I post this thread with full knowledge that it is likely to lead to chaos, but I will try to keep it in line and on-topic and most hopefully civil.
The reason for this new thread is that out of respect for Pano, I don't want to disrupt his thread any more and out of respect to DDF, whose post there was thoughtful even though I thought that he was throwing out the baby with the bathwater.
The question that I will raise is at what point can we deviate from the ideal of "objective accuracy" and still be respectful of the title of "Hi-Fi" and respect for the audio art.
Many believe that anything that they do that makes things sound better to themselves is fine. I don't accept that, but then I am smart enough to know that there are no absolutes in this world and that there will always be exceptions.
One except that I accept is that the HF response should roll off slightly towards the HFs. This is widely accepted practice and is supported by many many experts. It took me a long time to come around to accepting this change as "correct".
Recently is was proposed that the center phantom image is dull because of comb filter effects. I questioned the validity of correcting at the playback because it may (should!) be corrected at the source. I did not get a convincing argument that this was not the case.
So when does a "trend" become "correct"? And when is it not "Hi-Fi" and disrespectful of the art?
Let the games begin!
The reason for this new thread is that out of respect for Pano, I don't want to disrupt his thread any more and out of respect to DDF, whose post there was thoughtful even though I thought that he was throwing out the baby with the bathwater.
The question that I will raise is at what point can we deviate from the ideal of "objective accuracy" and still be respectful of the title of "Hi-Fi" and respect for the audio art.
Many believe that anything that they do that makes things sound better to themselves is fine. I don't accept that, but then I am smart enough to know that there are no absolutes in this world and that there will always be exceptions.
One except that I accept is that the HF response should roll off slightly towards the HFs. This is widely accepted practice and is supported by many many experts. It took me a long time to come around to accepting this change as "correct".
Recently is was proposed that the center phantom image is dull because of comb filter effects. I questioned the validity of correcting at the playback because it may (should!) be corrected at the source. I did not get a convincing argument that this was not the case.
So when does a "trend" become "correct"? And when is it not "Hi-Fi" and disrespectful of the art?
Let the games begin!
Last edited:
Pooh - if you are just going to try and make a farce out of this then please don't come here.
FredrikC - If there is a room curve at the mixing studio then this curve should also be in our playback system. That's how I understand it.
It gets more complicated when we differentiate between the room frequency response and the direct field frequency response and this gets tied in with the directivity at HFs. A wider directivity will need more falloff in the direct field to get a "natural" falloff in the room response.
FredrikC - If there is a room curve at the mixing studio then this curve should also be in our playback system. That's how I understand it.
It gets more complicated when we differentiate between the room frequency response and the direct field frequency response and this gets tied in with the directivity at HFs. A wider directivity will need more falloff in the direct field to get a "natural" falloff in the room response.
Last edited:
OK, I just brought the objective reality to your sales pitch for the speakers nobody buys.
OK, I just brought the objective reality to your sales pitch for the speakers nobody buys.
Really? I thought you were just on full troll mode.
I notice these sorts of threads generate a lot of (at best) impatience to (at worst) hostility on DIY boards. I don't understand that response as its incredibly important to understand the goals that need achieving. IMO, implementation is far easier than answering the question about specification because there are so many open questions.
Put another way, to paraphrase what Michael Jordan said about shooting. You can practice 8 hours a day but if what you're doing is wrong, you become really good at being really bad.
I'm glad you started this thread Earl. I'll start by repeating my earlier post, for comment
Originally Posted by gedlee
I am all for standardization, so lets just agree that it is the way we should go.
Certainly agree that the recording side should, and the playback side should have full access to reproducing this option. Do I personally think we should design to be as accurate to a standard as possible? Yes, absolutely. Would I be willing to trade off a small amount of realism for the best 5% of recordings to make the worst 60% of recordings sound notably better? Absolutely. But my opinion isn't so interesting, I think your premise is objectively false
Quote: Originally Posted by gedlee
Now lets say that we got there (somehow!) would you now feel that you had the right to change the art because it didn't suit your taste.
