What's great about typing on a forum is that you can do it while listening to music, can you imagine having this conversation while music is playing?
Re, tilt/tone controls, I've just built Doug Self's adjustable tone control preamp and it gives very useful flexibility and helps me a lot with my deteriorating hearing
Re, tilt/tone controls, I've just built Doug Self's adjustable tone control preamp and it gives very useful flexibility and helps me a lot with my deteriorating hearing
Slihjtly OT: does anybody here have a datasheet for the vintage 2SC1637 low noise bipolar? Google let me down.
Jan
Jan
It is interesting to me that engineers in this field are very aware of biases which effect comparisons with respect to the quality of equipment (A/B testing and so forth), yet not so much in the comparisons which effect their own logic and how they interact with people rather than inanimate objects.
However, I suppose this is also a type of bias or phenomena, probably in the way an engineer might utilize the same problem solving techniques professionally and personally.
I feel an engineer should rightly have this perfectionist mentality. If you didn't what would you have to aspire to? Everyone needs to feel competent in their profession. Many more people would surely be electrocuted if it were not this way.
This also breeds a reductionist mindset, in which others or the "out group" are seen as misguided or annoyances or so forth. This happens in politics frequently (please let's not talk about politics) and all sorts of other areas.
It's very easy to be a reductionist, but one shouldn't confuse that with a strategy toward a perfection or the approach of an apex. That's like a mole imagining it's flying.
My personal opinion is that the greatest knowledge comes from attempting to understand those with which we are most naturally at odds with, because that is a clue towards our own failings rather than theirs.
In this arena I'd imagine it would be the two camps of science/tech and the more casual hobbyists and "audiofools" as they are called.
Some of these terms might be interesting for conversation:
https://en.m.wikipedia.org/wiki/Distinction_bias
https://en.m.wikipedia.org/wiki/Curse_of_knowledge
https://en.m.wikipedia.org/wiki/Anchoring
https://en.m.wikipedia.org/wiki/Functional_fixedness
There's many more that could have relevance in this discussion, I just kind of pulled some that immediately came to mind, but I'll let you pull from the list what you find to be most intriguing:
https://en.m.wikipedia.org/wiki/List_of_cognitive_biases
https://en.m.wikipedia.org/wiki/List_of_fallacies
So, you're saying we aren't well versed with Kahneman, Haidt, Ariely, Tetlock, Thaler, Stanovitch, etc.?
😎🙂
BTW I've heard them at Yamaha's own showcase listening room where they demo to vendors and customers. I never read this stuff, a very silly story.
😎🙂
Well, what do you say RNM? Can you hear the difference between properly dithered 16-bit and 24-bit files of the same high-quality, high-rez, source? (I'm hoping you say you can.)
I've been trying for a while to him to agree to participate in a serious DBT. You will need to qualify "proper dither" I suspect. I'd be interested in how you will eliminate "turn it up at quiet parts" noise cheat.
Last edited:
Maybe we will have to find a suitable wav file segment with no quiet parts. Maybe also set playback level with an SPL meter to some standard value. Have to look into those kinds of things and see what we can come up with that works.
Probably if you just give people files, and someone wants to cheat in private, they can. Therefore, a serious study probably wouldn't be viable as a take home test. Have to get people together to be reasonably sure of an accurate outcome.
Last edited:
can I hear the difference...yes. Under these conditions ---> Modest CD player... mid price Sony consumer model for decades and now OPPO. Download an HD file of 24b run thru Araliti DAC and yes a big difference in increased resolution and detail in the music.
Now a wav file in 16 or 24 bit played thru BenchMark I have not tried. Only 16 b via CD player.
I play on the loud side of things. and Not often any classical... mostly blues or well recorded world music.
With the JBL M2 speakers I gained in many areas but the Crown amps lost something. But could be tried with dblt, I suppose. The new amp has the first batch of chassis arriving this month... so I could redu any dblt after they are installed here at my home.
THx-RNMarsh
Now a wav file in 16 or 24 bit played thru BenchMark I have not tried. Only 16 b via CD player.
I play on the loud side of things. and Not often any classical... mostly blues or well recorded world music.
With the JBL M2 speakers I gained in many areas but the Crown amps lost something. But could be tried with dblt, I suppose. The new amp has the first batch of chassis arriving this month... so I could redu any dblt after they are installed here at my home.
THx-RNMarsh
Last edited:
Mark this is falling very very quickly into "no true Scotsman" arguments.
