Listener fatigue

Discussion in 'Sound Science' started by SilverEars, Nov 3, 2017.
  1. bigshot
    The wild card is the room. If you can tame that with EQ, any decent speakers can outperform headphones.
  2. 71 dB
    Difficulty to locate bass is because of early reflections and reverberation (room modes actually) messing up with the ITD information. Yes, transition is gradual and typical cross-feeder operate that way. Yes, room acoustics affects the sound significantly, which by the way is a problem since recordings are mixed in studios where the acoustics are usually much more controlled and dead. It's comical how some people listen to their "expensive as hell" - speakers with snake oil cables inside a living room with acoustics of an echoic chamber.

    Maybe I wasn't clear enough about it, but I assumed gradual transition all the time. However, the 3 dB separation holds up to quite high frequencies, like 500 Hz or so, before the shadow effect of human head kicks in stronger. In acoustics the "big changes" happens when the physical size of an object is roughly the same as wavelength. In the case of shadow effect of human head, it happen between 800-1600 Hz. Human head is tiny compared to wavelength both at 20 Hz and 200 Hz, so the change in shadow effect is insignificant.

    It doesn't matter when the reduction of separation is done as long as it is done. If it isn't done while mixing music (very rarely is), I do it myself using crossfeed. My allowance doesn't matter. My hopes, opinions and advices don't change anything. They mix the music the way they do. However, I can use these principles when I make my own music. I try to mix it so that it works on headphones and speaker as it is. I call it omnistereophonic sound. Yes, that's my own definition.

    I don't think the limitations of separation limits much creativity, because there is so much more you can do. People are just used to do things in a certain way and are reluctant to change the habits. I question things and I am one of those guys who steps in a room an says "Why do you do thing X this way and not another way?" For me answer "Well, it's been done always this way." is not a good answer. It is an excuse. Anyway, It's just frequencies below 1 kHz or so and above that you have a lot of freedom with separation.

    Separation is overrated. In the early days of stereo it was considered a "virtue", but in real life we don't need much of it, at least below 1 kHz. People are just still stuck on that mentality. What we hear in everyday life around us is more mono than people think. Our hearing is good at detecting tiny spatial cues and that's why headphone listening with huge separation overloads our heads.

    You can create spatial effects with delay alone at low frequences where you need to limit separation. If you delay a bass tone by ~600 µs on right channel, and listen to it on speakers, you should hear the sound located near left speaker. Even when volume of right channel is the same! If you attenuate right channel by 3 dB, the effect works even better. In real life you have high frequency content on instrument sounds, transients when bass is plucked etc, that those have greater separation and the localization is much easier. So it is about using spatial cues cleverly.

    That's why omnistereophonic sound is difficult, but also possible, because such recordings exists (those "infamous" 2 %). It is about having a clever compromise between stereophony that works on speakers and stereophony that works with headphones. The main principles seem to be:

    - Limit separation at low frequences (3 dB at bass, increasing to about 10 dB at 1 kHz)
    - Use large ITD instead of large ILD at low frequencies to create spatial effects, if needed.
    - Use large ILD at high frequencies (above 1600 Hz) to keep the soundstage wide with speakers.

    Mind you, I am in the beginning of studying this subject, and I don't claim mastery on this issue. I like a little separation on bass more than total mono bass with headphones. it gives a sensation of a room around me, that something is "happening" around me acoustically.

    [A] Of course not. This is acoustics 101 stuff.
    No, I am aware of that and I never said it wasn't. I was teached this stuff in the university over 20 years ago, in 1992 maybe...
    [C] Of course not, but how the knowledge is applied is another story. I was 41 when I discovered crossfeed! Before that I was "spatially ignorant."
    [D] I believe there might be differing opinions in little details, but the large picture should be agreed upon.

    No, I did not answer "yes" to all those questions. I'm not as arrogant as you think.

    I'm not a dictator telling how all music on Earth must be remixed and crossfed to my liking. I'm figuring out myself what is the smartest way to do things and I apply that understanding in my music making as a hobby. I'm for neutral audio and this is about finding out how we get into "neutral spatiality." Omnistereophonic recordings would mean no need for crossfeeders. The spatiality would be much better controlled in all listening conditions because the spatial information in the recording is smart.

    HRTFs are measured in anechoic chamber, at least they were measured that way in the acoustics lab I worked in. My HRTFs have been measured in the past, but I don't have them because they are "intellectual property" of Nokia. What happens in room at low frequencies is up to room modes and the result is that spatial information is mostly lost. You need higher frequencies for spatial cues. For that reason it doesn't matter much what the separation at bass is when we listen to speakers, but it matters A LOT with headphones, so it should be optimazed for phones. At the lowest frequences monophonic bass helps speakers, because the sound sources add up amplitude-wise meaning a little bit more sensitivity (1-2 dB).

