Home |
Search |
Today's Posts |
#1
|
|||
|
|||
subjectivity versus objectivity
This is aimed more at JA (et al) than AK (et al), but both should read and
consider. Last Saturday morning around 8AM I was cleaning up my CD collection. I decided to turn on the system and play some background music. (I chose an unheard disc of Thomas Hampson singing artsy-schmartsy settings of various Whitman poems. Whitman would have derided most of them as "frou-frou". Absolutely /not/ recommended.) Anyhow, the sound was absolutely exquisite -- simply in terms of "beauty", the best I'd ever heard from my system. Do I need to add that, a few hours later, the system still sounded "good", but nowhere nearly as "pretty"? Subjective reviewers rarely, if ever, acknowledge that the time of day, ambient noise, their mood, etc, affects both what they perceive and how they perceive it. (When I wrote for Stereophile, I did my most-critical listening Saturday morning, when ambient sound was at a minimum and my hearing was most acute.) The same is true, to a lesser extent, of AB testing. If your mind and hearing are not at their most acute, you might miss legitimate differences. |
#2
|
|||
|
|||
You raise an interesting and valid point. However, the only way to
remove listener variability (with respect to time of day, mood, internal chemistry, etc.) is to average it over many sessions under all variants of those parameters. And, of course, to try to be self-aware of them. Kal On Sat, 25 Jun 2005 06:35:21 -0700, "William Sommerwerck" wrote: This is aimed more at JA (et al) than AK (et al), but both should read and consider. Last Saturday morning around 8AM I was cleaning up my CD collection. I decided to turn on the system and play some background music. (I chose an unheard disc of Thomas Hampson singing artsy-schmartsy settings of various Whitman poems. Whitman would have derided most of them as "frou-frou". Absolutely /not/ recommended.) Anyhow, the sound was absolutely exquisite -- simply in terms of "beauty", the best I'd ever heard from my system. Do I need to add that, a few hours later, the system still sounded "good", but nowhere nearly as "pretty"? Subjective reviewers rarely, if ever, acknowledge that the time of day, ambient noise, their mood, etc, affects both what they perceive and how they perceive it. (When I wrote for Stereophile, I did my most-critical listening Saturday morning, when ambient sound was at a minimum and my hearing was most acute.) The same is true, to a lesser extent, of AB testing. If your mind and hearing are not at their most acute, you might miss legitimate differences. |
#3
|
|||
|
|||
"Kalman Rubinson" wrote in message
... You raise an interesting and valid point. However, the only way to remove listener variability (with respect to time of day, mood, internal chemistry, etc.) is to average it over many sessions under all variants of those parameters. And, of course, to try to be self-aware of them. Right. |
#4
|
|||
|
|||
William Sommerwerck wrote: "Kalman Rubinson" wrote in message ... You raise an interesting and valid point. However, the only way to remove listener variability (with respect to time of day, mood, internal chemistry, etc.) is to average it over many sessions under all variants of those parameters. And, of course, to try to be self-aware of them. Right. Or make an objective measurment. Mark |
#5
|
|||
|
|||
On 25 Jun 2005 18:47:53 -0700, "Mark" wrote:
Or make an objective measurment. An objective and relevant measurement. Chris Hornbeck "I can build you a test that will show either one. Which would you prefer me to demonstrate? --scott |
#6
|
|||
|
|||
Chris Hornbeck wrote:
On 25 Jun 2005 18:47:53 -0700, "Mark" wrote: Or make an objective measurment. An objective and relevant measurement. There's the rub, the relevance part. Interpreting objective measurements is at once difficult and easy. On the one hand, nonlinear distortion that measures below 0.05% over the entire audible range +/- an octave or so is pretty well guaranteed to be inaudible. However, loudspeakers still commonly have far more than 0.05% nonlinear distortion in typical use. So, how do you characterize the way that loudspeaker sounds? |
#7
|
|||
|
|||
Arny Krueger wrote:
Interpreting objective measurements is at once difficult and easy. Now there's a decisive statement. Can't dispute those findings. So, how do you characterize the way that loudspeaker sounds? Well you could start by taking the pencils out of your ears and LISTENING? But that might confuse you Arny, so nevermind. VB |
#8
|
|||
|
|||
Vinyl_Believer wrote:
