Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Meridian Audio's new... sub-format called MQA. (Read 145453 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Meridian Audio's new... sub-format called MQA.

http://musicischanging.com/

I see more descriptions on 3rd party sites.

http://www.whathifi.com/news/meridian-audi...h-res-streaming

http://www.stuff.tv/meridian/meridian-s-mq...lity-music/news


From the description, it looks like a hybrid approach - a normal CD-quality stream with Meridian's algorithm/information called MQA. If the receiver cannot decode the algorithm, it will be just played as a CD-quality stream. Otherwise, it will be restored back to high-res music.

I am not even sure if this whole process is lossless or lossy. But I really do not like to have a new proprietary format.

Meridian Audio's new... sub-format called MQA.

Reply #1
All I'm getting from this is yet another hymn to hi-res music, to be honest. The usual "it sounds so much better than regular CD!".
The only "advantage" is that it supposedly takes as much space as a regular audio CD, so you get "more" for "free". Kind of like HDCD?

Meridian Audio's new... sub-format called MQA.

Reply #2
Just a guess: redbook lossywav with a lossily compressed difference file (from the hi-res) stored in the freed-up space.

Meridian Audio's new... sub-format called MQA.

Reply #3
Just a guess: redbook lossywav with a lossily compressed difference file (from the hi-res) stored in the freed-up space.
The correction file will be huge, because lossyWAV will make it plainly apparent that actual dynamic range of most music material is way below 16 bits.

reading http://www.realhd-audio.com/?p=3851, this was quoted from an actual paper by Stuart!:
Quote
When recording, the ideal is to capture a performance so that the highest possible sound quality can be recovered from the archive. While an archive has no hard limit on the quantity of data assignable to that information, in distribution the data deliverable depends on application-specific factors such as storage, bandwidth or legacy compatibility. Recent interest in high-resolution digital audio has been accompanied by a trend to higher and higher sampling rates and bit depths, yet the sound quality improvements show diminishing returns and so fail to reconcile human auditory capability with the information capacity of the channel. By bringing together advances in sampling theory with recent findings in human auditory science, our approach aims to deliver extremely high sound quality through a hierarchical distribution chain where sample rate and bit depth can vary at each link but where the overall system is managed from end-to-end, including the converters. Our aim is an improved time/frequency balance in a high-performance chain whose errors, from the perspective of the human listener, are equivalent to no more than those introduced by sound traveling a short distance through air.
It's only audiophile if it's inconvenient.

Meridian Audio's new... sub-format called MQA.

Reply #4
According to this Meridian graph I should stick to my Studer tape recorder, if I can live with its low convenience level.


Meridian Audio's new... sub-format called MQA.

Reply #5
Maybe they are going to build a network open reel machine which can download analog audio from internet and record to tapes.

Meridian Audio's new... sub-format called MQA.

Reply #6
According to this Meridian graph I should stick to my Studer tape recorder, if I can live with its low convenience level.


Wow - interesting graph. From it, I see that:
  • LP is higher quality and more convenient than DVD-A
  • DVD-A is extremely inconvenient. (How so? Is it because not everyone has a DVD-A capable player?)

I always had a fair amount of respect for Meridian - they've done some genuinely useful things in the domestic audio playback arena. (In particular, promoting the idea of digital active speakers).

But honestly - what have they been smoking?

Meridian Audio's new... sub-format called MQA.

Reply #7
And reel-to-reel is the best quality.
I think my parents still have their reel-to-reel machine somewhere. The tapes looked like this:


Meridian Audio's new... sub-format called MQA.

Reply #8
This may be based on the ideas revealed in a recent AES paper...
http://www.hydrogenaud.io/forums/index.php...mp;#entry882550


When I saw that quality/convenience graph I nearly wept. I assume it means studio quality reel-to-reel machines running at 15 or 30ips with professional tape stock, not the 7.5ips rubbish sold to consumers in the 1960s, but even so...

Cheers,
David.

Meridian Audio's new... sub-format called MQA.

Reply #9
And reel-to-reel is the best quality.
I think my parents still have their reel-to-reel machine somewhere. The tapes looked like this:


Analog tape is an audibly flawed medium. Even one generation on very high quality analog tape gear is audible in ABX testing.



Meridian Audio's new... sub-format called MQA.

Reply #11
Not had time to read it (and the ref. to a newer patent) properly yet, but this may be relevant.

Meridian Audio's new... sub-format called MQA.

Reply #12
What's this?
http://patentscope.wipo.int/search/en/deta...p;maxRec=599628
The file format looks like a fixed-bitdepth lossyWAV, with correction data stored in the LSBs of the signal. The approach seems to be similar to HDCD. Incompatible hardware will be audibly OK with most data since the correction info is in the LSBs beyond the upper 16 bits. Fidelity will be slightly worse than true 16 bit files because of the 3 bit lossy part in the stream. Compatible hardware will be able to restore the original file. Have I deduced that correctly from the graph in the patent?

