Blue or red pill?

[sandyk]

30 minutes ago, kumakuma said:

 

As mansr said, files are always played from system memory.

 

What you're referring to is whether or not the program loads the entire track or just part of the track into system memory before starting playback.

 

As fas42 pointed out , this results in a great deal more unnecessary processor etc. activity. and most members appear to prefer the sound without all this extra RF/EMI and HDD/SSD activity. The lower the noise generated at the time of playback results in improved performance in the DAC .

You may not be able to hear the difference, but I sure can.

Next you will be trying to tell me that playing music with Power DVD etc. , or even from a sound editing programs sounds just as good as a file that has been played solely from System Memory using a dedicated Audio software player ?

Dream on ! 

 

 P.S.

 Please also note the comments made attributed to jRiver in a recent post.

[kumakuma]

15 minutes ago, sandyk said:

 

As fas42 pointed out , this results in a great deal more unnecessary processor etc. activity. and most members appear to prefer the sound without all this extra RF/EMI and HDD/SSD activity. The lower the noise generated at the time of playback results in improved performance in the DAC .

You may not be able to hear the difference, but I sure can.

Next you will be trying to tell me that playing music with Power DVD etc. , or even from a sound editing programs sounds just as good as a file that has been played solely from System Memory using a dedicated Audio software player ?

Dream on ! 

 

Not sure how you got that from what I wrote

[mansr]

17 minutes ago, sandyk said:

Next you will be trying to tell me that playing music with Power DVD etc. , or even from a sound editing programs sounds just as good as a file that has been played solely from System Memory using a dedicated Audio software player ?

I don't know what PowerDVD does, but editing programs generally keep as much as possible in memory.

[sandyk]

40 minutes ago, kumakuma said:

 

Not sure how you got that from what I wrote

 Based on previous discussions with you !

[sandyk]

36 minutes ago, mansr said:

I don't know what PowerDVD does, but editing programs generally keep as much as possible in memory.

 The audio from Media players such as Power DVD, VLC Media player and GOM, while perhaps acceptable due to their video capabilities, sound quite lacklustre in comparison with the same audio (after extraction by demuxing) from a dedicated Audio SW player.

 Neither does Sound Forge F9 or Audacity sound as good either.

[kumakuma]

5 minutes ago, sandyk said:

 Based on previous discussions with you !

 

Don't think so. You appear to be confusing me with someone else.

[manisandher]

I've just had the pleasure of hosting @mansr for the day. Let's just say he's definitely on my 'really good guy' list...

 

The day didn't go as I had expected, but I hope some interesting insights will be gleaned in any case.

 

More to come in the next posts...

 

Mani.

[pkane2001]

4 minutes ago, manisandher said:

I've just had the pleasure of hosting @mansr for the day. Let's just say he's definitely on my 'really good guy' list...

 

The day didn't go as I had expected, but I hope some interesting insights will be gleaned in any case.

 

More to come in the next posts...

 

Mani.

 

Can't wait!

[manisandher]

Here's the procedure that Mans and I agreed to follow for the day:

 

1. take a quick listen together to the system and have a general chat

 

2. conduct the A/B/X

- we would play the same file back in two different bit-identical ways (by changing a parameter in XXHighEnd, which keeps everything bit-identical, but which seems to change the sound nevertheless... to my ears at least)

- Mans would be sitting in my office (with door tightly closed), controlling playback from there, and I'd be sitting in the listening room across the hall (with door tightly closed)

- we'd have a Tascam recorder set to auto-record, sitting in the basement next to the audio PC and DAC, capturing the digital output of the audio PC in real time

- we'd have my phone sitting on the desk in the office to video the PC screen as Mans conducted the test

3. ensure that the digital captures are identical

- Mans would take the digital captures away with him to ensure that they are indeed bit-identical

 

4. capture analogue outputs from the DAC, using the Tascam (listening track plus test tones)

 

5. analyse analogue outputs

- Mans would take the analogue captures away with him and attempt to identify any differences

 

We agreed that there was no point in going on to step 3 unless Mans was convinced there really were audible differences (either because he heard them too, or because I managed to demonstrate that this was the case in the A/B/X).

 

 

More to come...

 

Mani.

[Sonicularity]

Thank you both for putting this together.  Should be interesting and entertaining to read about the results.

[manisandher]

Earlier in this thread, I suggested the A/B/X should be conducted as follows:

 

On 3/1/2018 at 3:40 PM, manisandher said:

What I have in mind is the following:

 

1. set software to playback configuration A

2. play file in config A for 30 secs or so

3. switch to bit-identical playback configuration B (will take 5-10 seconds at least)

4. play file in config B for 30 secs or so

5. randomly switch between playback configs A and B, playing back file for 30 secs or so each time (ensuring that there's no way I can tell which playback config is being used)

6. edit: repeat 5. for 10 or so times (I don't think I'll need to remind myself of A and B again)

 

On each random playback, I'll simply jot down which of A or B I think I've just listened to.

 

So that's how we did things (although we used the denotation '1' and '2' instead of 'A' and 'B').

 

[For those of you familiar with XXHighEnd: '1' was with SFS=0.1 and '2' with SFS=200, all other parameters kept absolutely constant. Playback was 'Attended', in 'MinOS' mode.]

 

The track I had chosen was the 24/176.4 version of 'Tchaikovsky Hopak from Mazeppa' on Reference Recordings' 'Exotic Dances'. I asked Mans to play just the first 15 seconds in each case. And my results? 4/10!

 

I was disappointed, of course. But something felt really wrong. When Mans played '1' and then '2', the differences though certainly subtle, were clear to me nevertheless. And they sounded exactly as I expected them to sound - '1' more focused and more strident than '2'. But when Mans then played them randomly 10 times, I felt totally lost. When listening to a sample, I had no reference against which to compare, apart from the previous sample.

