The rational brain is the part of the brain good for rationalizing, or in more modern terminology, confirmatory thought. Nothing beats experimentation and the scientific method, even though it works in fits and starts and with errors along the way.
The people who focus on the frequency domain really shouldn't be let lose on equalising speakers. At least IMO. But this is where having a digital crossover is great. Want to try a Harsch 'transient perfect' crossover, no problem. An NTM crossover, go for it. Time aligment, no problem. As long as you can describe it in biquads you can build it in minutes. Won't turn a sow's ear into a nice purse, but does give freedom to experiment, optimise and find out what makes a difference.
Do it passive or analog and you generally don't try all the options and may miss something. Or just give up and say 'good enough'.
Should add, having spent a good chunk of this week manually decoding base64 messages reality is a little warped for me right now
Do it passive or analog and you generally don't try all the options and may miss something. Or just give up and say 'good enough'.
Should add, having spent a good chunk of this week manually decoding base64 messages reality is a little warped for me right now
Last edited:
Agreed. But, if at all possible only go through a D/A once. And make it a really good one. If the source material is already digital, no harm done, assuming well written crossover DSP.
Mark: Once upon a time I would have believed you, but seeing how places like AIR run digital out, through the Neve or SSL and then redigitise I hate to think how many conversions some recording go through.
But I am optimising for digital playback. Records are fun, the ritual is fun and sometimes
astound me by how good they can sound given the flaws, but for me digital is superior in sound quality. And I wont pay £30 for a new vinyl pressing when a CD is available for £10.
But I am optimising for digital playback. Records are fun, the ritual is fun and sometimes
astound me by how good they can sound given the flaws, but for me digital is superior in sound quality. And I wont pay £30 for a new vinyl pressing when a CD is available for £10.
Hi Mark, I just got around to taking a listen.
The differences in the files are abundantly clear on my laptop internal (harmon/kardon) speakers and same on my lounge room system *.
Without spending too much time I came to some observations.
The original source clarity is apparent.
G.wav is constantly 'dirty' compared to Q.wav which whilst 'cleaner' has some annoyances of it's own, mostly a high mids/highs flaring/shoutiness when things get busy.
So whilst Q.wav is generally cleaner the artefacts are 'moving' and draw attention and are thereby fatiguing.
Q.wav bass goes lower and overall sound is cleaner up through to the mids then it goes sour with unnatural essing/flaring on high mids/highs which ultimately drives me out of the room.
G.wav dirt is constantly dirty with no/little untoward frequency dependant dynamic/distortion behaviours and sounds duller/dirtier but this noise can be 'listened through' for an overall more 'musical' but less detailed listening.
Conclusion - the two interfaces are clearly different, and with quite different behaviours.
If I had to pick, I wouldn't, I don't find either recording satisfying.
Mark, I reckon there is a constant noise in your system sitting on both recordings.
I have had very good success with battery powered USB isolator/regenerator pcb from Ebay to keep PC noise/PC ground noise out of the equation.
This gives an overall cleaning and new 'stability' in the sound, akin to lowered and 'nicer' jitter, highly recomended for $25.00 or so plus a bit of tweaking.
Dan.
* - Current lounge system signal path - Foobar>Soundwire Link Software>USB WiFi Adaptor>Household modem/router>Galaxy Note 8.0>3.5-3.5 cable>Sony mini shelf system speaker output>Behringer active speakers.
Even over this convoluted and arguably hopeless (it actually sounds bloody good) signal path the file differences are clearly audible and reliably identifiable.
I have Foobar on repeat running minimised and when I check I have been 100% correct in my G.wav/Q.wav 'blind' decisions.
When you get your system running properly it might be worth the experiment to compare.
Dan.
Maybe I wasn't clear. I said that I would avoid it if at all possible. I have gone though extra data conversions when the benefits outweigh the damage inflicted from doing it. For example, when mastering a digital mix, if analog mastering processing will do enough good, then I am willing to accept the damage due to extra data conversions. In such cases some people prefer to use DSD to get to analog and back because they think it sounds better, and that may well be preferable in many cases.
The other thing that is going on with modern pop record production is there is so much digital processing and so much damage from it, that listeners are just getting used to it, like they have gotten used to the sound of limiting and clipping to make music louder. If the public will accept damaged audio, I guess some people find no reason to be ultra careful with it. Besides, conversion to mp3 or AAC often removes or lessens some of the objectionable distortion.
Last edited:
Actually, not my recordings. They are some of the recordings I listened to before deciding to purchase the HEDD, which was my 2nd choice. The first choice was too expensive to justify, with the HEDD being more than painful enough in that regard.
No this is logic inverted. In the situation you picture, the only logical conclusion is that the tester is biased in some way.
After all, in a properly set up test, it is the test subject who determines whether there is a difference or not, not the tester (assuming you mean the tester is the guy who administers the test) . That's the whole point Ed.
Jan
You should not get so hung up on ABX which is just one of many protocols.
So just for the sake of clarity. You are listening to say two amplifiers and the differences are glaring.
But when someone shoves a visual screen between your eyes and the amps, and you can't see what's playing at any time, your incredible sensitive hearing collapses and becomes non-existent.
