "de-blur" DSP step in the mastering: since we're dealing with a reference that is 192K, the reference would already have essentially perfect time domain performance to well above the audio range, right? So the de-blur would have to apply to any time domain issues with the microphones, analog preamps, EQ, etc? That would require very detailed knowledge of the recording process to implement successfully ... Or maybe I'm missing something.

    ReplyDelete
    Replies
    1. Hi Scott. Yes, I believe you are correct.

      For a fully optimal "de-blur" assuming we accept Meridian/MQA's suggestion that this is even doable, I would imagine significant work will need to be done to measure and identify each device in the recording chain. Furthermore, in a studio project, I imagine plugins, EQs, and other digital signal processing will need to be examined to look for relative latency and phase effects. If you recall, this was one of the questions I brought up in my last MQA post for Jason Serinius to ask.

      Then somehow this information is presumably fed into the encoding system. I don't know how much of the above is accurate of course since we're not privy to the proprietary technique... Maybe Bob Stuart/MQA/Meridian can clarify the process at some point.

      Delete
    2. It would be interesting to know if there are consistent differences in phase relationships between the MQA and reference files. Are there ways in Audition to analyze this sort of thing? And could one then derive a transform to apply to the reference file, and compare to the MQA? I wish I had the equipment and expertise to answer such questions for myself ...

      Delete
    3. Hmmm... Interesting idea. I see the comment on REG about phase delaying the low frequencies to create a "clearer" sound. Boy that would be a let-down after all the discussion about how sophisticated the process is supposed to be if this is so!

      Well, I'm away from my main computer for a few days so can't get at the audio tools to see if there's anything there that can help. One suggestion that can keep things simple, perhaps the folks on REG can have a look at the 2L MQA for the Carl Nielsen piano 16/44 "DAT" sample versus the top 16-bits of the MQA file if there is a way known to detect hints of this being done. I hear the same "clarity" in that file which would come straight from 2L and without the extra complexity of 24/192 as evaluated here.

      Delete
  10. De-Blur after a already final master

    But if I mix and master a 192 kHz recording for a 192 kHz master, isn't it the case, that I use small eq in either minimum phase or linear phase (up to the task and my taste) that I am happy and satisfied with the final sound?

    So I deal with the timing and phase response of the source components anyway, to get the best out of it. So why an additional "de-blur" process on the final master, that the mastering engineer, producer, artist, ... signed already for "this is the sound we wanted".

    Juergen

    ReplyDelete
    Replies
    1. Yeah... Exactly. You would think that in the studio one would just produce the final master sounding the way it *should* sound with high quality accurate gear!

      Hopefully there will be more details forthcoming from MQA/Meridian/Bob Stuart about the whole process before they simply expect audiophiles to just "accept" the claims. It would be great if they could actually even release example 24/192 files of what the de-blur process sounds like free from the need to decode the "encapsulation" to help us understand the benefits of this process and the kind of sonic difference it makes.

      Delete
    2. I just posted the following to REG's audio forum. Maybe I'm too cynical :-)

      --------------------------------------

      In any case, it is unlikely we will learn details from Stuart/MQA directly. Given that MQA mastering almost by definition must use well known (and readily available) core technologies for measurement and DSP, the only way MQA has a chance to create a brand and monetize their approach is by aggressively protecting their process. And the mostly sycophantic audiophile press (present company excepted) surely won't force the details out of them :-)

      Delete
  11. Hi Scott

    Just one word concerning your thought, about differences in the bass response.

    When I compare the integrated magnitude of the above mentioned song, of the 192 kHz PCM file, the DSD64 file, the MQA non encoded file and the MQA 192 kHz encoded file between 5 Hz to 5 kHz, within an accuracy of + / - 0.5 dB, then only the DSD64 files does differs (in magnitude).

    And when I do compare those files with DiffMaker (with and without sample rate drift compensation) then I was not aware of any phase difference in the bass area, because this would have lead to a larger part of difference in the bass area, compared to the midrange.

    The most difference in the signal content within Diffmaker was in the area of upper midrange to lower treble. Above 10 kHz every of the above mentioned files where different (the large part in the noise of the sound floor, but also greater differences in amplitude than + / - 0.5 dB).

    So it looks like that the phase, and so the timing, of the signal parts between upper midrange and lower treble have been de-blurred most (and as I said, above 10 kHz, the differences are too large, to make any judgement based on a song. There we would need some test signals.)

    Juergen

    ReplyDelete
    Replies
    1. Thank you for your observations, Juergen. It is interesting that you note response differences "greater than +/- 0.5 dB", which is certainly audible if over a large enough frequency range, regardless of any phase differences.

      Delete
  12. Hi Demian

    Just one word about the recording interface

    You are right, that the noise floor of the RME FireFace 802 slowly begin to rise at around 30 kHz, but still is below -130 dBFS at 30 kHz and so still 10 dB lower, than of the above mentioned recording at 30 kHz.