AES E-Library Sampling Rate Discrimination: 44.1 kHz vs. 88.2 kHz
5 recorded pieces were simultaneously converted to 24/44.1 and 24/88.2. A third set of samples were made by software downsampling the native 88.2 to 44.1 (via Pyramix 6 software). Sennheiser MKH8020 mics in a simple stereo arrangement and RME equipment through the chain (branched after the mic preamp to identical RME ADC's running 24/44 and 24/88, respectively), Macbook Pro via firewire for I/O (and a separate recording box for the 44 line). No Windows involved. Dedicated listening room with B&W 802's. Music grad students and experienced staff of the music department were their test subjects. 3 ABX tests were run: native 44.1 vs native 88.2, native 44.1 vs downsampled 44.1, and native 88.2 vs downsampled 44.1.
Only thing "audible" was a resampling of the 88.2 to 44.1 (neither the native ADC's to 44.1 or 88.2 were seperable) on 1 of 5 tracks by 3/16 listeners. Which, in short, means sub-subgroup analysis to get a positive.
Summary:
Null test between 44.1 and 88.2 full stop. If there's anything of note, the Pyramix downsampling algorithm might be dodgy.
Discussion: It's as thorough a test as you can ask for, and, well, the likelihood of this null being turned over as a positive in a larger-scale study is slim-to-none. Usually regression to the norm goes the other way (positive -> null).
------------------------------------------------------------------------
AES E-Library The Audibility of Typical Digital Audio Filters in a High-Fidelity Playback System
Summary: Meridian looking at filters and dither from a 24/192 purchasable track of Haydn that they cut up into chunks and applied a -90 dB lowpass filter over 23.5 - 24 kHz (to model 48 kHz sampling) and 21591-22050 Hz (44 kHz sampling), respectively. Then they applied 1. no resampling 2. "16 bits" via zero populating the bottom 8 bits of the 24 (maintaining a 24/192 data stream) and 3.) same truncation as 2 with a rectangular dither applied. That makes 6 for tests in total.
Meridian's own equipment, with attendant 24/192 pipeline, was fed from a Macbook pro. A dedicated listening room was used. Filters were confirmed to be operating correctly by loopback.
8 experienced (recording studio engineers) subjects were allowed extensive time to train for the tests and listen to the selections in their various forms of filtration before testing. Forced choice pair comparisons were made between the 6 different filter/resampling methods and the native file over multiple test blocks to mitigate fatigue.
Testers were able to differentiate p <0.05 5/6 tests vs native 24/192, save the non-truncated 24 kHz low-pass. But BARELY. There was absolutely no difference in "success"/effect size between the 5 tests that were positive--all those box/whisker plots overlapped. While this is certainly interesting, the track snippets were short AND the longer filter tap length on the 24 kHz lowpass was identified better than the 22.05 kHz filter. Which makes me think this was more a clever test (and kudos to the testees for noticing) about pre-ringing than anything else.
Words were spilled about hearing differences between dithered and non-dithered, but zero tests were shown about a forced pair between those two selections so I take it all with a grain of salt. The test results do not lend much confidence to real differentiation between any of the 6 tests against each other, which would be contrary to Mark's commentary about 16 vs 16 dithered vs 24.
Ah, I was wondering about DSP in the Crown amps. Do you think that is the issue?
Anyway, should be easy to make some test files. I think somebody on DIYAudio already did it, but not a problem to make new ones.
May I ask if you tried listening to the A/D converter files I posted? They do sound different though my DAC-1 and NS-10s, but it can take some practice and careful listening to notice what to listen for. If any interest in giving it a try, here is a link: https://www.dropbox.com/sh/9r5kiiptr00seub/AAC5zEynkg-ZmriNSkQxFPaEa?dl=0
Daniel, using RME converters already has me worried. Which model? Also, what amplifiers and speakers? How far were listeners from the speakers, in the very near field? And I don't know what "Meridian's own equipment" means. Also, did they have someone on the research team who could hear the difference and verify the test setup? These are not silly questions.
Look, the thing is this: Some reputable mastering engineers claim to be able to hear the difference, I make the same claim. However, in my experience not all equipment and listening environments work for this type of testing despite specs that would make it seem like they should.
This has nothing to do with logical fallacies. There is no fallacy to say that doing the testing properly is trickier than some engineers are likely to expect. Actually, exactly what equipment and environment is used may not matter to me at all, so long as someone who can reliably hear can verify the test setup. That means someone who can hear needs to be identified and recruited. I would start looking with reputable people, not audiophiles, who can reliably identify test files on the experimenters equipment or their own equipment. The experimenters can provide the files and operate the equipment to make sure the recruitment test is honest.