Arny Krueger wrote: Interpreting objective measurements is at once difficult and easy. Now there's a decisive statement. Can't dispute those findings. So, how do you characterize the way that loudspeaker sounds? Well you could start by taking the pencils out of your ears and LISTENING? That's quite an insult there, VB. Can you prove that: (1) I ever listen with pencils in my ears (2) That I don't evaluate loudspeakers by listening tests? But that might confuse you Arny, so nevermind. Where's the confusion VB? I'm doing small volume production and distribution of recordings all of the time. By noon today I will have pulled 7 live sound gigs in 7 days. If I don't listen, how is it that I do that? Oh, I get it, VB. On your planet people do audio production and live sound without listening. Must be horrible! |
#9
|
|||
|
|||
Well your question Arny..... "So, how do you characterize the way that
loudspeaker sounds?" ...... Is ridiculous. You listen and describe the sound ..... Measurements are nice but they usually stay in the bottom of the box. Glad you're working but you'd better hit the hay if you're gonna make that noon show at the school for the deaf..... Doh!!! g VB |
#10
|
|||
|
|||
Vinyl_Believer wrote:
Well your question Arny..... "So, how do you characterize the way that loudspeaker sounds?" ...... Is ridiculous. Where did I ask that question in the post you are pretending to replyt to, VB? Are you as delusional as Harry Lavo, who finds all sorts of things in my posts that I never said and even vociferously deny? You listen and describe the sound ..... You said I don't, now you say this VB? Are you a little confused or a lot confused? Measurements are nice but they usually stay in the bottom of the box. What does that mean? Glad you're working but you'd better hit the hay if you're gonna make that noon show at the school for the deaf..... Doh!!! g As usual VB you are making it up as you go along. There ain't no noon show, and this ain't no school for the deaf. Sorry to disappoint you, VB. |
#11
|
|||
|
|||
Arny Krueger wrote:
There's the rub, the relevance part. Interpreting objective measurements is at once difficult and easy. On the one hand, nonlinear distortion that measures below 0.05% over the entire audible range +/- an octave or so is pretty well guaranteed to be inaudible. However, loudspeakers still commonly have far more than 0.05% nonlinear distortion in typical use. So, how do you characterize the way that loudspeaker sounds? This just shows why using a single scalar number for distortion is misleading. It doesn't tell the whole story. --scott -- "C'est un Nagra. C'est suisse, et tres, tres precis." |
#12
|
|||
|
|||
So, how do you characterize the way that loudspeaker sounds?
Well you could start by taking the pencils out of your ears and LISTENING? But that might confuse you Arny, so nevermind. Forgive me for defending Arny, but Mr. Krueger was talking about finding a relationship between what we hear and what we measure. Simply listening does not produce a correlation. |
#13
|
|||
|
|||
"Arny Krueger" wrote in message ... Vinyl_Believer wrote: Well your question Arny..... "So, how do you characterize the way that loudspeaker sounds?" ...... Is ridiculous. Where did I ask that question in the post you are pretending to replyt to, VB? Are you as delusional as Harry Lavo, who finds all sorts of things in my posts that I never said and even vociferously deny? snip Since you dragged me into this Arny, you said: "So, how do you characterize the way that loudspeaker sounds?" in post: NNTP-Posting-Date: Sun, 26 Jun 2005 04:55:50 -0500 From: "Arny Krueger" Newsgroups: rec.audio.pro References: . com Subject: subjectivity versus objectivity I rest my case. |
#14
|
|||
|
|||
Scott Dorsey wrote:
Arny Krueger wrote: There's the rub, the relevance part. Interpreting objective measurements is at once difficult and easy. On the one hand, nonlinear distortion that measures below 0.05% over the entire audible range +/- an octave or so is pretty well guaranteed to be inaudible. However, loudspeakers still commonly have far more than 0.05% nonlinear distortion in typical use. So, how do you characterize the way that loudspeaker sounds? This just shows why using a single scalar number for distortion is misleading. It doesn't tell the whole story. To clarify, a scalar number for distortion is definitive if that number is an upper limit, below which all equipment-generated distortions lie, and if that upper limit is suitably low. OTOH, a single scalar number is not definitive if that number is too high, or not an upper limit. |