    You are also right, that the EMU “M” interface does not show this slow rise of the noise floor at 30 kHz, but here, the noise floor begin to rise slowly in the bass area (I have a good handful of different audio interfaces).

    All interfaces do have a bit different sound, a bit of sonic signature, but in my impression, the RME FireFace 802 is on the neutral side, with a good transparency and just a tiny tendency on the "bright side" (very slightly).

    But this is only minimal, compared to the differences between MQA and PCM and DSD

    Juergen

    ReplyDelete
  13. Hi Demian

    And just one last word about the recording microphone

    You are right with the above mentioned B&K microphone (that I do also have for measuring loud speakers). This tuned mainly for free field measurements and does really drop off at off axis.

    But some good small diagram condenser microphones for recording music are tuned to a sort of diffuse field compensation, meaning that the average of between + / - 30 degree is “relatively neutral”, and the “rest” will be done during mixing.

    For good distances I am happy with the Sennheiser MKH 8000 mics and the Earthworks mics (depending on the need) and they are ok “neutral” in the above octave, when recording violins, trumpets, harpsichord, and other instruments, with high spectral content in the above octave. Even I do in now way hear higher than 17 kHz (or so), those microphones do not change the tone within certain degree and also not so much the phase response, so I am feeling on the safe side.

    And as for recording with higher sample rates I totally agree with you that the higher sample rate do mainly shift the phase and ringing behavior of the digital filters in an area, to feel save (just time wise), even in the frequency pane, I feel no need of that frequency area, just to feel “save” in the time and phase behavior. So far, from my side.

    But this only as a side note, in this MQA blog.

    Juergen

    ReplyDelete
    Replies
    1. Thanks Juergen for your insights and expertise!

      Everyone, I think it'll be interesting in the months ahead what the public reception of MQA looks like. So far, looking at message boards and such, other than specific audiophile places posting articles on it, the general music loving public seems to not be noticing it much. Possibly fatigue given the time duration since the splashy party at the Shard in late 2014.

      I suspect the rather secretive nature of the technique and lack of details doesn't help and perhaps adds to the aura that maybe it isn't everything it was sold to be (as I suspect based on the objective findings and piecing together the little we know).

      BTW: Computer Audiophile had this message thread to "Ask Bob Stuart":
      http://www.computeraudiophile.com/f8-general-forum/your-chance-ask-bob-stuart-anything-about-mqa-27412/

      It was opened Jan 28th. At about a month since then, has anyone seen any answers?

      Delete
  14. That is a bit of a bummer.. Thanks btw for explaining MQA a bit better, this was my first time looking at it in this level, just heard about that it exist. My thoughts were sceptical from beginning, the way it was marketed made it sound like it's lossless while i know that they can not fit that information. I was hoping for 20+4 scheme but alas.. Not suitable for archiving, where this was the best application..

    Would be nice to get the typical IM distortion and heat when talking about ultrasonic content; imho the best cure for hires syndrome is to ask what happens to the energy that is not outputted, what happens to 96kHz signal that is fed to system capable of doing 30k? It is just an easy, copy paste sentence after each article about hires ;)

    Agree on Fireface, transparent, no need to doubt it. Haven't measured it's ultrasonic signature, i am bandpass advocate (everything has to be bandpassed, no matter what resolution.. so i take as high as i can, bandpass internally, i only care about higher resolution in timedomain, don't care about >20k content..) ;)

    ReplyDelete
    Replies
    1. Thanks for the note Squid.

      Yeah, I was at least hoping for 18+6 scheme to ensure that the "true" lossless portion would be better than 16-bit CD resolution. Who knows, maybe it's still possible to encode like this depending on how rigid the encoding algo has to be... Considering version 1 of MQA is already out, the answer to this question is known; just a matter if MQA wants to spill the beans!

      Delete
    2. Thanks for the note Squid.

      Yeah, I was at least hoping for 18+6 scheme to ensure that the "true" lossless portion would be better than 16-bit CD resolution. Who knows, maybe it's still possible to encode like this depending on how rigid the encoding algo has to be... Considering version 1 of MQA is already out, the answer to this question is known; just a matter if MQA wants to spill the beans!

      Delete
  15. Pretty nice post. I have in fact enjoyed reading your website posts. I have been googling blogs and sites in related manner recently and i have to state you have a nice
    Shawls Manufacturer in Delhi

    ReplyDelete
  16. Linked this site from Benchmark's site. I recently purchased new headphones and an Explorer2. I returned the first one. Just got the second one back from Meridian. It tested OK. I downloaded 2L files on a Dell laptop and a ASUS flip and also my Win7 Desktop. Despite using Foobar2000 and ASIO, I have never gotten the MQA light to change. Also, tried the download play of files from Ellis website listed on MQA facebook page. So, how does one get a bit perfect file download. Or is the Realtek chip in the desktop and Dell with its software volume control ruining the file? John Atkinson uses a MAC. The laptops have the latest WIN10 and ANNIV. update. There must be a secret. And if you can get rid of the volume control, the headphones will be too loud and one must use the mini output jack to a playback system.

    ReplyDelete