Anyway, I wonder what the point is. If you deliver 24 bit files anyway, why do this compression in the first place. The argument that you can truncate the LSBs to get lower bitdepth files in low bandwidth cases is still possible with plain 24bit PCM.

My suspicion is the following: The filesize and bitrate will still be huge enough, so that audiophiles can be content about huge numbers on their displays. Audio quality will be worse than true 16 bit delivery, and maybe just bad enough to be noticeable on selected material, so that there is an incentive for people to buy the expensive meridian decoder hardware. This format only exists to intentionally deliver low fidelity music to consumers without Meridian hardware, to sell them expensive Meridian decoders.
It's only audiophile if it's inconvenient.

Meridian Audio's new... sub-format called MQA.

Reply #13
Anyway, I wonder what the point is. If you deliver 24 bit files anyway, why do this compression in the first place. The argument that you can truncate the LSBs to get lower bitdepth files in low bandwidth cases is still possible with plain 24bit PCM.


.... because it is a new bespoke format that *must be better* than those that have gone before, however if simple bit-depth reduction on playback were to be used, uers would try to notice the difference between a 24-bit PCM stream and a (rounded or truncated and dithered) 16-bit PCM stream and would presumably be unable to at normal listening levels?

Does the new format allow any form of DRM to be included?
lossyWAV -q X -a 4 -s h -A --feedback 2 --limit 15848 --scale 0.5 | FLAC -5 -e -p -b 512 -P=4096 -S- (having set foobar to output 24-bit PCM; scaling by 0.5 gives the ANS headroom to work)

Meridian Audio's new... sub-format called MQA.

Reply #14
Not had time to read it (and the ref. to a newer patent) properly yet, but this may be relevant.
I'm sure it is.

For those who haven't bough the other Meridian paper from the recent AES conference, the patent related to it might be helpful...
http://worldwide.espacenet.com/publication...677A1&KC=A1

I don't know whether MQA is based on the patent co-authored with Peter Craven alone, the one co-authored with Malcolm Law and Peter Craven, or both.

There's also
http://www.theabsolutesound.com/articles/r...o-meridian-mqa/

Cheers,
David.

Meridian Audio's new... sub-format called MQA.

Reply #15
This may be based on the ideas revealed in a recent AES paper...
http://www.hydrogenaud.io/forums/index.php...mp;#entry882550


When I saw that quality/convenience graph I nearly wept. I assume it means studio quality reel-to-reel machines running at 15 or 30ips with professional tape stock, not the 7.5ips rubbish sold to consumers in the 1960s, but even so...

Cheers,
David.
Do you mean this paper.
http://www.aes.org/e-lib/browse.cfm?elib=17501

I stumped uo the $20 and downloaded it. I have to confess that I found it very confusing.  I have only an enthusiasitc amateur understanding of sampling based on the standard  undergraduate texts, but I have to say I found bits of it very difficult to follow, especailly the stuff about time blur. Also the explanantion of the limitation of time resolution using inter sample diracs seemed very odd. Obviously Staurt and Craven are clever and expert, and the paper was peer reviewed. But somehow talking about trasnients which start and finish between samples (as opposed to locatign the peak of an event) seemed a bit dubious to me. Do these exist in music, or in the world?(assuming 44khz sample rate) Isn't an analysis of sampling a dirac "against the law" as it is of course not band-limited?

"Can a sampled system convey time differences that are
shorter than the periods between successive samples? An
intuitive answer might be ‘no’ [60], but we note that even
when convolved with a sinc function, an arbitrarily small
displacement of an impulse can be detecof waveform comparison, assuming one has sufficient
signal-to-noise ratio.
Instantaneous sampling without any filtering is not
recommended, for the sampling would then be vulnerable
to high-frequency noise (even to the megahertz region).

Further, a Dirac impulse would not be registered at all if
it happened to occur between the sampling instants.
Intuitively one would at least integrate over one sample
period, as illustrated in Figure 11 (upper). Here a
transient falling entirely within the sample period
corresponding to Sample 0 will be integrated and the
value of Sample 0 will represent the area of the transient

If the transient moves to the right, there will be no change
in the sample values until the transient crosses into the
adjacent territory of Sample 1. Positional information
has been lost, indeed quantized, so the above ‘intuitive’
answer was correct for this case.

The information loss can be avoided by using an
integration kernel in the form of a triangle or dual ramp
that spans two sample periods, as shown in Figure 11
(lower). By comparing the values of Sample 0 and
Sample 1, both the area and the position of the transient
can now be unambiguously determined."

Can someone explain to me where they are going with this?