When someone moves a light blocker between the amps and your eyes, your incredible hearing acuity drops to zero.
When someone shoves a light screen between your eyes and the amps, your fantastic golden ears crumble to dust.
Unbelievable.
Jan
Last edited:
Jan,
If you follow the report after disclosing the noise difference to the subject they ran 20 more trials. However due to a testing mistake all 20 trials were the same. The subject under those conditions did report an asymmetric number of differences. Now read the author's conclusions. Either that shows inexperience or a strong bias.
As the issue was does inserting the digital gizmo make a difference the answer must be yes if any observer can note accurately the difference.
Now I may recall incorrectly but the digital gear used in that report was only 9 bits linear.
Later testing does show that level of encoding is not transparent.
As you are aware if you run enough tests there will be an outlier that can support almost any theory.
BTY thanks for the AX reprint comment.
If you follow the report after disclosing the noise difference to the subject they ran 20 more trials. However due to a testing mistake all 20 trials were the same. The subject under those conditions did report an asymmetric number of differences. Now read the author's conclusions. Either that shows inexperience or a strong bias.
As the issue was does inserting the digital gizmo make a difference the answer must be yes if any observer can note accurately the difference.
Now I may recall incorrectly but the digital gear used in that report was only 9 bits linear.
Later testing does show that level of encoding is not transparent.
As you are aware if you run enough tests there will be an outlier that can support almost any theory.
BTY thanks for the AX reprint comment.
I have actually been listening to a lot of vinyl from the mid-50's to mid-60's lately. My wife has finally sold her parents' house and is clearing it out, and I have received a lot of her father's LP's (her cousin got the classical, I got the jazz
).An excellent example would be a recording on the Verve label called "Little Jazz" featuring trumpet player Roy Eldridge along with Oscar Peterson, Ray Brown, Herb Ellis, and Buddy Rich. Musically wonderful, and sonically very good in spots. Of course I know the vinyl was played on some substandard systems over the years that could have affected it, and I certainly cleaned all the albums very carefully, but it is very interesting to listen to. The early tracks on side A sound distorted and overloaded at times, even though Eldridge was using a mute. Still it sounds like the engineers were struggling to get the levels right, and moved Eldridge around to avoid overload. So it briefly overloads, then the level of the trumpet goes down (they turned it down), then the trumpet level goes down again (they moved him back from the mic), then the level goes up again but cleaner. By the second side they seem to have got it figured out.
So my point is that the music was great but the sound was all over the map, and that was ON ONE RECORD. I have also heard recordings on Prestige from the 50's or 60's that were terrible, though the music was good enough to justify listening through the garbage. There is no question that pressing quality (vs recording quality) went downhill horribly post OPEC oil shock in the late 70's. So when evaluating older analog recordings we have to somehow distinguish the quality of the recording vs the quality of the pressing vs the quality of what is left after decades of abuse or neglect. There was an awful lot of dreck pressed into vinyl back in the day.
That test, overall and in all separate parts, shows convincingly what will/can happen if the perceived differences are not due to actual differences.
Jan
Later testing showed...
It has been shown that...
Everyone knows that...
Funny, I always feel left out when I read such unsubstantiated stuff
Jan
Later testing showed...
It has been shown that...
Everyone knows that...
Funny, I always feel left out when I read such unsubstantiated stuff
Jan
Are you really going to argue 9 bits is enough?
If so try reading JJ's papers.
Yes hearing differences when there are none is an issue. Putting the subject on the spot by telling them there are differences when there are none is a valid experiment but not for what was the goal in the experiment in question.
A psychology experiment is I suspect easier to design than a listening one.
I consider the cited experiment to be accurate for determine the subject's sensitivity and motivations. But not for testing the hypothesis of is digital transparent. (In actuality it showed a trained listener could detect differences.)
Your are of course welcome to your own opinion, works for the Grand Negus.
Grand Negus.
Had to look that up. Interesting role model you have, Ed.
Do you require people to kiss your staff too?
Jan
Had to look that up. Interesting role model you have, Ed.
Do you require people to kiss your staff too?
Jan
Not me, apply a bit if logic and you will figure out who the appellation applies to!
Using the ABX method, participants internal mental processes are different compared for example to an A/B - protocol. That could and did led to different results dependent on the protocol used, researchers did note that already in the 1960s.
We´ve discussed the example that jan.didden linked quite often, and as i´ve said before, there was a positive control (by accident, call it hidden positive controls; in fact even two) and Ivor Tiefenbrunn missed it. So it is safe to conclude that he did not reach the highest sensitivity level under the specific conditions of that experiment.
I beg to differ; controlled listening tests aren´t per se immune against interference from listeners expectations that is one of the reasons why it is quite often better to let participants do training under the specific test conditions.
Sometimes protocols were used, where the participants even don´t know what the EUT really is/was, just to avoid a certain expectation. But in any case it is difficult for humans to not expect something, usually they have to learn that too .
The best way is to carefully express which question/hypothesis should be examined and then work out which test procedure to use.
Don´t believe in any test results per se, ask if the experiment was objective, valide and reliable. If you can´t evaluate that due to a lack of information remain sceptic.
- Status
- Not open for further replies.