Blue or red pill?

[manisandher]

5 hours ago, adamdea said:

This part troubles me. Aside from my concern about switching tells...

 

There were no switching tells. My responses in the A/B/X were based purely on the differences in sound between A and B. I played A and B to Mans afterwards and talked him through exactly what I was hearing.

 

We took a number of analogue captures of A and B after the A/B/X. Once I'm back at home, I'll upload these for anyone to take a look at. (Mans will be able verify that they are indeed the same analogue captures he has and that were taken whilst he was there.) You will see that there were no drop-outs or 'tells' of any kind.

 

Mani.

[mansr]

1 minute ago, manisandher said:

There were no switching tells.

You can't know that for sure.

[manisandher]

1 minute ago, mansr said:

You can't know that for sure.

 

I know what I heard!

 

Cut this out guys. Apply your thinking to how bit-identical playback can sound different...

 

Mani.

[manisandher]

6 hours ago, pkane2001 said:

Do you have a calibration scan of a 10KHz sine wave using your ADC? Something we could compare these scans with without Mani's PC and DAC in the circuit?

 

Mans has some analogue captures of a 10k tone being played through the Phasure NOS1, taken with the same ADC.

 

Mani.

[manisandher]

1 hour ago, jabbr said:

Is there a summary thread?

 

I'll put something together at some point...

 

Mani.

[pkane2001]

38 minutes ago, manisandher said:

 

Mans has some analogue captures of a 10k tone being played through the Phasure NOS1, taken with the same ADC.

 

Mani.

 

That might be useful to compare to, to eliminate the ADC as the possible source of noise, jitter and 100Hz harmonic content.

[semente]

On 27/03/2018 at 10:01 PM, manisandher said:

I was listening to the piano transients - were they 'incisive' or 'blurred'/'soft'? With the female vocals, I was listening to her sibilance - acceptable, or too sharp and annoying.

 

Generally, it's much more about 'focus' than it is about the 'quantity' of any particular thing. For example, I didn't hear more bass, or more highs in one over the other, or anything like that.

 

What kinds of distortion could produce these effects besides jitter?

Phase distortion, IM?

[adamdea]

1 hour ago, mansr said:

You can't know that for sure.

 

57 minutes ago, manisandher said:

 

I know what I heard!

 

Cut this out guys. Apply your thinking to how bit-identical playback can sound different...

 

Mani.

No. You don’t necessarily know what you heard. And whilst it is obvious what channel of thought you would like us to go down,  it just doesn’t work that way.

[austinpop]

Sigh. I allowed myself to think this time it would be different. My mistake.

[Ralf11]

10 minutes ago, semente said:

 

What kinds of distortion could produce these effects besides jitter?

Phase distortion, IM?

 

bump for @jabbr's question re 1/f noise...

 

[mansr]

1 hour ago, manisandher said:

Mans has some analogue captures of a 10k tone being played through the Phasure NOS1, taken with the same ADC.

That's news to me.

[Ralf11]

Do we need a thread on Jitter ??

[manisandher]

48 minutes ago, mansr said:

That's news to me.

 

PM sent to you via CA on 4th April:

 

"I've taken some 24/176.4 captures from the output of my Phasure DAC. I thought you might find the following two useful in determining the performance of the Tascam ADC..."

 

Let me know if you need me to send the links again.

 

Mani.

[manisandher]

50 minutes ago, adamdea said:

No. You don’t necessarily know what you heard.

 

There was a 1% probability of achieving what I did in the A/B/X through guessing alone. Now, if I achieved what I did through hearing some 'tells', where were these 'tells' during the ABXXXXXXXXXX, where I scored 4/10?

 

Adam, it's clear you're really struggling with all this. As I said before, the red pill can be bitter. But I really have no sugar coating for you.

 

Mani.

[Confused]

58 minutes ago, Ralf11 said:

Do we need a thread on Jitter ??

A good idea I think.  Reference clocks, clock chains, clock cables, ever better imbedded clocks, all these seam to be big topics currently.  Jitter has to be at the heart of all this.

[esldude]

9 hours ago, mansr said:

Sure.

 

 

Here's another plot showing a minor difference:

Those noise floors for a 16 meg or 4 meg FFT look awfully high.  Converted to the bandwidth it looks like not even 16 bit performance and that is without looking at the spikes every 100 hz interval.  