Its a loaded question as the premise of the question is false. It IMO gives up on the pursuit of the illusion of realism (pursuing "art" instead). But lets even suspend that and go with it for a moment. It falsely presumes there is an objective way to do this that does not require individual tailoring. This is not how the human mind works and the science backs my statement up.
It falsely assumes the same boundary conditions for every replay situation: everyone has the same HRTF, every replay room is the same, and everyone has the same level of training. None of these are or ever will be true and so lead to different acceptable solutions for the individual.
HRTF: as I showed earlier, stereo has an inherent tonal error over lateral angle that is HRTF dependent. If the producers fixes this in eq, it is fixed for the producers hrtf, not the listeners. It may not be a huge error, but it is one and would require me to eq to hear what the producer even heard. Are there other errors created by stereo that are individualized? Probably, would have to think about it some more
Room affect on balance/imaging: if there was one standardized room, would I want to emulate in replay? Maybe, but I can tell you for certain it would be less enjoyable. As many studies showed, visual biases affect sonic perceptions. Why would I want to subject myself to a visually oppressive or less desirable room in the name of "objective" reproduction accuracy when my own perceptions (which are all that matter) will judge it a poorer experience or maybe even less accurate?
Here's where I'm pretty certain we will never agree but I believe the most worthwhile goal is that the illusion sounds more accurate to the individual listener. I thinks it oppressive (and the science proves it dead false) to insist that one person's perception or a technical standard can achieve that without individual tailoring.
For example, Olives tests show that training significantly improves a persons ability to discriminate and notice differences. We knew this in telecom forever and applied training in JND testing. The only reason to do this is to get the mean right. But asking people who buy stereos for enjoyment (not to be lab tools) to train themselves up to the level of the producer just so that what seems accurate to them is closer to the theoretical ideal strikes me as ludicrous. I agree that they should train themselves before making declarations regarding absolute accuracy, but don't confuse the two. If they are untrained and have perceptions that accuracy is better attained through technical inaccuracy, then technical inaccuracy is the right answer for them as it better achieves the end goal of the perception of accuracy for them.
As people with an interest in the science (ie making this predictable) we need to be able to distinguish the two, and I'm one the biggest fans and earliest adopters of Toole/Olive/Queen/Gressingers/Blaurts/name drop/name drop etc etc work but that in no way invalidates the perceptions of the untrained listener to them.
Quote: Originally Posted by gedlee
I just have to assume that the producers of the music that I listen to are smart enough to know about the circle of confusion and attempt to get to some standard that will reproduce the best possible reconstruction that I can get. Now we all know that not all of them do, but many do.
I think it's wishfull thinking to think their tastes don't affect the rendition. For example, many 70s DG classical recordings have a horrible 1st row perspective to the tonal balance. I'm sure it was "realism" to the conductor but ghastly to the me as an audience member. Multi-mic studio recordings are artificial and recording is tweaked to taste.
Different speaker design types even betray an obvious bias towards main floor (stand mount) or balcony (line source).
For me personally, I strive to train myself to be able to detect accuracy and true differences (I have CD training material and headphones at my bed side: Critical Listening Skills for Audio Professionals by F Alton Everest), but I would never presume to be authoritarian and to expect this to be ubiquitous, or to make light of other's perceptions if they don't pursue this path
Put another way, to paraphrase what Michael Jordan said about shooting. You can practice 8 hours a day but if what you're doing is wrong, you become really good at being really bad.
I'm glad you started this thread Earl. I'll start by repeating my earlier post, for comment
Originally Posted by gedlee
I am all for standardization, so lets just agree that it is the way we should go.
Certainly agree that the recording side should, and the playback side should have full access to reproducing this option. Do I personally think we should design to be as accurate to a standard as possible? Yes, absolutely. Would I be willing to trade off a small amount of realism for the best 5% of recordings to make the worst 60% of recordings sound notably better? Absolutely. But my opinion isn't so interesting, I think your premise is objectively false
Quote: Originally Posted by gedlee
Now lets say that we got there (somehow!) would you now feel that you had the right to change the art because it didn't suit your taste.
