Sonore MicroRendu Hardware Teardown and Review

[RayDunzl]

I'm holding a seminar here next week.

$3500 per seat, includes snacks and party favors.

 

[Sal1950]

I'm holding a seminar here next week.

$3500 per seat, includes snacks and party favors.

I get a 25 yo smokin hot hooker with that?

 

[RayDunzl]

Sure, bring her along...

 

[Sal1950]

Sure, bring her along...

Cheapskate.

 

[fas42]

I hear no difference 8/9.

Time for fas42 to try an ABX, I think.

This is exactly where a typical ABX will fail - people keep forgetting that the brain is always trying to reconcile differences in sounds which have meaning, and will do an excellent job of correlating two versions with a slight variation. I've done this sort of things a number of times - and have been aware of how the differences go away the more I listen. The test is guaranteed to fail, because it's a dumb test - it doesn't take into account the "cleverness" of the hearing system in being able to fill the gaps ...

By all means do an ABX to confirm your prejudices ... and thereby learn, nothing ...

 

[Thomas savage]

This is exactly where a typical ABX will fail - people keep forgetting that the brain is always trying to reconcile differences in sounds which have meaning, and will do an excellent job of correlating two versions with a slight variation. I've done this sort of things a number of times - and have been aware of how the differences go away the more I listen. The test is guaranteed to fail, because it's a dumb test - it doesn't take into account the "cleverness" of the hearing system in being able to fill the gaps ...

By all means do an ABX to confirm your prejudices ... and thereby learn, nothing ...

 

[fas42]

I'm happy to be confused, that's my goal in fact - hands up those whose systems will confuse me into thinking that their playback of a pipe organ recital recording is in fact the real thing ...

 

[amirm]

I'm happy to be confused, that's my goal in fact - hands up those whose systems will confuse me into thinking that their playback of a pipe organ recital recording is in fact the real thing ...

Frank, do you expect two bottles of water out of a six pack at the store to taste markedly different?

 

[fas42]

Frank, do you expect two bottles of water out of a six pack at the store to taste markedly different?

No ... and relevance?

 

[fas42]

As an exercise, one I've done a number of times over the years, I did a bit of playing in Audacity to synchronise the waveforms of the original, and one of the output versions. This is essentially doing the work of DiffMaker manually, resampling to very high rates and then very, very slightly varying the rate of one so that they correspond, to the sample, over significant period. Fine tune the amplitude of one, subtract and you get some waveform which looks nothing like the original, in particular, lots of high frequency mess ...

Now, the big question is, what is that "distortion" doing to the subjective listening?

 

[RayDunzl]

Now, the big question is, what is that "distortion" doing to the subjective listening?

Nothing.

You manufactured it. It is not resident in either file.

 

[manisandher]

Ray, thanks for all the hard work.

The channels are reversed in 8/9 compared to original, files 2 and 3 are not reversed.

This simply is not the case! The channels are correct in all the files. I've checked all my ICs and they are all correct. But it's easy to verify by listening to the first 20 seconds or so - the piano starts playing, there is a cymbal slightly to the left, then one to the right, then another to the left again. All the files do exactly the same.

Did you reverse something at your end perhaps?

Mani.

 

[manisandher]

Glitch in both 8 and 9 at 0.045 seconds from the beginning of the data, not in 2 and 3.

Ah, this will be because I started to use the Tascam's 'auto start' recording feature. It was set to -54dB. Sorry if that added some unnecessary complexity...

Mani.

 

[manisandher]

OK, listened to versions 8 and 9. Again, no contest - 8 has a piece of cloth draped over the mic, that's the subjective impression. All the attack elements in the sound have been dulled, it's boring to listen to.

Wow, talk about putting your neck on the line Frank - I haven't said which is which. But it's really interesting that you're hearing a difference at all because the only thing that has changed between files #8 and #9 is the digital transport.

Of course, I'll reveal which is which at some point...

Mani.

 

[RayDunzl]

Did you reverse something at your end perhaps?

Merely downloaded and dragged the files into audacity.

Looking again, I can hardly believe it - reference post #48:

Before the glitch, the channels are reversed (which is where it caught my eye), after the glitch, they aren't. Follow the squiggles straight across in #1 vs #8 or #9. They change tracks L/R at the glitch.

Ah, this will be because I started to use the Tascam's 'auto start' recording feature. It was set to -54dB. Sorry if that added some unnecessary complexity...

I don't mind looking at stuff I choose to look at... You never know what you'll see...

 

[RayDunzl]

Here it is:

 

[manisandher]

Could it be the Tascam's auto start being a bit 'broken'?

 

[manisandher]

Ray, how difficult would it be to get a difference file between #8/#1 and #9/#1? It would be really useful to me to know how accurate my DAC/ADC chain is.

Is this something I could do in Audacity, or is there a better piece of software (I've heard you guys mention 'DiffMaker' a few times)?

Mani.

 

[RayDunzl]

I'd send a bug report...

 

[RayDunzl]

Ray, how difficult would it be to get a difference file between #8/#1 and #9/#1? It would be really useful to me to know how accurate my DAC/ADC chain is.

I'm using Audacity, don't know Diffmaker. Diffmaker may be better. Will look.

Basic steps:

Drag both files into a new Audacity window.

Zoom in on a prominent feature, and time-align the tracks. The tool that looks like ][ is for selection, the tool that looks like <-> is for moving selections in time.

Invert one track.

Select both tracks, Mix and Render to a New Track

Comparing digital to analog (redigitized) it is unlikely to produce a true null ( null = no difference) due to amplitude differences, a few milliseconds of overall time difference, etc. Being off by even one sample time makes a mess. You have to fool around with it and see what happens.

Analog doesn't have "sample times" so you are unlikely to be re-sampling the analog at the same nanosecond as the sample that was decoded. Even with "no difference" you will see differrence. Ready to give up yet?