    Now I have to go. I continue later...
  3. pinnahertz
    Unfortunately, it does matter. Recordings are made to convey creative intent as universally as possible. Building in crossfeed would assume only headphone listening. Built in crossfeed is not removable. However, it can be added during playback, and there are many methods. Those playing back the recording are aware of their playback methods, and can apply crossfeed if they desire and affect no one else but them.
    The implication here is that you are right and every engineer mixing for stereo on speakers without crossfeed is wrong. I'm glad you realize your opinions won't change how music is mixed.
    Ever seen a pan pot? I assume you must have at some point. Now, ever seen a pan pot that varies ITD vs frequency and ILD vs frequency? Of course not, they don't exist. Because that wouldn't work as universally as the simple ILD pan pot. I've advocated the idea above for years, actually proposed a design for this in the 1980s, it's even entirely possible now that mixing is fully done in software, but the fact is, it's impractical because it's not a universal solution for every listener.
    The problem with the above statements is they are based on your opinion of what's right. The industry is already mixing using a compromise it feels adequate.

    I'm going to quote you some other statistics. 80% of all existing recorded stereo music doesn't benefit from rudimentary cross-feed. 40% of that group is actually damaged by rudimentary cross-feed. 15% of all recorded music is at least somewhat benefitted by rudimentary cross-feed, and the remaining 5% is mono. Those figures are MY opinion with the same verification as yours (none).
    Nobody would know any of that from your posts. If that wasn't your intention, you may want to rethink how you say things.
    But you are saying that 92% of all mixes are "wrong", you've invented statistics and terminology....etc., etc....

    Nobody's arguing HRTF or how hearing works, you can stop explaining the obvious any time now. The argument is about your cross-feed mandate for 92% of all recordings ever made, and that your cross-feed should employed in mixing. Wrong for both.
    Please don't continue.
  4. bigshot
    All those blue stripes sure do look pretty on the page!
  5. 71 dB
    What human hearing does is a good "pre-correction" target if the rest of the chain does next to nothing (headphones without cross-feed). A studio set-up doesn't reveal problems such as excessive stereo separation at low frequencies. Of the 98 % a lot are "mild" cases which I don't "hate", but I still think can be improved with crossfeed. Ping pong recordings are in the "hate" category.

    I believe that excessive stereo separation is one of the least understood aspects of audio and that a lot improvents can be done on that front. I was also "deaf" to it because of speakers hiding problem it with acoustic crossfeed and thinking headphones sound unnatural/tiring because they are headphones. Headphone listening has become more popular thanks to "mobile music devices", so maybe people who mix music become wiser day by day and we go toward omnistereophonic sound.

    Yes, crossfeed doesn't improve every recording. For the 2 % it is harmful and for the mildest cases of the 98 % it is a matter of taste and even daily mood. However, in my opinion crossfeed benefits without doubt easily over 50 % of recordings, maybe 2/3 of all. I have said several times that off-switch is needed on a crossfeeder, because every now and then you need it!

    What then if "everyone else" was wrong? That's only human. I was wrong/ignorant about this myself up until 2012.

    I have a strong faith in what I say based on the rather intensive and impassioned study of the subject for half a decade. Is there something I can learn? Absolutely! My understanding isn't complete at all (for example I haven't studied the "transitional octave" much), but I sense it is quite advanced compared to the general understanding of the subject.

    Habits can be rational or "cultural". Most of people do "cultural" headphone listening (you but your Beats Audio cans to your head and blast out bassy EDM or hip hop tracks without thinking much about it). I try to do rational headphone listening which try to take into account how our hearing works to make the "sonic flow" to my mind as rational as possible. The problem is that "cultural" habits are mistaken as rational when they rarely are.

    The 2 % / 98 % is an estimate, which I don't have reasons to disagree. Crossfeed doesn't improve everything, but most of the time it imo does.
  6. 71 dB
    Some people have more to say than just comment on blue stripes! :L3000:
  7. 71 dB
    If you can make an omnistereophonic mix then that's what I suggest you to do. If you can't, then mix for speakers. That's what I think about this.

    Of of course pan pots should be ITD+ILD based in the 21st century. It's not 1958 anymore, and simple ILD panning is naivistic. How is it universal even for speakers? Many hard panned ping pong recordings sound plain silly on speakers and remixing such recordings using the principles of ITD+ILD results in much better sound image on speakers (and headphones too!)

    I have processed Dave Brubeck's Jazz Impressions of Eurasia for speakers (crossfed for speakers) to have reasonable soundstage and it was a clear improvement even if I say it myself. Hard panned stereo makes the sound come from speakers only (keyhole audio), but if both speakers participate on every sound, the soundstage gets spread over the area where the speakers are, a bit behind the loudspeakers. ILD-only panning is naivistic and hard panned ping pong stereo is silly. Nothing "universal" about those. It's 2017 and we have powerful computers to do panning much better. Omnistereophic recordings are universal.

    Except recordings are mixed differently! Some have huge separation at bass, others hardly any. Some recordings need brutal crossfeed, others hardly any. There is no standard. It's up to who is doing it. The principles of omnistereophonic stereo would set limitations and in that way create some kind of standard so that recordings would have more coherent stereophony.