Its a loaded question as the premise of the question is false. It IMO gives up on the pursuit of the illusion of realism (pursuing "art" instead). But lets even suspend that and go with it for a moment. It falsely presumes there is an objective way to do this that does not require individual tailoring. This is not how the human mind works and the science backs my statement up.
It falsely assumes the same boundary conditions for every replay situation: everyone has the same HRTF, every replay room is the same, and everyone has the same level of training. None of these are or ever will be true and so lead to different acceptable solutions for the individual.
HRTF: as I showed earlier, stereo has an inherent tonal error over lateral angle that is HRTF dependent. If the producers fixes this in eq, it is fixed for the producers hrtf, not the listeners. It may not be a huge error, but it is one and would require me to eq to hear what the producer even heard. Are there other errors created by stereo that are individualized? Probably, would have to think about it some more
Room affect on balance/imaging: if there was one standardized room, would I want to emulate in replay? Maybe, but I can tell you for certain it would be less enjoyable. As many studies showed, visual biases affect sonic perceptions. Why would I want to subject myself to a visually oppressive or less desirable room in the name of "objective" reproduction accuracy when my own perceptions (which are all that matter) will judge it a poorer experience or maybe even less accurate?
Here's where I'm pretty certain we will never agree but I believe the most worthwhile goal is that the illusion sounds more accurate to the individual listener. I thinks it oppressive (and the science proves it dead false) to insist that one person's perception or a technical standard can achieve that without individual tailoring.
For example, Olives tests show that training significantly improves a persons ability to discriminate and notice differences. We knew this in telecom forever and applied training in JND testing. The only reason to do this is to get the mean right. But asking people who buy stereos for enjoyment (not to be lab tools) to train themselves up to the level of the producer just so that what seems accurate to them is closer to the theoretical ideal strikes me as ludicrous. I agree that they should train themselves before making declarations regarding absolute accuracy, but don't confuse the two. If they are untrained and have perceptions that accuracy is better attained through technical inaccuracy, then technical inaccuracy is the right answer for them as it better achieves the end goal of the perception of accuracy for them.
As people with an interest in the science (ie making this predictable) we need to be able to distinguish the two, and I'm one the biggest fans and earliest adopters of Toole/Olive/Queen/Gressingers/Blaurts/name drop/name drop etc etc work but that in no way invalidates the perceptions of the untrained listener to them.
Quote: Originally Posted by gedlee
I just have to assume that the producers of the music that I listen to are smart enough to know about the circle of confusion and attempt to get to some standard that will reproduce the best possible reconstruction that I can get. Now we all know that not all of them do, but many do.
I think it's wishfull thinking to think their tastes don't affect the rendition. For example, many 70s DG classical recordings have a horrible 1st row perspective to the tonal balance. I'm sure it was "realism" to the conductor but ghastly to the me as an audience member. Multi-mic studio recordings are artificial and recording is tweaked to taste.
Different speaker design types even betray an obvious bias towards main floor (stand mount) or balcony (line source).
For me personally, I strive to train myself to be able to detect accuracy and true differences (I have CD training material and headphones at my bed side: Critical Listening Skills for Audio Professionals by F Alton Everest), but I would never presume to be authoritarian and to expect this to be ubiquitous, or to make light of other's perceptions if they don't pursue this path
I don't think the premise is false, not at all. I think that we are getting closer and closer to universal better sound in both the home and the studio. I see the goal as possible as long as people are indeed interested in pursuing it. I see nothing wrong with a pure pursuit of the "art" and see no reason why this should interfere with the "illusion".
Taken to an extreme, it is clear that taking a position that any modification to the playback that makes "me" feel it sound better is not "Hi-Fi" and does not show much respect for the "art" being promoted by the artist.
The other extreme, that nothing in the playback chain should be anything but accurate, also does not seem to be supportable. Some things do make sense like a HF roll-off (although that clearly is not universally accepted.)
My point is that it takes more than just "It sounds better to me!" to be a valid support for altering the art. Sure you can alter the art to your hearts content, but at some point you have to understand that this is not showing respect for that art.
