That is an amazingly good result, I'm impressed. I'll see if I can post same, but perhaps not in this thread, it might not be such a good idea.
But the issue, to me, is not if its audible, of course it is. My point is that it is easily correctable in the original mix by the engineers. Do they all do this correction, I'm sure they don't. Do good ones make this correction, I'm equally sure that they do.
It's critical to understand this effect, Toole talked about it so we should all understand it. The question is "Who should 'correct' this problem?" To me it belongs in the recording studio, not in my playback setup.
Yes, that is the important question. And the answer isn't easy.
There is not a wealth of recordings in my music library where the same voice or instrument is heard in more than one position. So the effect is not likely to jump out at you, as most music recordings don't move things around much.
The recordings I have that do move things around, have an uneven tonal balance. That unevenness bothered me, so I sought to correct it, or at least understand it.
Interestingly, in the vast majority of recordings where the sounds stay put, using the shuffler did not adversely affect the tonal balance, as long as the slight brightening in the center was taken into account. The shuffler did not ruin the tonal balance of any recording I put thru it.
We did find all versions of the shuffler to have some side effects, but tonality was not a major one. It didn't alter the overall mix enough to be noticeable.
I.E., it fixed one problem without creating too many others. The trade off is up to the individual listener.
This is a simple amplitude pan done in software. It should be smooth and even. Perhaps check it on headphones?
Listened to both count files, two counts on the left, two at center, two at right, center, etc.. the raw file sounded a bit lower in level with slight change in tone balance, the convolved file sounded a bit louder than the sides with a better tone balance match.
Not on headphones. Am I hearing the correct pan?
I'm not fixing the effect it has on the phantom centre. In fact my EQ is close to a B&K curve or the Toole/Olive proposals. I'm only fixing the slight tonal imbalance I do notice in the side panned content compared to that centre.
Am I fixing the recording or fixing the positional differences in my room compared to what was heard in the studio?
All I know is side panned content sounds more real (more body to panned background vocals) and more consistent with the phantom voices. It doesn't take much. And it works for the majority of recordings I have tried as a sample set, consisting of a lot of different genres.
Introducing a diffuse, lateral arriving late return (15 to 20 ms after the initial impulse), lower in SPL and band passed, (Haas Kicker) makes it sound more 3D or holographic like and immersive*.
Not using the Haas kicker but instead the cross talk cancelation did get me the same 3D like imaging, but only in a small sweet spot and made me cramp up. Giving me head aches over a longer course.
Everywhere in the room except in the sweet spot centre line you get a different comb filter effect in each ear, in the sweet spot you get the same comb filter effect (or dips at the same place) in each ear. Yes, you're going to notice that!
Unless (early or late) reflections/diffractions fill in the dips within the Haas limit.
You're still only arguing about it, such a waste of time. Listening to it makes more sense to me.
Look at your IR for the first 20 ms, is it largely reflection free? If it isn't you probably won't be bothered by this effect.
Very early reflections would influence imaging in a bad way. Once you get past 7 ms they are less severe for the imaging and staging, but might make the difference between the "they are here" or "you are there" type of discussions, another long debate
.(*) immersive only if there's a longer reverb like you get in bigger rooms or spaces.
You don't get a real reverb in a small room. I add a reverb tail to the (L-R) and (R-L) ambient signals (based on Lexicon's Random Hall) as it helps out to hide my small room dimensions. It's down in SPL by over ~30 dB and more but still helps to hide my room boundaries.
The lovely part is I can shut off each part separately. Not that I ever do, except in testing.
Thanks for the suggestion. I tried it and found Toole.
Looks like the ancient findings using bowling balls and two mics and Bauer correction. Nobody doubts bowling balls have real problems with inter-aural differences.
People turn to Toole for his very competent tests on people. No tests of human hearing?
Maybe we really need a thread about correcting the resonance in the external ear canal. Nobody doubts the resonance in every human ear is very substantial and got to be a bit different between a persons own ears too! El Greco all over again.
I don't know if we have a "problem" here or not.
Thanks.
Ben
Last edited:
Toole would not agree on that
(*if* speakers with coherent off axis response are used, and preferably with only lateral reflections)
Toole would not agree on that
(*if* speakers with coherent off axis response are used, and preferably with only lateral reflections)
I know, that's why I asked Dr. Geddes about his preference compared to Toole's point of view right here in this thread:
The reason I think Toole prefers some early reflections is also helping to hide/fill in those dips caused by the cross talk in the first place. His 'early' reflections are not that defined though. I bet he means later than 5 to 7 ms. Especially if I look at the listening room at Harmon:
An externally hosted image should be here but it was not working when we last tested it.