 

Meridian Audio's new... sub-format called MQA.

Reply #17
Can someone explain to me where they are going with this?
It seems to me he's saying that there exist acoustical signals (dirac) that can be missed during sampling. Haven't we heard that before from Dr. Kunchur ?
Quote
Unless a different interpretation of minimal temporal separation is taken, it is completely fallacious to assert that a CD can resolve less than 5 microseconds when its individual samples are separated by periods of 23 microseconds. (Note that it is true that small alterations in temporal profiles can be indirectly encoded through variations in adjacent levels and that this is certainly aided by having more bits; however, a true translation in time of a temporal feature can only take place in quantized sample periods.)
[/size]It would be helpful if someone could demonstrate with a test how to generate such a signal and how it completely escapes the AD converter.

Meridian Audio's new... sub-format called MQA.

Reply #18
Perhaps Amir, who is an industry insider and from his posts has worked in this field, will have an opportunity to read the patents and other announcements, and offer his synopsis of what this Meridian technology does.

Meridian Audio's new... sub-format called MQA.

Reply #19
Do you mean this paper.
http://www.aes.org/e-lib/browse.cfm?elib=17501
Yes.

AFAICT they're proposing what most people would call non-ideal downconversion. Compared to what's normally done, they are radically shortening the anti-alias / anti-image filters, which means these filters now allow some aliasing and introduce some time-invariance to the sampling. The benefit they claim is that it allows them to shorten the impulse response dramatically. Because of the first two effects, you "have" to use a higher sampling rate than CD, otherwise you would get some nasty aliasing in the audible band. They claim to ensure it's below the noise floor.

Cheers,
David.

Meridian Audio's new... sub-format called MQA.

Reply #20
AFAICT they're proposing what most people would call non-ideal downconversion. Compared to what's normally done, they are radically shortening the anti-alias / anti-image filters, which means these filters now allow some aliasing and introduce some time-invariance to the sampling. The benefit they claim is that it allows them to shorten the impulse response dramatically. Because of the first two effects, you "have" to use a higher sampling rate than CD, otherwise you would get some nasty aliasing in the audible band. They claim to ensure it's below the noise floor.

So it seems it all comes down again to eliminate ringing. I wonder if this ringing only is used because this impulse response picture meanwhile is that strong manifested it must be the holy grail.
Is troll-adiposity coming from feederism?
With 24bit music you can listen to silence much louder!

Meridian Audio's new... sub-format called MQA.

Reply #21
I have to confess that I found it very confusing.

From reading your quote of the paper I'm also confused.

If you properly lowpass a signal which contains an "inter sample dirac", wouldn't this dirac ripple out into the neighboring samples in the form of a sinc? No need for any other magic. Correct?

Meridian Audio's new... sub-format called MQA.

Reply #22
I have to confess that I found it very confusing.

From reading your quote of the paper I'm also confused.

If you properly lowpass a signal which contains an "inter sample dirac", wouldn't this dirac ripple out into the neighboring samples in the form of a sinc? No need for any other magic. Correct?

That is my understanding- at least that it would spread out -after all a band limited signal cannot be time limited- but the shape of the remaining signal would I suppose depend on the filter. Then again I am not an expert.

Meridian Audio's new... sub-format called MQA.

Reply #23
Do you mean this paper.
http://www.aes.org/e-lib/browse.cfm?elib=17501
Yes.

AFAICT they're proposing what most people would call non-ideal downconversion. Compared to what's normally done, they are radically shortening the anti-alias / anti-image filters, which means these filters now allow some aliasing and introduce some time-invariance to the sampling. The benefit they claim is that it allows them to shorten the impulse response dramatically. Because of the first two effects, you "have" to use a higher sampling rate than CD, otherwise you would get some nasty aliasing in the audible band. They claim to ensure it's below the noise floor.

Cheers,
David.

There seem to be a number of different points floating around one seems to be a sort of perceptual coding which changes the amount of information for the higher frequencies allowing some like 24/96 or possibly 24/192 at a data rate of only about 1 mbps. The other seems to be the filtering. But there also seems to be some idea about the "sampling kernel". I have to confess that I don't really understand this part which seems to include the passage I quoted. 

In my ignorance I srt of thought that if you were sampling by avaeraging across samples you would be reducing the time resolution and I've never quite understood what was supposed to be so good about shortening the impulse response (except for reducing latency). Is the triangular kernel thing orthodox?

As I understand it the system if I can call it that is intended to be applied end to end from A/D to storage to D/A

Meridian Audio's new... sub-format called MQA.

Reply #24
That makes no sense.

Make the A/D filter steep, non-aliasing. What matters for playback is the D/A or resampling filter used, which can allow imaging, be a lot less steep, even be minimum phase ...
"I hear it when I see it."