I respect the art more than I respect audiophiles whims at changing it.
Taken to an extreme, it is clear that taking a position that any modification to the playback that makes "me" feel it sound better is not "Hi-Fi" and does not show much respect for the "art" being promoted by the artist.
The other extreme, that nothing in the playback chain should be anything but accurate, also does not seem to be supportable. Some things do make sense like a HF roll-off (although that clearly is not universally accepted.)
My point is that it takes more than just "It sounds better to me!" to be a valid support for altering the art. Sure you can alter the art to your hearts content, but at some point you have to understand that this is not showing respect for that art.
I respect the art more than I respect audiophiles whims at changing it.
Would anyone have noticed a possible improvement like the above without experiment? Like you said: as long as you know that is what you are doing.
For instance, earlier in the other thread I said I'm experimenting with ambient back channels. That could be seen as deviating from the reproduction Art. But after reading a lot of old material on Lede studio design, which at some point included Haas Kickers I had to experience that myself. Lacking the space in my room to solve that with physical means I do it with indirect firing speakers in the back which are delayed and attenuated just as a Haas kicker would have sounded in the Studio. The difference being that I can chose the signal that's coming out of them.
On gearslutz there have been multiple discussions on the Haas kicker's effect.
Basically it was noted as a pleasing way to terminate the reflection free zone.
It seems to be abandoned in Studio use as it made most recordings sound better, more pleasing. I can get that something like this would detract from the objectives in a Studio. But for a listening room, wouldn't that be the point? Making it sound better? I just had to try, and can turn it off at any time with just one button. Not that I did turn it off yet.
Even Linkwitz experimented with something very similar: Surround stereo system
For me, I have no problem with trying something. But with measurements to guide me as to know what I'm listening to. Basically as a learning process. For instance I have read that Toole did similar tests with ambient signals and did not find it an improvement. Should I just shut up and listen to him? Or is it OK to make up my own mind. I try to avoid/minimise early reflections while Toole finds them useful. Why not try and educate yourself.
As long as you know what you add or detract, and learn from it. Isn't that a way forward? I know I have learned quite a few things that make much more sense to me now.
Just an example of why I personally think it is ok to experiment. Not planning on making anyone's Art sound worse. I still think the better your speakers measure, the more they are enjoyable with all kinds of music.
Yet I read a lot of people's opinion that seem to think/point otherwise. Building speakers to suit a certain style of music.
Now that doesn't make much sense to me personally. But I won't stop them from pursuing their dreams.
Bypassable tone controls should not be looked as bad. As long as you can go back to flat. Take two simple examples of the 3-5KHz region which changes perceived depth and 8KHz(ish) that changes perceived image height. For orchestral recordings you might want a different perspective and purely IMO being able to adjust for that is no sin.
However building a system with those deviations build in and no way of going back to a flat curve to me is veering away from hifi and into 'musical instrument'.
For multi-track pan potted music then I assume the sound engineer will have done a lot of that frequency tweaking to build up their sound stage during the mix. But I would need that confirmed.
So from my point of view, I would dearly love to have a cello palette at my disposal and to enhance my enjoyment, but I would never try and persuade someone that my frequency deviations were 'better' or 'higher fidelity'. As for respect to the "art" that is a tricky question. It's not respecting the art to play it on a bluetooth squawk tube, but people do, so I guess the artist is not that fussed as long as they get paid.
However building a system with those deviations build in and no way of going back to a flat curve to me is veering away from hifi and into 'musical instrument'.
For multi-track pan potted music then I assume the sound engineer will have done a lot of that frequency tweaking to build up their sound stage during the mix. But I would need that confirmed.
So from my point of view, I would dearly love to have a cello palette at my disposal and to enhance my enjoyment, but I would never try and persuade someone that my frequency deviations were 'better' or 'higher fidelity'. As for respect to the "art" that is a tricky question. It's not respecting the art to play it on a bluetooth squawk tube, but people do, so I guess the artist is not that fussed as long as they get paid.