#15
|
|||
|
|||
Harry Lavo wrote:
"Arny Krueger" wrote in message ... Vinyl_Believer wrote: Well your question Arny..... "So, how do you characterize the way that loudspeaker sounds?" ...... Is ridiculous. Where did I ask that question in the post you are pretending to replyt to, VB? Are you as delusional as Harry Lavo, who finds all sorts of things in my posts that I never said and even vociferously deny? snip Since you dragged me into this Arny, you said: "So, how do you characterize the way that loudspeaker sounds?" in post: NNTP-Posting-Date: Sun, 26 Jun 2005 04:55:50 -0500 From: "Arny Krueger" Newsgroups: rec.audio.pro References: . com Subject: subjectivity versus objectivity I rest my case. Harry, I'm quite sure that you don't realize that your post muct be absolutely meaningless to me. But it is, because the purported quote from me has been stripped of all context. |
#16
|
|||
|
|||
William Sommerwerck wrote:
Arny Krueger wrote: So, how do you characterize the way that loudspeaker sounds? Vinyl Posture-Matic spewed: Well you could start by taking the pencils out of your ears and LISTENING? But that might confuse you Arny, so nevermind. OK VB, so you have a pathological fear of measurements. Forgive me for defending Arny, but Mr. Krueger was talking about finding a relationship between what we hear and what we measure. That would probably be way over VB's head. Simply listening does not produce a correlation. Exactly. |
#17
|
|||
|
|||
Arny Krueger wrote:
Vinyl_Believer wrote: ...snip.. Measurements are nice but they usually stay in the bottom of the box. What does that mean? ..snip... I'd guess at an implied education gap. Most folks (perhaps V.B. included) don't know how to read AND interpret tech specs and measurements. [ YMMV ] Later... Ron Capik cynic in training -- |
#18
|
|||
|
|||
Ron Capik wrote:
Arny Krueger wrote: Vinyl_Believer wrote: ...snip.. Measurements are nice but they usually stay in the bottom of the box. What does that mean? ..snip... I'd guess at an implied education gap. Most folks (perhaps V.B. included) don't know how to read AND interpret tech specs and measurements. [ YMMV ] Later... Ron Capik cynic in training Sure they do. They just understand that it doesn't end there. The AKG C1000 specs out pretty good, doesn't it? |
#19
|
|||
|
|||
Joe Sensor wrote:
Ron Capik wrote: ...snip.. I'd guess at an implied education gap. Most folks (perhaps V.B. included) don't know how to read AND interpret tech specs and measurements. [ YMMV ] Later... Ron Capik cynic in training Sure they do. They just understand that it doesn't end there. The AKG C1000 specs out pretty good, doesn't it? Don't know, I left the specs in the bottom of the box. ;-) Later... Ron Capik cynic in training -- |
#20
|
|||
|
|||
"Arny Krueger" wrote in message ... Harry Lavo wrote: "Arny Krueger" wrote in message ... Vinyl_Believer wrote: Well your question Arny..... "So, how do you characterize the way that loudspeaker sounds?" ...... Is ridiculous. Where did I ask that question in the post you are pretending to replyt to, VB? Are you as delusional as Harry Lavo, who finds all sorts of things in my posts that I never said and even vociferously deny? snip Since you dragged me into this Arny, you said: "So, how do you characterize the way that loudspeaker sounds?" in post: NNTP-Posting-Date: Sun, 26 Jun 2005 04:55:50 -0500 From: "Arny Krueger" Newsgroups: rec.audio.pro References: . com Subject: subjectivity versus objectivity I rest my case. Harry, I'm quite sure that you don't realize that your post muct be absolutely meaningless to me. But it is, because the purported quote from me has been stripped of all context. It serves its purpose. It shows that you said what you just denied above you said. |
#21
|
|||
|
|||
Joe Sensor wrote:
Ron Capik wrote: Arny Krueger wrote: Vinyl_Believer wrote: ...snip.. Measurements are nice but they usually stay in the bottom of the box. What does that mean? ..snip... I'd guess at an implied education gap. Most folks (perhaps V.B. included) don't know how to read AND interpret tech specs and measurements. [ YMMV ] Later... Ron Capik cynic in training Sure they do. They just understand that it doesn't end there. The AKG C1000 specs out pretty good, doesn't it? Based on specs, it looks like a mic with a rough, bright mid-high end, quite a bit of proximity effect, and a thin low end for sources at some distance. Based on specs, it could easily sound pretty ugly. |
#22
|
|||
|
|||
Measuring sound is of limited use and numbers can and are manipulated.
