davewantsmoore Posted May 11, 2016 Posted May 11, 2016 The 352/24 file is referred to as DXD. It's a straight PCM file and is the original recording format (they confirm that on the right hand side of the table). The other files are produced from that master. You'll see that the majority of the recordings are in that format. The MQA files have been prepared from the source, that is, the format listed on the far right. Thanks. Yeah... it took me a minute to work it out also. Heh. Having said that, you probably have to play the lot to get any real impression of what is going on, and know your DAC's limitations (in case the DXD is downconverted before playback, which some do). That brings up an important aspect re: the "point of MQA". Why could down conversion be a problem? (eg. from 384 to 192) The common misunderstanding is to simply say because 384 (DXD) is better than 192 ...... but all we know for certain is that 192 can't store frequencies above 96khz .... whereas 384 can...... this says nothing concrete about quality. MQA contend that it may sound different because the audio was 'damaged' when it was converted by the DAC. Not because it was converted to an inferior format, but because a damaging conversion process was used. MQAs approach is two-fold: Transparently render the audio in a format best for the target DAC (ie. don't feed the DAC 384, and let it convert it to something else) Apply filters to the audio which correct for other known problems (like damaging conversion processes, either in production, or at playback)
davewantsmoore Posted May 11, 2016 Posted May 11, 2016 Hi John, the MQA file plays back at 352/24. The MQA file is said to be the original source resolution (not file type). on the 2l site. DXD is a high resolution PCM format. Yes. DXD (352/24) encoded with MQA will result in 24/44.1 When decoded, it could again be 352/24 ..... but the MQA decoder, could also render it at a lower sampling rate. For example if the target DAC only worked internally with rates up to 176.4. Then the MQA decoder could render it as 176.4/24.
Guest Eggcup The Daft Posted May 11, 2016 Posted May 11, 2016 Yes. DXD (352/24) encoded with MQA will result in 24/44.1 When decoded, it could again be 352/24 ..... but the MQA decoder, could also render it at a lower sampling rate. For example if the target DAC only worked internally with rates up to 176.4. Then the MQA decoder could render it as 176.4/24. I'd got that far regarding the playback rate. MQA has to know the capability of the target DAC to produce the relevant datastream. While it plays back at whatever rate, it is not playing back the same as a prepared straight PCM file at whatever rate though... it is already "deblurred" before the MQA decoding happens (see the preamble at the top of the 2L test bench page, and elsewhere). It has had some high frequency and bit depth information removed to allow for the control information such as the adaptive filters. I've been assuming that any MQA approved DAC has to play at whatever speed allows the (explicit) impulse response they feel the need for their product to reproduce - is that how you read it? Clearly, it has to "target" a commonly used lower bitrate to playback on a non-MQA DAC. I vaguely remember that Stuart has talked about requiring an octave above hearing range, so there's a LOT to discard there in a 352/24 file. Hi John, the MQA file plays back at 352/24. The MQA file is said to be the original source resolution (not file type). on the 2l site. DXD is a high resolution PCM format. It isn't identical to the original though. See the above. I've made the same mistake here: the MQA file plays at 24/44.1 on a non-MQA DAC. I've pulled that figure from a couple of Windows players and an Oppo 105. @@blairy - can I ask a favour - could you download the Carl Nielsen Chaconne from there and report what the playback rate is?
davewantsmoore Posted May 11, 2016 Posted May 11, 2016 Clearly, it has to "target" a commonly used lower bitrate to playback on a non-MQA DAC. Without an MQA decoder, it plays back as whatever the PCM container is. So for the MQA released thus far, that is 24bit and 44.1khz. or 48khz. is that how you read it? Compensating for something in production is cooked into the file. Compensating for the playback path, is done when the file is "rendered". This is where the MQA decoder decides how to "expand" the initial PCM (eg. 24/44.1) to a new depth and rate. If it knows about the DAC, it could choose a depth and rate which suited the DAC, and it could y apply filters to the audio to improve things.
blairy Posted May 12, 2016 Posted May 12, 2016 @@blairy - can I ask a favour - could you download the Carl Nielsen Chaconne from there and report what the playback rate is? File name indicates it is 44k-24b On my XDP it plays back at 44.1kHz (bit rate unfortunately is not displayed).