I've come to my own conclusions and add later reflections (later than 15 ms) which still fall within the Haas limit. Even though I've read Toole, does not mean I actually agree on all of his points.
I trusted in Geddes's point of view and do not regret that. Most of my library also consists of studio recordings. Though with an adjustable Haas Kicker I can be flexible.
The 20 ms reflection free zone was more common in Studio engineering. So I spent a lot of time on forums like GearrSlutz to see what they thought was optimal. The Haas Kicker has been abandoned in Studio's long ago but for a listening room it isn't a bad idea at all. The old LEDE concept has brought me a lot of inspiration.
Last edited:
It should be, I think. But what about the tonality?
I"m going to make you a new counting file. I'll post it here.
What people call "optical illusions" are visual images that people see wrong.
The phenomenon discussed in this thread might possibly be heard under special test circumstances. That would be like "optical illusions" that can be seen under special circumstances. Both (if and when they might possibly occur) are wrong perceptions because both lead to perceptions that are incorrect.
In the case of "optical illusions", your visual processing is using its usual heuristics, but still is led astray.
If the thing discussed here is taking place, the sound is not changed but - due to having a head between your ears - you get the tonal colour wrong. So that is an "aural illusion" and is precisely analogous to a visual error.
If you follow this line of reasoning, you'll notice I've turned the discussion on its head. The phenomenon experience* is an error, not the recording.
Ben
*if it is true and occurs anywhere but under special test circumstances
The phenomenon discussed in this thread might possibly be heard under special test circumstances. That would be like "optical illusions" that can be seen under special circumstances. Both (if and when they might possibly occur) are wrong perceptions because both lead to perceptions that are incorrect.
In the case of "optical illusions", your visual processing is using its usual heuristics, but still is led astray.
If the thing discussed here is taking place, the sound is not changed but - due to having a head between your ears - you get the tonal colour wrong. So that is an "aural illusion" and is precisely analogous to a visual error.
If you follow this line of reasoning, you'll notice I've turned the discussion on its head. The phenomenon experience* is an error, not the recording.
Ben
*if it is true and occurs anywhere but under special test circumstances
Last edited:
In that case the only real thing I should get out in my room is sound coming from the 2 spots where my speakers are? Because all of the rest of it, even the phantom centre IS an illusion.
My goal is to make that illusion stereo can bring as strong as possible. Not to hear two different sound sources
.It is the magic of stereo sound that your hearing system takes the information and creates pretty good soundscapes.
Somehow in your mind you can close your eyes and hear the back wall of your music room disappear and a large orchestra play in a hall behind your house even though many cues tell you unambiguously your room is X by Y feet.
The auditory cues you are taking in conflict and confound one another but still the image works. Like in the movies when actors in a corn field have the reverberation of a room, but we take it as meaningful sound. Or when the producer cranks up the oboe at the right moments so you can hear it like Beethoven meant you to if it were live in a hall.
Hope that starts your education in psychoacoustics.
Ben
Last edited:
Oh boy.... you've thought me so much in this single post...
Do you read my posts with your eyes closed too? I take as much of my room's queues away as I can to maximise the stereo illusion in my room.
All you seem to do around here is to make insinuations and snide remarks. Why were you following this thread exactly?
Last edited:
A performing violin player sitting on a chair right in front of you hasn't got any natural build in dual transducers, thats a reproducing method created by humans long time ago and why should such interface just be absolut perfect and not add some small distortion products here and there.
And for physical theory HRTF as you call bowling ball with microphones curves still looks distorted using known psychoacoustics or whatever smoothing method, look legend in below graphs to see smoothing schemes.
Attachments
It's a bit difficult for me to pinpoint tonality because the counts are short. Normally testing tonal changes would be better using longer sounds. For example, golden ears training content. But I could try.
The confusing thing is even in real life, if a device shifts to the side, the tone balance would change, wouldn't it? So we never know what is totally current.
Sent from my iPhone using Tapatalk
Last edited:
Not as much as it changes from a single speaker to two speakers playing the same content.
I've made a new version of the counting file for you. There is a normal version and a shuffled version using my phase only shuffler V2. In the version posted 2 years ago, I had dropped the center channel by 6dB to give an even volume across left, phantom center, right. But after your comments, I ended up dropping the center by 7dB, because it seemed a better balance. 8dB was too much.
Have a listen and let me know. I will look for something with more continuous speech and have a try with that.