I didn’t ask for the marketing. I wanted to know how you knew the sound difference was from some “deblurring” done initially and how you were able to separate the sound changes made by origami folding and this “deblurring” to definitively tell someone that’s what they are hearing with MQA content that has been unfolded only?
1 Like
You should read up on the shambles that is royalty payments to artists from streaming.
Assuming the origami unfolding is transparent enough (they have to), then the next obvious different is the de-blurring effects applied during recording and playback that is likely to contribute to the SQ. I’ve listened to different type of digital filters in the past, all these have different tonal quality at the playback end.
When comparing to Hi-Res PCM and MQA of the same master, both sound different.
‘Stealing’ is ripping a CD or DVD that you don’t own (or you do own and then sell the original disc), including discs from Netflix.
‘Stealing’ is also using software which downloads tracks from streaming services without permission.
The origami folding could explain the change in high frequency energy. The layed back that is mentioned by some listeners. So I wouldn’t say it’s transparent.
The other observation is a slight volume change, so what was done that causes a volume change?
I’m not saying MQA is doing the following but personally I’ve applied .5 db of izotope ozone’s maximizer and experienced some of the descriptors used when describing MQA. It’s special transient preserving capabilities has an effect that everyone seems to describe with MQA, improved instrument clarity, spacial queues and a slight widening of the sound stage.
I believe, the origami folding or simply lossy compression applied during encoding must be as transparent compared to a lossless compression right? This is their ultimate goal in designing this.
Slight change in gain is due to the mastering process.
What many are hearing is the de-blurring process; different type of digital filters used. MQA uses minimal phase slow roll-off similar to Ayer implementation. This type of filter gives the most natural sound; it has no pre and 2 cycles of post ring effects.
One of the things I really like about this thread is the fact that its TITLE has as many as three different tenses in it. Simple present, present progressive and present perfect…
4 Likes
I understand the filters in the renderer. But we aren’t there yet, this discussion stops at Roon’s unfold. There is no way to know what or if anything is being done beside the MQA fold because it’s tightly coupled with the fold. Archimago brought this very point up in one of his articles.
And a slow roll off minimum phase apodizing filter is a personal preference thing and very system dependent not automatically the most natural sounding. Saying so is marketing and just Chuck and Bob’s preference. Ask Mike Mofatt or Rob Watt’s their opinion and compare results. Anyway you get my point.
1 Like
I wouldn’t have noticed if you hadn’t brought it to my attention (if-clause type 3).
3 Likes
Not exactly, take at the playback end for example, Roon can do the first unfold then output to the non MQA DAC instead of a renderer. If one put the renderer in, such as the Dragonfly then the additional process of selecting the appropriate up-sampling de-blurring filters (16 of them) can be done. This complete the whole ‘de-blurring’ process from the recording right down to the playback end. This also shows that unfolding and de-blurring may not be necessarily tightly coupled. If they do, it will be virtually impossible for anyone to make this discovery; 16 type of de-blurring filters found in Dragonfly and Mytek Brooklyn DAC!
You quoted one sentence and then go on a tangent including the renderer. Roon can only unfold, I’m discussing what you get from MQA at that point with a non MQA DAC.
1 Like
Started out as simple present about 1 week before V1.5. Just trying to stay up to date.
1 Like
I wanted to do a simple test to compare MQA rendering with ordinary Roon upsampling after Roon unfolding. As you will see, I found the content limitations make rendering largely irrelevant, and the results were undramatic.
Basically, I compared 1) Roon passthrough and Meridian 818v3 unfolding+rendering, with 2) Roon unfolding and upsampling, and 3) Roon unfolding with no upsampling. I can’t test Roon unfolding plus MQA rendering because the 818v3 can’t do that, your DAC may be able to.
A key problem: I wanted to use an MQA 192k track to get some rendering, a 96k track wouldn’t do rendering, and a 352k track would do more rendering than I can do with Roon upsampling because the 818v3 can’t accept more than 192 input. Turns out, this was difficult. Of my 189 MQA albums, almost all were 88k or 96k, a few 44k or 48k, and just a couple 352k (2L). Only a few were 192k, and they were all from the 1970s: Joni Mitchell, Herbie Hancock… This puts a twist on the test. No modern albums I could find at MQA 192k. Maybe this is special for my library. But it essentially makes the whole rendering filter discussion moot, because there is nothing to render.
On to the test.
I turned off all DSP other than upsampling.
My system has some special characteristics, you may need some tweaking.
I used USB input on the 818v3, which supports 192 (the network input is limited to 96). Because Roon knows about the 192 limit I didn’t have to limit Roon’s upsampling, if your DAC supports higher input you may have to set up custom upsampling for case 2).
I used Headphone mode which uses analog outputs on the 818v3, into a Bryston headphone amp; using the DSP speakers would invalidate the test because they upsample to 768. The 818v3 would upsample if fed 44k, but I don’t believe it upsamples a 96k or 192k feed before the DAC.
Settings: 1) DAC set to MQA unfolder and renderer. 2) DAC set to no MQA, Roon set to upsampling, tried various filters. 3) Roon set to no upsampling.
Results: Nothing. Nada. Zilch. All sounded great, no differences. At least discernible to my ears. At least with this music.
You try it.
Let us know if you have any modern 192 albums, the 70s had some great music but recording t3chnology of 50 years ago…
2 Likes
I have since rediscovered Brothers in Arms, you will all be glad to hear. Some idiot (me) had managed to hide it. Back in place, along with Sia’s This Is Acting and The Cure’s Greatest hits that I managed to misplace a few months ago.
My first comment in this and other threads:
I would like to refer you this articke : http://iar-80.com/page170.html
It is verbose in the extreme and from time to time very repetitive, BUT I have referred it to a well know mathematician who is a musician and reviewer, since I have not asked his permission to use this as a quote I will not mention his name here, but the long and short is that the mathematics referred to here are correct.
I t would really seem that a lot of digital development over the last many years is fundamentally flawed. I do not lay claim to golden ears myself, but this is definitely worrying. Please do not let the style of presentation makw you miss the actual message which seemsa to absolutely correct mathematically speaking.
We have heard quite q few voices raising the idea that the Nyquist–Shannon theorem is insufficient but it looks like this is NOT the case.
I look forward to your comments.
Robert
2 Likes
@Robert_Jorgensen that’s quite the read. Repetitive at times and a bit difficult to follow because of it’s layout in the browser but interesting never the less.
Does Roon have a “brick wall” filter? I read through that and I wish I had more knowledge but I don’t understand the science or lingo used for the most part. What is the optimal Roon settings for PCM with regard to that article?
Roon will leave the PCM alone unless you enable DSPing, volume leveling or cross fade.
Absent sample rate conversion, DSP operations mentioned above will not apply digital filtering.
As for a brick wall filter, I do not have Roon in front of me, but a fast/steep roll off, linear phase filter would fit the bill.
AJ
1 Like
Brick wall filters are about analog-to-digital or digital-to-analog conversion, specifically for 44k.
Roon doesn’t do that.