A graph is only 2D, and we listen in a 3D world with other senses and cues at play (feeling for example). You can measure volume, time, distortion, S/N, but what else? ...... You can't measure depth of sound, dimension or even naturalness of sound, prescence, or many other complex sonic experiences. Although Arny seems to have invented a technique to measure 'glare'. VB |
#23
|
|||
|
|||
Vinyl_Believer wrote:
Measuring sound is of limited use and numbers can and are manipulated. The same applies to listening evaluations. Listening tests are of limited use because they are or can be profoundly affected by the state of body and the state of mind of the listener. It is quite easy to manipulate the outcome of listening tests by a number of well-known means: (1) Don't match the levels well enough, or intentionally increase or decrease the level of one of them. (2) Displace the sonic alternatives in time so far that small real differences can't be reliably detected. A common means for implementing this scam is called "The Single Presentation Mode". (3) Before the listening evaluation, convince the listener that he's going to hear an improvement during one or more phases of the test. If the listener is already a True Believer in some form of technology such as SETs or vinyl, so much the better. A graph is only 2D, and we listen in a 3D world with other senses and cues at play (feeling for example). Actually, a lot of technical tests involve what amount to be graphs of 3D (or more) functions. The 2D graphs are just generally-accepted approximations of them. You can measure volume, time, distortion, S/N, but what else? .. In fact what other parameters does an audio signal have? You can't measure depth of sound, Sure you can. There are some very specific technical operations that we perform or situations we contrive in order to increase the perception of depth in sound. One well-known example is stereo. These technical operations and situations leave measurable technical artifacts. Their results can be easily measured. dimension or even naturalness of sound, This is known as an absence of audible distortion and noise. prescence, Those of us who know how to spell presence mostly know that it is due to a certain kind of spectral balance, Joe. or many other complex sonic experiences. They are all technical effects or the results of preserving certain technical properties of signals. Although Arny seems to have invented a technique to measure 'glare'. Glare is usually a combination of a certain spectral balance and certain kinds of nonlinear distortion. I can create a range of kinds of glare with my DAW. Certain mics in certain situations can also create this perception. It's real obvious to me Joe that a lot of your problem is that you don't know enough about audio production, and/or don't have enough experience doing audio production. Get thee into a control room! |
#24
|
|||
|
|||
Arny Krueger wrote:
Listening tests are of limited use .... That's about the most asinine statement I've ever heard coming from someone in the audio field. It's real obvious to me Joe..... You're talking to your 'voices' Arnie....... This is VB you're responding too. VB |
#25
|
|||
|
|||
Vinyl_Believer wrote:
Arny Krueger wrote: Listening tests are of limited use .... That's about the most asinine statement I've ever heard coming from someone in the audio field. Why don't you deal with the critical points I raised, VB? Here they are again: Listening tests are of limited use because they are or can be profoundly affected by the state of body and the state of mind of the listener. It is quite easy to manipulate the outcome of listening tests by a number of well-known means: (1) Don't match the levels well enough, or intentionally increase or decrease the level of one of them. (2) Displace the sonic alternatives in time so far that small real differences can't be reliably detected. A common means for implementing this scam is called "The Single Presentation Mode". (3) Before the listening evaluation, convince the listener that he's going to hear an improvement during one or more phases of the test. If the listener is already a True Believer in some form of technology such as SETs or vinyl, so much the better. |
#26
|
|||
|
|||
Arny Krueger wrote:
To clarify, a scalar number for distortion is definitive if that number is an upper limit, below which all equipment-generated distortions lie, and if that upper limit is suitably low. If we could all agree on what "suitably low" is, I'd agree with that. But we cannot. It would have to be very, very low. --scott -- "C'est un Nagra. C'est suisse, et tres, tres precis." |
#27
|
|||
|
|||
Scott Dorsey wrote:
Arny Krueger wrote: To clarify, a scalar number for distortion is definitive if that number is an upper limit, below which all equipment-generated distortions lie, and if that upper limit is suitably low. If we could all agree on what "suitably low" is, I'd agree with that. But we cannot. It would have to be very, very low. Nahh, and a few good clean DBTs can make that quite clear. Just get all forms of nonlinear distortion below 0.05%... Easy for electronics and digital, mission impossible for speakers. |
#28
|
|||
|
|||
Who uses pencils now anyway?