I think that this is quite a reasonable position. And yes, I have found (along with Toole and Olive) that the better your speakers measure the better they sound. The question here is "When do you deviate from that?" Experimenting is fine to answer that question and when someone has a solid concrete example of when to make that deviation I will be interested in trying it. But all too often there isn't much foundation for the claim to deviance and all that is given is the catch all "Well it sounds good to me.", which, OK, it ends all conversation, as it was meant to do, but I don't find it a very respectable position.
Why care?, because I think a lot of unknowledgeable people come here looking for guidance and they need to get both sides. No one can claim that the subjective side isn't well represented around here - chezz' its dominate in that regard. So I think that the purely objective point-of-view needs to be represented even though it is clearly in the minority.
Altering art is art. Some of build our systems to enjoy the sound of the music knowing our playback is better (we hear more than they did in the booth) than what gear was used to create the recording. Knowing that and most recordings we may love and cherish are flawed I find the "sounds great" to me route is a surer path to happiness than "hi fidelity" - high fidelity to what?
Even more, I made a commitment to DIY some once. Ask me if it was worth it.for the speakers nobody buys.
I've read this a half dozen times and still don't know what it means, specifically the part "towards the HFs." Towards the what?
i.e., tilting down as you go up in frequency. Take for example the frequency response of Wesayso's towers and Mitchba's speaker - they both arrived at same tilt which sounds more pleasing.
I could talk about a lot of things tonight, like bafflestep and power loss at crossover with even order filters.
But let's talk about the legendary BBC dip. AKA rolloff at higher frequencies.
In fact Doctor Earl Geddes explained it in a simple way that stuck in my mind and ensured I wasted no more time on it. High dispersion tweeters, like domes, need a BBC dip. Lower dispersion tweeters like cone types appreciate a little lift.
It's to do with power response. Which is what you hear in the far field or off axis of a speaker. So with my cone tweeters, I lift them. With domes, I drop them. Seemples.
But let's talk about the legendary BBC dip. AKA rolloff at higher frequencies.
In fact Doctor Earl Geddes explained it in a simple way that stuck in my mind and ensured I wasted no more time on it. High dispersion tweeters, like domes, need a BBC dip. Lower dispersion tweeters like cone types appreciate a little lift.
It's to do with power response. Which is what you hear in the far field or off axis of a speaker. So with my cone tweeters, I lift them. With domes, I drop them. Seemples.
Attachments
I thank you for that. It is an important topic, and deserves its own thread.
As has been said before, what are we being faithful to?
This may be a strawman argument. I don't believe I know anyone who thinks that. Most people do have a priority of "sounds good" but no one I've met - inside or outside audio circles - thinks that "anything they do that makes things sound better" is fine. Certainly many don't care much about accuracy, but they do understand a certain fidelity. It's not anything goes.
This is a real effect that you can verify for yourself. You might also read the paper I linked in the thread and some of the references in that paper. It's well documented. More important with dialog than music, but real none the less.
You did, but chose to ignore it. I even posted some examples.
Who really can answer that? It's not a relevant question.
No one here believes you to be so naive as you pretend. You are not that obtuse. We don't believe it. Even you know that the reproduction of art, visual or auditory, or terpsichorean is always an interpretation of the original because no reproduction is perfect and compromises are always made. Often by the artists themselves. I've dealt with 100s who don't want a strict reproduction of their work. They usually want something "better", even if they don't know what that "better" is. Anyone who works in the audio or fine art mastering business deals with that on a daily basis. Reproductions that closely match the original are usually rejected by the artist. We learn that lesson quickly thru lost time and money. We do still try, tho.
At least in cinema there are some standards. They aren't followed as well as they should be, but they do exist. In music recording - not so much. It's all over the map. Which leads to many throwing up their hands and declaring "All I can do is objectively reproduce the signal". That's as much a cop-out as saying "None of it matters, I'll do what I want".
The reproduction of art is difficult and imperfect. Science and good engineering help immensely in getting us close to real. But there still remain imperfections to be dealt with. Those require some creativity and ingenuity to get us closer to a feeling of reality. The illusion IS the point.