Guest Eggcup The Daft Posted May 12, 2016 Posted May 12, 2016 File name indicates it is 44k-24b On my XDP it plays back at 44.1kHz (bit rate unfortunately is not displayed). Thanks
blairy Posted May 26, 2016 Posted May 26, 2016 @ eggcup the daft I've now got hold of a cable that enables connection of my XDP-100 to my Oppo HA1. The HA1 of course shows both resolution and bit rate. In my prior post I advised that on my XDP the Carl Nielsen Chaconne played back at 44.1 kHz and that bit rate was not displayed. I can confirm that it plays back at 44.1/24
Guest Sime Posted May 26, 2016 Posted May 26, 2016 Not a good start, Schiit have pulled the plug. http://www.digitalaudioreview.net/2016/05/schiitting-on-mqa/
powerav Posted May 26, 2016 Posted May 26, 2016 Not a good start, Schiit have pulled the plug. http://www.digitalaudioreview.net/2016/05/schiitting-on-mqa/ Yeah, just read that on my Facebook feed. Schitt's stance does make sense though. I think it is good that they have made a stance as it really is a case that the latest audio and video formats are always the 'best" until the next one comes along. 1
Guest Eggcup The Daft Posted May 27, 2016 Posted May 27, 2016 @ eggcup the daft I've now got hold of a cable that enables connection of my XDP-100 to my Oppo HA1. The HA1 of course shows both resolution and bit rate. In my prior post I advised that on my XDP the Carl Nielsen Chaconne played back at 44.1 kHz and that bit rate was not displayed. I can confirm that it plays back at 44.1/24 Thanks. MQA only really plays back at 16 bit resolution anyway - the lower 8 bits contain only the control information in this case. The fact that the Oppo shows 24 bit means that this file won't play back on old 16 bit DACs. Care to comment about the sound?
betocool Posted May 27, 2016 Posted May 27, 2016 Not a good start, Schiit have pulled the plug. http://www.digitalaudioreview.net/2016/05/schiitting-on-mqa/ It reflects my thoughts exactly. Maybe I'm not audiophile enough?
Jumbuck Posted May 27, 2016 Posted May 27, 2016 I am just waiting for everyone to sell off their high end DAC's so they can get one that one that supports MQA. Should be some real bargains around then 2
davewantsmoore Posted May 27, 2016 Posted May 27, 2016 Thanks. MQA only really plays back at 16 bit resolution anyway - the lower 8 bits contain only the control information in this case. It isn't "fixed" like this.... so you can't say "MQA plays at 16bits" Some amount of bits are used for the "regular" audio.... and the rest of the bits are noise, and available to store encoded MQA information (the high frequencies and whatever else). This decision is made by the encoder. It can be influenced by settings passed to encoder. All MQA released so far is 24bit (any bitrate could be released) ..... So. In general. If you encode audio that has a noisefloor which requires 18bits to represent. 18bits will be used for the audio .... and 6bits will be noise, available for the MQA encoded data. If you encoded audio which needed 13bits to represent .... then there would be 11bits available for MQA.
davewantsmoore Posted May 27, 2016 Posted May 27, 2016 The fact that the Oppo shows 24 bit means that this file won't play back on old 16 bit DACs. All MQA released so far is 24bit PCM .... So to play any of them, you need a DAC which can accept 24bit audio. Many 16bit DACs will accept 24bit as an input format, by truncating the least significant bits.
davewantsmoore Posted May 27, 2016 Posted May 27, 2016 Helping people get to the best generation of the original recordings is a very good thing. Packing higher frequencies into lower rate PCM is a welcome backwards compatibility feature. These two things could be considered a "revolution" (or at least a welcome evolution). ... are the filters that MQA are applying now (or will apply in the future) a tangible benefit? Are they open to abuse? Deep rabbit hole. MQA say a resounding "yes, and not on our watch". 1
gcgreg Posted May 27, 2016 Posted May 27, 2016 Not a good start, Schiit have pulled the plug. http://www.digitalaudioreview.net/2016/05/schiitting-on-mqa/ To be fair, this is really a reiteration of what Schiit have said all along. They offered DSD due to supposed high demand, under declared sufferance at the time and ultimately had their position vindicated.