|
#29
|
|||
|
|||
studiorat wrote:
Who uses pencils now anyway? Analog bigots and other Luddites? |
#30
|
|||
|
|||
Had the OP listened to something half decent, he might have not have
had to write his post. I often notice that something sounds must better when you are not actually critically listening to it. I noticed this recently when I remarked on how a piano sounded beautiful on a vocal mic I was using at the time, I then read endless posts about mic placement etc. when I was really just pointing out that if you are not listening for problems you won't notice them. It means a difference between work listening and listening for preasure. I've never had an 'entuastiasts' hifi, and anyone who would consider owning one I believe deserves to hear every % of distortion the thing makes. Paying $4000 or whatever for a bloody cd player! What the ****? And Walt Whitman, would kick his ass for listening to that ****e in the first place. So there. It's about the message and not the medium, surley the proliferation of Mp3 has shown that by this stage. |
#31
|
|||
|
|||
"I've never had an 'entuastiasts' hifi, and anyone who would
consider owning one I believe deserves to hear every % of distortion the thing makes. Paying $4000 or whatever for a bloody cd player! What the ****? And Walt Whitman, would kick his ass for listening to that ****e in the first place. So there. It's about the message and not the medium, surley the proliferation of Mp3 has shown that by this stage. " very subjective reactionary statement. now how about something objective and visionary.... the OP closed with this "If your mind and hearing are not at their most acute, you might miss legitimate differences. " in your case you missed it ! |
#32
|
|||
|
|||
On Sun, 26 Jun 2005 13:28:24 -0400, "Arny Krueger"
wrote: To clarify, a scalar number for distortion is definitive if that number is an upper limit, below which all equipment-generated distortions lie, and if that upper limit is suitably low. OTOH, a single scalar number is not definitive if that number is too high, or not an upper limit. This might ultimately prove to be true, but I'm still unconvinced by current number crunching. Easy enough to foresee a brighter day soon when believable thresholds have been tested and a single number could be calculated as the sum of various error probabilities. Maybe something along the lines of: the probability of the average trained observer detecting error A plus(?) the probability of t.a.t.o. detecting error B, plus, etc. Thanks, Chris Hornbeck "I can build you a test that will show either one. Which would you prefer me to demonstrate? --scott |
#33
|
|||
|
|||
Chris Hornbeck wrote:
On Sun, 26 Jun 2005 13:28:24 -0400, "Arny Krueger" wrote: To clarify, a scalar number for distortion is definitive if that number is an upper limit, below which all equipment-generated distortions lie, and if that upper limit is suitably low. OTOH, a single scalar number is not definitive if that number is too high, or not an upper limit. This might ultimately prove to be true, but I'm still unconvinced by current number crunching. It's not the number crunching that proves it, its the listening. We know from listening that the ear's sensitivity peaks around 4 KHz. We know that the ear is vastly less sensitive at 20 Hz and 20 KHz, than it is at 4 KHz. Nonlinear distortion is most audible when the consequences of the nonlinear distortion end up at frequencies where the ear is more sensitive to the distortion than it is sensitive to the signal that caused the distortion. There are two cases where this happens - harmonics created by a low frequency signal, and intermodulation difference tones created by a high frequency signal. We also know that in general, modern equipment is generally free of distortion due to steps or sharp discontinuities in its input/output amplitude response. Distortion of that kind in digital convertors died when sigma-delta converters became the rule. Distortion of that kind in analog equipment was almost always due to biasing errors in push-pull amplifiers, but the art of properly biasing power amps became well-understood some decades ago. The major exception to this is clipping, but modern equipment has enough dynamic range that there is really no excuse to clip it. Nevertheless, audio gear still has distortion due to curvature of its input-output amplitude response. This distortion generally generates more low-order distortion than high-order distortion. Because the most likely form of distortion in equipment is of a lower order, we can conclude that generation of distortion products in the 4 KHz range by a lower frequency signal is far less likely than the generation of distortion products in that same range by means of intermodulation of higher frequency signals. I spent quite a bit of time listening to music whose basic nature would tend to generate a lot of this kind of distortion, with and without added distortion of various kinds. I also looked at masking curves of the human ear to see exactly how sensitive the ear could be reasonably be expected to be in these kinds of scenarios. The results of both of these kinds of investigations gave similar results - distortion that is much below 0.1% just can't be heard, even under the most ideal conditions for hearing distortion. |
#34
|
|||
|
|||
now how about something objective and visionary.... the OP closed with this "If your mind and hearing are not at their most acute, you might miss legitimate differences. " in your case you missed it ! I think you missed the fact that he was talking about A/B testing there Dale. Are we listening to the equipment or the music here? |
#35
|
|||
|
|||
" Are we listening to the equipment or the music here? "
both. for neither can exist for this discussion without the other. and most music today has musicians who can not play without the equipment! and cannot judge the equipment as good or bad! |
#36
|
|||
|
|||
On Tue, 28 Jun 2005 06:56:18 -0400, "Arny Krueger"
wrote: carefully considered, well presented and completely correct analysis clipped 'cause ya already read it Yes. And yet... Competent profession observers still report hearing differences between very competent and expensive A/D and D/A converter boxes, meaning those with 80 dB performance above artifacts. It's this dichotomy that interests me in the topic. There's something interesting here; not yet sure what it means. Thanks, as always, Chris Hornbeck "I can build you a test that will show either one. Which would you prefer me to demonstrate?" --scott |
Reply |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Forum | |||
Basic Gain Staging and +4 versus -10 | Pro Audio | |||
Behringer EP-1500 & 2500 versus QSC RMX controversy | Tech | |||
A comparative versus evaluative, double-blind vs. sighted control test | High End Audio | |||
Great Web Site- Subjectivism in the world of Objectivity. | Audio Opinions |