I too can say a lot about this. Being an aspiring med student I spend as much time as I can with M.D.'s and I am reminded constantly of the balance between art and science.
But I'll just give this one idea.
Go to any exhibit of fine art at a 'proper' gallery, and lets say you are looking at an original Van Gogh. The lighting they use is unbiased, so you get the true and accurate image in front of you. But what if Vince was sitting down in the late hours in a candlelit room and had been drinking wine and his vision was a bit blurry when he painted that picture? But that's when the art came out, that was the moment of inspiration, captured on canvas.
Just a thought. Like many sciences there are quantitative and qualitative aspects. Especially when dealing with human perception.
Don't get me wrong here, I am very much one for the quantitative information. It appears to me this science is young and it would be very good to continue perusing a strong scientific foundation of information as Dr. Geddes is arguing for. It seems like a forum like this is an amazing opportunity for knowledge to be shared. I am really glad to be a part of this.
Shawn
But I'll just give this one idea.
Go to any exhibit of fine art at a 'proper' gallery, and lets say you are looking at an original Van Gogh. The lighting they use is unbiased, so you get the true and accurate image in front of you. But what if Vince was sitting down in the late hours in a candlelit room and had been drinking wine and his vision was a bit blurry when he painted that picture? But that's when the art came out, that was the moment of inspiration, captured on canvas.
Just a thought. Like many sciences there are quantitative and qualitative aspects. Especially when dealing with human perception.
Don't get me wrong here, I am very much one for the quantitative information. It appears to me this science is young and it would be very good to continue perusing a strong scientific foundation of information as Dr. Geddes is arguing for. It seems like a forum like this is an amazing opportunity for knowledge to be shared. I am really glad to be a part of this.
Shawn
Last edited:
That would be the legendary dip that wasn't in the specifications and was a production anomaly?
Another question that might fit this discussion: what about time coherency?
If we want to be true to the recording, does that also imply time coherent behaviour? Currently we have much more access to tools that can supply us with corrections that can give us time coherency in our replay devises. Personally I believe it is a way forward to hear the signal as recorded. If a drum goes BAM while recording, shouldn't we reproduce BAM instead of KaBoem?
I take it many of studio monitors are of the non time coherent designs. Yet that seems to change. Slowly but surely professional monitors with active DSP are surfacing that have a time coherent output. Making the impulse and STEP look more like the signal that was put into the speaker. It has been one of my personal goals as well. For an example of time corrected monitors you could look at Kii, but more examples in commercial speakers are surfacing as well like GrimAudio.
I also expect the new B&O flagship to be on the same path, though I haven't seen an impulse yet. One of the designers Geoff Martin responded to my question on time behaviour:
FIR correction and tools like RePhase, Acourate, Audiolense, DRC-FIR among others give us the ability to try this. Yet I see a lot of doubt on the merits on this forum. Is this really audible?
I am convinced it is, by testing it. Should I put on a flame suit now?
If we want to be true to the recording, does that also imply time coherent behaviour? Currently we have much more access to tools that can supply us with corrections that can give us time coherency in our replay devises. Personally I believe it is a way forward to hear the signal as recorded. If a drum goes BAM while recording, shouldn't we reproduce BAM instead of KaBoem?
I take it many of studio monitors are of the non time coherent designs. Yet that seems to change. Slowly but surely professional monitors with active DSP are surfacing that have a time coherent output. Making the impulse and STEP look more like the signal that was put into the speaker. It has been one of my personal goals as well. For an example of time corrected monitors you could look at Kii, but more examples in commercial speakers are surfacing as well like GrimAudio.
I also expect the new B&O flagship to be on the same path, though I haven't seen an impulse yet. One of the designers Geoff Martin responded to my question on time behaviour:
Original response hereGeoff Martin said:
FIR correction and tools like RePhase, Acourate, Audiolense, DRC-FIR among others give us the ability to try this. Yet I see a lot of doubt on the merits on this forum. Is this really audible?
I am convinced it is, by testing it. Should I put on a flame suit now?
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.