By the way, I just had some recent communication with someone who claimed to manage 6 recording studios and a mastering room. He indicated they couldn't hear the difference between properly dithered 16-bit and 24-bit, without looking. The only thing that seemed make sense to me is that I suspect they didn't recruit people with listening tests. Or maybe they didn't choose the right equipment out of what they had on hand. On the other hand, I understand Bob Ludwig does use listening tests for employee recruiting. And I believe the people he hires can hear the difference. Such people exist, so I say let's get one or more of them on the research team.
Look, the thing is this: Some reputable mastering engineers claim to be able to hear the difference, I make the same claim. However, in my experience not all equipment and listening environments work for this type of testing despite specs that would make it seem like they should.
This has nothing to do with logical fallacies. There is no fallacy to say that doing the testing properly is trickier than some engineers are likely to expect. Actually, exactly what equipment and environment is used may not matter to me at all, so long as someone who can reliably hear can verify the test setup. That means someone who can hear needs to be identified and recruited. I would start looking with reputable people, not audiophiles, who can reliably identify test files on the experimenters equipment or their own equipment. The experimenters can provide the files and operate the equipment to make sure the recruitment test is honest.
By the way, I just had some recent communication with someone who claimed to manage 6 recording studios and a mastering room. He indicated they couldn't hear the difference between properly dithered 16-bit and 24-bit, without looking. The only thing that seemed make sense to me is that I suspect they didn't recruit people with listening tests. Or maybe they didn't choose the right equipment out of what they had on hand. On the other hand, I understand Bob Ludwig does use listening tests for employee recruiting. And I believe the people he hires can hear the difference. Such people exist, so I say let's get one or more of them on the research team.
Last edited:
No, no, no.
Not making claims about the general breadth / diversity of knowledge of various parties here.
I'm sure you're all insightful and learned people.
Also, this wasn't directed at your listening test in any respect whatsoever. I'm not even totally clear on it's particulars, I'll have to dig back through the thread.
It was regarding a sidebar / derailment on what is considered "real" musical reproduction, a "good" system, etc. and how these are fairly useless discussion points, and in a way show the differences in goals between designers and consumers, and also the influence of different biases on the part of both parties.
I was reflecting on Jan's comment about ongoing misunderstanding and conflict in this area and offered some general talking points of what could influence these issues and maybe help contain them.
yeh, the DSP is an issue. I don't know of anyone using Crown amps in studio for recording/mix down or mastering.
Don't recall.... if I did, it was in Bangkok when I did it .... and that case I heard differences.
I might use headphone/amp instead of speaker system.
But won't do anything soon... tomorrow I drive down to race track near Las Vegas to partake in high performance car driving/handling classes (2017 Cadillac CTS-V).
My B-Day present to myself compliments of Cadillac who is footing the bill.
-Richard
Last edited:
Okay. I think too much has been made about use of the word "good." We often speak somewhat loosely with an expectation that our words will be understood in the context at hand. If not, clarification can always be sought.
In my experience, it is impossible to write anything and be completely sure that no one can misinterpret anything. Natural language is inherently more ambiguous than mathematical language. And even in mathematics, symbols are sometimes reused (overloaded, in computer speak) and usage may need to be understood from context.
So, as a practical matter, if anyone feels that terminology is not sufficiently well defined in the discussion of some subject, it should be fine to say so and ask for clarification.
In addition, making a point that "good" may imply moral judgement is not particularly helpful, since it would seem rather unusual for an engineer to somehow judge something like loudspeakers in a moral sense. I would probably wonder more about someone making such an inference.
I will skip discussion regarding appropriateness of the use of the word masterbatory.
Last edited:
Regarding discussion of mental biases and heuristics, there is probably not going to be much interest in this forum. People come here because of some preexisting interest in audio, and not cognitive psychology. Trying to get other people interested in something that interests you, even if it is something that affects everyone, is unlikely to gain much traction. And even when people do learn about biases, it generally offers little or no immunity to the effects thereof. We are living in a time shortly after the discovery of many biases, when researchers are working to find effective ways of debiasing. For a few biases, such as anchoring, it is fairly easy. But for the most part, we are still at a fairly early stage. And the foregoing is assuming that people want to debias. Many people are quite attached to their biases -- they might not like hearing they are unduly affected by confirmation bias, for example. And there are some good reasons for that. I could go on and on, but I will quit here.
Last edited:
- Status
- Not open for further replies.