 

I asked Mans if we could repeat the test - maybe I simply wasn't relaxed enough. So we repeated the test. And the result the second time around? 4/10 again!

 

This was tough. I knew I was hearing a difference between '1' and '2', but simply wasn't able to identify them when played back randomly 10 times.

 

It was then that I remembered a couple of earlier posts in this thread:

 

On 2/10/2018 at 2:03 PM, mansr said:

1. Play sample A.
2. Play sample B.
3. Play A or B randomly, you decide which it is.

 

On 3/1/2018 at 6:42 PM, SoundAndMotion said:

- Try to ignore past trials and focus on the current trial. I assume an objective randomization will be used (e.g. computer), since people are notoriously bad at trying to create a random sequence on their own. That means the "correct" answer could be A (or B) 3 or 4 or more times in a row. Don't assume "I answered B the last 3 times; the next one is more likely A". Treat each trial as separate from previous ones.

 

(I'd particularly like to thank @SoundAndMotion for this advice.)

 

So I asked Mans if he'd be kind enough to conduct one more test (I promised him it'd be the last and final test before I gave up). He agreed. I decided to use another track for this final test - 'Persephone' from Patricia Barber's 'Mythologies' album (16/44.1 upsampled to 176.4 in XXHighEnd).

 

This time we did things the way Mans had suggested in the first place.

 

More to come...

 

Mani.

[mansr]

6 minutes ago, manisandher said:

I decided to use another track for this final test - 'Persephone' from Patricia Barber's 'Mythologies' album (16/44.1 upsampled to 176.4 in XXHighEnd).

Didn't we use that track for the second run as well?

[manisandher]

During the 3rd test, I was noting my results in a notebook. Meanwhile, Mans was using a random generator on his phone, and also noting the order of playback on his phone. Once we had done 10 A/B/Xs, we compared the two lists. I ticked all those correct. Here's the page from my notebook:

 

 

Yep, that's 9/10!

 

The 6th result is interesting. I had initially written '2?'. But when I listened to the 7th sample, it was clear that it must have been '1', so I crossed out the '?' and wrote '1' over the '2'.

 

More to come...

 

Mani.

[manisandher]

8 minutes ago, mansr said:

Didn't we use that track for the second run as well?

 

Yes, I think you're correct.

 

Mani.

[mansr]

He's not making this up. While 9/10 in one trial wouldn't impress a statistician, it's enough to make me intrigued. We recorded a bunch files, both music and test signals, and I'll be analysing those. There has to be an explanation.

[manisandher]

4 minutes ago, mansr said:

There has to be an explanation.

 

I'm really looking forward to seeing what you come up with.

 

Mani.

[esldude]

Not trying to be a spoil sport, but I will point out Mani's collective 17 of 30 correct doesn't meet the 95% level of confidence.  

 

OTOH, if you aren't one who has done some of these you might not have been 'ready' or comfortable or might have been trying to hard.  Others always ask for more of course.  Would be interesting to know if you had done it twice more if your results would have continued to be 9 of 10 or 10 of 10.  It might indicate a learning curve in handling the procedure on Mani's part. And yes I know completing two more of these would be asking quite a lot. So not complaining just wondering out loud. 

 

In any case, that someone will step up and do such tests is good and to be commended.  Much more interesting than what happens so often in audio forums.  So kudos to Mani and mansr.  

[manisandher]

9 minutes ago, esldude said:

Not trying to be a spoil sport, but I will point out Mani's collective 17 of 30 correct doesn't meet the 95% level of confidence.  

 

It's not 17/30, it's 9/10. The first two test were done 'incorrectly'... and the result showed.

 

Mani.

[esldude]

5 minutes ago, manisandher said:

 

It's not 17/30, it's 9/10. The first two test were done 'incorrectly'... and the result showed.

 

Mani.

So what was incorrect about it?

 

Also, did mansr think he heard a difference?

[manisandher]

29 minutes ago, mansr said:

While 9/10 in one trial wouldn't impress a statistician, it's enough to make me intrigued.

 

And that was the objective.

 

Mani.

[Superdad]

5 minutes ago, manisandher said:

It's not 17/30, it's 9/10. The first two test were done 'incorrectly'.

 

You could have gotten 10/10 or 100/100 and still there will be a bunch who will call the result statistically invalid or the test flawed. 

[esldude]

5 minutes ago, Superdad said:

 

You could have gotten 10/10 or 100/100 and still there will be a bunch who will call the result statistically invalid or the test flawed. 

I'm not calling it flawed just trying to understand what changed. 

 

I can respect Mani for actually doing the listening test unlike some. 

[esldude]

8 minutes ago, manisandher said:

 

It's not 17/30, it's 9/10. The first two test were done 'incorrectly'... and the result showed.

 

Mani.

Okay reading back over it, in the first two tests did you just play the track 10 times in random order and say if you thought it was 1 or 2. 

 

Then in the third test you played 1, played 2, then played X and chose whether it was 1 or 2?

 

If so, yes the first two runs were less than optimum, and the third one would have been the way to do it.  You have a fresh reference that way. 

[esldude]

And not being a user of your software XX Highend what is different about SFS .1 vs 200 in terms of what is happening to the playback. 

[manisandher]

2 minutes ago, esldude said:

Okay reading back over it, in the first two tests did you just play the track 10 times in random order and say if you thought it was 1 or 2. 

 

Then in the third test you played 1, played 2, then played X and chose whether it was 1 or 2?

 

If so, yes the first two runs were less than optimum, and the third one would have been the way to do it.  You have a fresh reference that way. 

 

Yes, exactly.

 

Mani.