malsound Posted May 27, 2016 Posted May 27, 2016 http://www.digitalaudioreview.net/2016/05/schiitting-on-mqa/
SilverPS3 Posted May 27, 2016 Posted May 27, 2016 http://www.realhd-audio.com/?p=5688 Interestingly, the technology skeptic and critic Mark Waldrep (aka. Dr Aix) couldn't get a test sample from Meridian to do a thorough analysis
blairy Posted May 29, 2016 Posted May 29, 2016 ...... Now first things first. On the 2L website the MQA file is 98MB....the stereo 352.8kHz file is 760MB. In terms of file size and resolution provided I would say this is a quantum leap forward .... certainly for downloads and streaming. The MQA file sounded very good. I'm not familiar with this music and have only listened to the first minute or two of each track. And probably pointless comparing the 352.8 file to the 176.4 file anyway. I was more curious about whether MQA worked on my XDP and if so what resolution does it provide. I wonder how many people in Australia have actually heard MQA Thanks. ..... Care to comment about the sound? Initial impressions gleaned from listening via XDP-100 thru Oppo HA2 and Audiofly IEM 180 was that MQA format seemed to be a little easier on the ear. And trust me this was probably just a feeling.
blairy Posted May 29, 2016 Posted May 29, 2016 My view thus far on MQA is that if it is to be game changing it will not have anything to do with sound quality but more to do with providing MP3 convenience (for want of a better expression) applied to HiDef music. Further testing to come :-)
davewantsmoore Posted May 29, 2016 Posted May 29, 2016 MQA format seemed to be a little easier on the ear That's the type of improvement MQA say we should expect.
blairy Posted May 29, 2016 Posted May 29, 2016 OK, I’ve been comparing the flac 352.8/24 with the MQA version of the same track downloaded from the 2L website: Mozart: Violin concerto in D major - AllegroMarianne Thorsen / TrondheimSolistene Source is Pioneer XDP-100 thru Oppo HA1 into Oppo PM1 headphones Obviously in this setup it is impossible for me to undertake blind testing…or anything close thereto J NB: Using the XDP-100, to get MQA files to play at 352.8/24 (via a USB connected DAC) I have to use the XDP-100 up sampling function and turn on DSP. Perhaps this is a quirk of the XDP or a quirk of MQA. (I note that Schitt have declared MQA a lossy format.) Using a number of points in the above mentioned track…. I compared solo violin and orchestra and everything in between …. And I can’t pick a difference between the flac track or the MQA version of the flac track. If you’re looking for the ‘wow smile’ featured on the MQA website sorry I can’t help you. I suspect that smile is people used to MP3 music J Is MQA a game changer; in terms of sound quality I don’t believe so. In terms of providing a high resolution quality file; based on the above MQA is as good as hirez flac but comes in MP3 size for streaming or download…. This is what might make MQA a game changer. They need to change their marketing tack…MQA is nothing to do with a new and better quality sound format. It’s a compression format. ​My earlier comments about MQA being easier on the ear are still relevant. This comment was based on a portable listening environment. In my desktop setup I did not experience this.
MusicEar Posted May 29, 2016 Posted May 29, 2016 Another DAC manufacturer refuses to support MQA and Why, read on... http://schiit.com/news/news/why-we-wont-be-supporting-mqa
wizardofoz Posted May 29, 2016 Posted May 29, 2016 http://www.realhd-audio.com/?p=5688 Interestingly, the technology skeptic and critic Mark Waldrep (aka. Dr Aix) couldn't get a test sample from Meridian to do a thorough analysis Mark is not a big fan on DSD either it seems http://www.3beez.com/video-introduction.html
davewantsmoore Posted May 30, 2016 Posted May 30, 2016 (edited) NB: Using the XDP-100, to get MQA files to play at 352.8/24 (via a USB connected DAC) I have to use the XDP-100 up sampling function and turn on DSP. Perhaps this is a quirk of the XDP or a quirk of MQA. (I note that Schitt have declared MQA a lossy format.) Using a number of points in the above mentioned track…. I compared solo violin and orchestra and everything in between …. And I can’t pick a difference between the flac track or the MQA version of the flac track. That seems strange (not that you couldn't hear a difference, but that you had to turn on upsampling DSP). Are you sure you're not just listening to the raw PCM 24/44.1 oversampled to 352.8khz? I would expect (not knowing specifically about the XDP-100) that when MQA is playing back, that the MQA light comes on ... and the playback format says "24bit and 44.1khz". I note that Schitt have declared MQA a lossy format >_< schitt Obviously to assess whether "lossy" is a problem or not, we need to quantify how much is lost. MQA is "lossy" below 20khz ... in so far as there are filters baked into the audio. The intention is an improvement, so is lossy "bad" in that case? MQA is "lossy" above 20khz .... in so far as not all frequencies and amplitudes are encoded. Considering the is mostly nothing above 20khz, then encoding all the empty areas of the content is very wasteful. So is lossy "bad" in this case? Edited May 30, 2016 by davewantsmoore
Recommended Posts