    Even your numbers suggest that people should have crossfeeders for the 15 % of their recordings.
    Mono recordings are of course not included in the 2 % / 98 % estimation.

    Too late...
  8. bigshot
    I guess if I read all those words, I might have more to say. I keep trying and I get one or two stripes in and I start thinking of better things I could be doing than wade through it all. It might be a good idea to put your best thoughts up front. I tend to pay attention to properly constructed paragraphs with a statement, supporting arguments and summation. I understand what people are trying to say better when they organize their thoughts for the convenience of the reader. If it's just for your own amusement, that's fine. I can always just admire the pretty blue stripes.
  9. 71 dB
    My original posts are more in the form you suggest, but then pinnahertz reads them and interprets them as me arrogantly stating my opinions as facts and it all escalates into pretty blue stripes for you to admire...
  10. bigshot
    One thing I've learned about posting in internet forums is that it's sometimes better to reply addressing the hundreds of lurkers reading than it is to reply solely for the person you're replying to. I know for myself that I enjoy reading things that are organized to make it clear and concise better than I do trying to sort out line by line and footnote replies. If something looks like a lot of work to parse, I'll read the first line or two and if there isn't anything to grab onto there, I'll skip over the whole thing to the next person's post. I imagine other people do the same. When I reply, I try to open up the discussion by replying in a broader sense with new information. I avoid closing in the discussion on details that are probably irrelevant to the larger discussion anyway.
  11. pinnahertz
    1. I completely disagree with your statistics. Please prove them.
    2. I believe your are wrong, stereo separation is good and well understood, and that we make choices to use it as required for our end goal. I believe we have all the terms we need. Your new term "omnistereophonic" is misapplied if you consider that the root greek word behind "stereo" means "solid, and 3 dimensional". Your omnistereophonic concept is not that at all.
    3. I completely disagree with your statistics. Please prove them.
    4. Proclaiming that "everyone else is wrong", without a shred of proof (and I'm not talking about discussing how hearing or HRTF works, I'm talking about statistical proof of preference), is the very height of arrogance. And now you've labeled everyone else as "ignorant" and "deaf" too. How do you expect to get people to listen to you when you do that?
    5. Yes, we all recognize that I'm sure. You also have zero respect for the opinion of anyone else who has worked on the same exact thing for several decades, and still works in the industry. Again, "arrogance" would seem to be the descriptor.
    6. You understand the subject, just not the practical application of it.
    7. I completely disagree with your statistics. Please prove them. (estimate or not).
    8. How's that for minimizing the blue lines?
  12. pinnahertz
    1. It's your term, and you haven't even clearly defined it. It's not going to happen unless you do the mix.

    2. But they aren't. Perhaps you should apply your talents to learning why. The ability to do an ILD/ITD pan pot has existed for at least 2 decades. It is not done at all. Figure out why, and you'll actually learn something.

    3. I'm afraid you're going to need to dig in and do your own research. Rather than try to disprove it, figure out why it is universal, and works in the industry. You don't listen to me anyway.

    4. What you call "silly" may just have been the creators intent. Now you're just being a critic. Opinions..opinions..opinions.

    5. You've cherry-picked your example. And there's no such thing as "omnistereophonic" except in your world.

    6. You can't standardize art! You have preferences, so do others. Yours are not global standards!

    7. What principles? Now your coined term is to be a global standard? Why do you think the word "delusional" keeps popping into my head?

    8. My numbers are complete fabrications. They are made up. They mean nothing, not even to me. Just like yours. Do you get the point?

    • You keep quoting fictional statistics
    • You keep stating opinion as fact
    • You make up a term and now want it to become an international standard
    • You are "right" and everyone else is "deaf" or "ignorant" or "wrong"

    What does all of that say... really?
  13. pinnahertz
    Um...I'm not the one stating your opinions as fact, dude.

    I rather like the blue line format. It likes up comment with the statement that stimulated it. But I'll concede to the needs of others if it makes their life easier.
  14. bigshot
    For stereo music playback on my main 5.1 speaker system I generally use the Yamaha Stereo to 7.1 DSP. It does a great job of spreading the sound around the room in a sophisticated manner. It's more than just pushing left front to left rear and right front to right rear. It creates a center channel that bridges the phantom center really well, while it creates a strong focused soundstage. The rear ambience fills in the room to keep it live without drawing attention to itself. It's probably pretty close to the speaker equivalent of cross feed.
  15. 71 dB
    2. Omni means "all / every". It states that stereophony is of the kind that it works well on both speakers and headphones (without crossfeed). What end goal is achieved by large ILD at bass? With speakers bass will be almost mono anyway in your ears, no matter what the separation is on the recording and our ears wouldn't like large ILD anyway. I think many live in the 1958 stereo utopia of huge separation, but maybe it's changing finally?

    8. Good job!

Share This Page