[mansr]

2 minutes ago, esldude said:

Those noise floors for a 16 meg or 4 meg FFT look awfully high.  Converted to the bandwidth it looks like not even 16 bit performance and that is without looking at the spikes every 100 hz interval.  

It's a 16-bit DAC chip from the late 80s. Go figure.

[mansr]

33 minutes ago, manisandher said:

PM sent to you via CA on 4th April:

Sorry, I was away from home and totally forgot about that. Downloading them now.

[Superdad]

2 hours ago, manisandher said:

 

I know what I heard!

 

Cut this out guys. Apply your thinking to how bit-identical playback can sound different...

 

Mani.

 

I wish the CA search engine worked right.  Because I'm quite sure I made a post at the beginning of this thread predicting that no matter how well you passed the test the results would be rejected. 

[Ralf11]

Who has rejected the results???

 

WHO ?

[fas42]

2 hours ago, semente said:

 

What kinds of distortion could produce these effects besides jitter?

Phase distortion, IM?

 

Referring to these comments,

 

Quote

 

I was listening to the piano transients - were they 'incisive' or 'blurred'/'soft'? With the female vocals, I was listening to her sibilance - acceptable, or too sharp and annoying.

 

Generally, it's much more about 'focus' than it is about the 'quantity' of any particular thing. For example, I didn't hear more bass, or more highs in one over the other, or anything like that.

 

 

 

 

This is exactly the type of thing I listen for when optimising a system - blurred/soft transients, or sharp and annoying sibilance are classic markers of faulty playback - a clear indicator of some degree of incompetence of the playback chain.

 

What kind of distortion? Ummm, I would be hesitant to apply some nicely defined term to it - somewhere, part of the circuitry is not working as well as it should, and I'm far less interested in getting some 'evidence' that reads well in a journal article, than in simply fixing the flakiness that's causing the problem ...

 

 

 

[jabbr]

54 minutes ago, mansr said:

It's a 16-bit DAC chip from the late 80s. Go figure.

 

At this point it sounds like we have an A/B/X test that indicates that the human ear can detect a difference. 

 

The electronic correlate has has not yet been determined and at this point one could:

 

1) Do a proper unblinded measurement of both A and B using a low noise, high resolution ADC. @PeterSt suggests with a resolution better than 702khz and you seem to be suggesting better than 16 bits. 

 

2) Use an input file with a generated pure 10 kHz tone. Make it 24 bits.

 

I think both are necessary. 

 

I would also:

 

3) use a NAS connected via fiberoptic and with an Intel Pro series NIC that writes directly to cache. (DirectIO)

 

4) use an iSCSI attached local drive with (3)

 

etc.

[manisandher]

45 minutes ago, esldude said:

Those noise floors for a 16 meg or 4 meg FFT look awfully high.  Converted to the bandwidth it looks like not even 16 bit performance and that is without looking at the spikes every 100 hz interval.  

 

In which case it's interesting that my ears could (relatively easily) distinguish between the two bit-identical playback methods, and yet the 24/176.4 captures of the track from the analogue outputs of the same 'not-even-16-bit' DAC, are too "chaotic" for any meaningful analysis.

 

Mani.

[jabbr]

26 minutes ago, Superdad said:

 

I wish the CA search engine worked right.  Because I'm quite sure I made a post at the beginning of this thread predicting that no matter how well you passed the test the results would be rejected. 

 

There is no basis for rejecting anything — a scientist would repeat the trial — “Science” would like to see the study repeated by a different lab. 

 

The electromagnetic correlate to the verified observation has not been determined. Such determination require laboratory study ie measurements otherwise there will be endless flailing hand waving. 

 

The first thing I would personally look for would be a difference in the FFT peak width of an isolated tone eg 10khz or 1khz but this would require a very high resolution ADC.

[jabbr]

2 minutes ago, manisandher said:

 

In which case it's interesting that my ears could (relatively easily) distinguish between the two bit-identical playback methods, and yet the 24/176.4 captures of the track from the analogue outputs of the same 'not-even-16-bit' DAC, are too "chaotic" for any meaningful analysis.

 

Mani.

No it simply might mean that your ears are better than 16 bits... ??‍♂️