I mean, this film made by BBC in 1965 show how big venues recordings just pretend to be like Kraftwerk's (master) pieces of electronic music, and are anything but an effort to capture the sound field of a live acoustic event: it's all man made fiction...
It's an electronic fiction about Solti's and other interpreters of Wagner's work, the acoustic of the live event being anecdotic to it, though is also important as a decorative element.
Last edited:
Then there's this—a plug for $99:
Makes the Smyth Realizer seem like a legit bargain at around $3K. Considering it could replace a very good hi-fi system and room, it would be worth it. I really need to get some time with one of these.
Nx ? Virtual Mix Room over Headphones | Waves
That looks excellent!
I wonder if it simulates the comb filtering in the phantom center though...
Probably!
I've given the phase shuffler a fair shake. Invested quite some time to see what it does and listen to it. I also tried to optimize it for my specific room etc.
At a first listen it really had me. Because it opened up the soundstage in the phantom center. As the days passed I started to get a bit irritated by some specific tonal balance issues. Issues that I didn't have before.
No matter what I did I could not get a handle on those issues. I couldn't EQ it out in other words. I tried all kinds of phase deviations but nothing I did got that tonal balance quite right. After spending even more time on it I concluded that it really does do something right, but also something wrong.
To me it seems it clears up the comb filter problem on one side (avoids the deep dips) but it makes it worse on the other side. But strangely enough I don't hear that as separate events.
Only after looking at early waterfall plots of the left and right side (after ~0.27 ms has passed) I noticed the improvement on one side, but more pronounced peaks on the other side.
Anyway, it was enough reason for me to abandon the shuffler as a viable solution to the "Phantom problem".
I even re-visited the cross talk cancelation I tried earlier by adding an inverted signal 0.27 ms after the main signal, this time applied only to the mid channel of a mid/side split stereo signal. This also didn't work out but it did make it clear to me that the shuffler was definitely adding something. There was indeed less strain while listening to the XTC version. Sadly, the cross talk cancelation trick also had some tonal balance issues, I did not try and solve those. But somehow less strain in listening was apparent.
So, for now I'm back to mid/side EQ. I might not have found the absolute settings yet but I can mimic some of the shufflers plus points. Tonal balance is way better with most songs, maybe not quite right yet though. I need more time to evaluate.
I really wished the shuffler would have worked out. The things it did do were worth the trouble to look a bit deeper. It seems after that initial wave front, the added balance at the left and right ear were much less sensitive to sound direction.
Meaning the positional queue seems to come from the two ears hearing the first sound. The sense of distance seems more related to the actual frequency spectrum heard over a longer time frame. Particular that ~2 KHz area does change the perceived distance to the phantom voices.
I hope Pano can start measuring with a dummy head. That should give us more clues, I really think that our head shading plays a large role. I would like to know more about that, where does the shading start etc...
At a first listen it really had me. Because it opened up the soundstage in the phantom center. As the days passed I started to get a bit irritated by some specific tonal balance issues. Issues that I didn't have before.
No matter what I did I could not get a handle on those issues. I couldn't EQ it out in other words. I tried all kinds of phase deviations but nothing I did got that tonal balance quite right. After spending even more time on it I concluded that it really does do something right, but also something wrong.
To me it seems it clears up the comb filter problem on one side (avoids the deep dips) but it makes it worse on the other side. But strangely enough I don't hear that as separate events.
Only after looking at early waterfall plots of the left and right side (after ~0.27 ms has passed) I noticed the improvement on one side, but more pronounced peaks on the other side.
Anyway, it was enough reason for me to abandon the shuffler as a viable solution to the "Phantom problem".
I even re-visited the cross talk cancelation I tried earlier by adding an inverted signal 0.27 ms after the main signal, this time applied only to the mid channel of a mid/side split stereo signal. This also didn't work out but it did make it clear to me that the shuffler was definitely adding something. There was indeed less strain while listening to the XTC version. Sadly, the cross talk cancelation trick also had some tonal balance issues, I did not try and solve those. But somehow less strain in listening was apparent.
So, for now I'm back to mid/side EQ. I might not have found the absolute settings yet but I can mimic some of the shufflers plus points. Tonal balance is way better with most songs, maybe not quite right yet though. I need more time to evaluate.
I really wished the shuffler would have worked out. The things it did do were worth the trouble to look a bit deeper. It seems after that initial wave front, the added balance at the left and right ear were much less sensitive to sound direction.
Meaning the positional queue seems to come from the two ears hearing the first sound. The sense of distance seems more related to the actual frequency spectrum heard over a longer time frame. Particular that ~2 KHz area does change the perceived distance to the phantom voices.
I hope Pano can start measuring with a dummy head. That should give us more clues, I really think that our head shading plays a large role. I would like to know more about that, where does the shading start etc...
Last edited:
Thanks for all you efforts, research and frequent posting of results and ideas in the thread.
And thanks for letting us know why the shuffler ultimately isn't working for you.
We know what the cause of the imbalance is, fixing it isn't easy. If the mid/side EQ works better in your listening position, that's a good way to go.
