As I have posted elsewhere, some of the comments are better reading (and more accurate) than the editorial itself
To be honest the best comments I have read are by Ken Newton on the non subjective aspects of MQA processing.
If we are to summarise the issues
MQA has joined with the major record labels to convert files to MQA
Many streaming sites have begun streaming MQA
MQA Involves reprocessing and filtering both in software, and hardware domain which attempts to preserve time domain information. It gives this precedence over antialiasing and ultimate frequency extension, and enfolds the high frequency data into the LSB of the 24 bits. This is believed to be subjectively below the noise floor, but unproven at this stage.
This system involves some elements of compression. To package it down to lower freq download.
Although this in theory may involve some technical loss of information especially in high frequency dowmain, and precision, and ultimate noise floor, in absolute terms, this as stated above, is sacrificed in return for great time domain accuracy.
The reasons for this engineering decision making appears somewhat unclear, but appears to relate to four competing engineering concerns
1. To be able to easily stream
2. A standard for downloads sites, which eliminates the need for multiple formats
3. A closed system
4. Priority of sound quality as perceived, over ultimate technical specifications in digital domain.
This system, is closed, in that it requires licensing and hardware/ firmware solutions for ultimate performance
A software only un-enfolding appears to give a more limited alleged benefit, over full un-enfolding which currently requires some hardware/firmware modifications.
There are many reports at demonstrations and home listening of improved quality.
There are also reports of more patchy benefit
I am unaware of any direct comparison of hires formats against MQA and would be enlightened of any reports.