I hope to be able to do some dual speaker and dummy head measurements soon, but there are big changes come at Rancho-Pano, so I need to act fast. Having a set of measurements in a format that we can all easily access would be handy. Maybe REW. I know the research has been done before (Geddes and others) but I'm not afraid of doing it again.
And thanks for letting us know why the shuffler ultimately isn't working for you.
We know what the cause of the imbalance is, fixing it isn't easy. If the mid/side EQ works better in your listening position, that's a good way to go.
I hope to be able to do some dual speaker and dummy head measurements soon, but there are big changes come at Rancho-Pano, so I need to act fast. Having a set of measurements in a format that we can all easily access would be handy. Maybe REW. I know the research has been done before (Geddes and others) but I'm not afraid of doing it again.
This information is available... DOWNLOAD PAGE
One could average many heads if desired.
Still trying to figure out the file naming scheme though.
Thanks for all you efforts, research and frequent posting of results and ideas in the thread.
And thanks for letting us know why the shuffler ultimately isn't working for you.
We know what the cause of the imbalance is, fixing it isn't easy. If the mid/side EQ works better in your listening position, that's a good way to go.
I hope to be able to do some dual speaker and dummy head measurements soon, but there are big changes come at Rancho-Pano, so I need to act fast. Having a set of measurements in a format that we can all easily access would be handy. Maybe REW. I know the research has been done before (Geddes and others) but I'm not afraid of doing it again.
I'm not done experimenting yet
. So far the shuffler proved there is potential for a solution, but it isn't going to be easy. The (mid) EQ alone does not mimic the shuffler completely, nor the cross talk cancelation. A center speaker makes a lot of sense. Though with the wide library of 2 channel music it would be great to find a working solution for the perceptual differences of the phantom vs the sides. I guess we do process the time differences between direct sound and the comb effect to determine the origin of the sound. I do believe the shuffler was softening that change, though only successfully doing so on one side. I can only agree. I've heard so many changes in all of my personal experiments and it's quite easy to "fall into the trap of change is good". One round of shivers down my spine isn't enough proof anymore
. I've had plenty of those moments, even today with using some mild mid/side EQ only.Just to add, by no means I consider my room or setup of high enough quality to be any kind of reference. Yet I am glad I can more often than not relate what I hear to what I see in the graphs. Listening came first, to not influence my perception. That's why I initially raved about the effect of the shuffler, but in long term could not figure out the tonal balance problems. Measurements and simulations showed me an acceptable cause/explanation (for me anyway) afterwards.
Last edited:
That is quite the library... seems quite interesting after skimming trough the protocol.
The IRCAM library is cool and has tons of info. I've use it a bit. The hard part is finding the head size/shape you need. And then I'm not sure about combining the files to get the comb filtering we want to look at.
FWIW, I worked at IRCAM about 30 years ago. I had nothing to do with the HRTF measurements, tho.
FWIW, I worked at IRCAM about 30 years ago. I had nothing to do with the HRTF measurements, tho.
Sure, I'm open for suggestions and criticisms.
My plan is basically this.
I think that will provide the basic info needed for head shading on the 60 degree stereo triangle looking at both sides of the head separate and combined, as well as a single mic measurement.
My plan is basically this.
Set up two speakers on stands about 2M above the ground and maybe 5M apart.
Set up a child size soccer ball (dummy head) on a stand 2M above the ground at the triangle point between the speakers.
Place 2 measurement mics, one on either side of the ball, tight against it to mimic ear position.
Usie software such as REW to run sweeps on Left, Right and both speakers.
Record on both mics at once to find frequency response and inter-aural delays.
Record both mics summed together playing sweeps on Left, Right and both for comb filtering.
Replace ball with single mic and record Left Right, Both as a reference
I think that will provide the basic info needed for head shading on the 60 degree stereo triangle looking at both sides of the head separate and combined, as well as a single mic measurement.
Alright, maybe throw in a measurement with one speaker straight ahead for reference?
I expect any "head shape" to alter the FR curve as well, so it might come in handy to see what that one would do, with the head and with a single mic, just as a reference.
What speakers do you plan to use? Are you planning on using DSP to get a flat signal at the mic?
I expect any "head shape" to alter the FR curve as well, so it might come in handy to see what that one would do, with the head and with a single mic, just as a reference.
What speakers do you plan to use? Are you planning on using DSP to get a flat signal at the mic?
Pano, if I may... one more request. I'd really like to see the difference between a near field setup and a far field one, just to see if there's a change in the FR curve. Head size will be the same of coarse, but the near field might react different in terms of head shading.
Just trying to get a feel of possible differences in a mixing desk setup and a home environment.
Just trying to get a feel of possible differences in a mixing desk setup and a home environment.
Good question, I'd say whatever is most commonly used at a mixing desk, though that may vary as well. Mike? Where are you . Just trust your own judgment Pano . Indeed the commonly used 60 degree setup.
I'm just curious if close up to monitor like speakers makes a big change compared to 2.5 to 3 meter in these head related measurements.
. Just trust your own judgment Pano . Indeed the commonly used 60 degree setup.I'm just curious if close up to monitor like speakers makes a big change compared to 2.5 to 3 meter in these head related measurements.