I´m following Obsessive Compulsive Audiophile and find his methods to be on another level. ( Hands down )
Is someone else in here using the same procedure?
Is someone else in here using the same procedure?
Heard enough of the video to believe it to be web nonsense, as much as I respect posts from esl 63 elsewhere.
Having invested (wasted?) lots of time working on time/phase alignment with easy-to-do DSP and read the opinions of others trying likewise, I think there's nothing much to be gained and that may be because the notion that you can "correct" phase is inherently illogical. This is just one more mistaken wannabee-engineer theory about human hearing. At the best, you can improve the alignment of speakers at very low crossover frequencies and with a lot of A-B testing, might find A an insignificant bit better than B.
At the risk of being considered provocative, I'd even say long, long experience with polarity showed me I can often not decide which polarity sounds better hooking up a separately located woofer.... and I am talking about careful testing and comparing FR traces using REW to my listening.
Just for starters, who believes the L and R stuff that comes out of the recording studio has any phase "truth"? But let's say you really love listening to mono sources, the sound bounces around, as the video notes, and scrambles the phase in a manner that can't be reverse engineered even if you had a good technique to capture it.
B.
Having invested (wasted?) lots of time working on time/phase alignment with easy-to-do DSP and read the opinions of others trying likewise, I think there's nothing much to be gained and that may be because the notion that you can "correct" phase is inherently illogical. This is just one more mistaken wannabee-engineer theory about human hearing. At the best, you can improve the alignment of speakers at very low crossover frequencies and with a lot of A-B testing, might find A an insignificant bit better than B.
At the risk of being considered provocative, I'd even say long, long experience with polarity showed me I can often not decide which polarity sounds better hooking up a separately located woofer.... and I am talking about careful testing and comparing FR traces using REW to my listening.
Just for starters, who believes the L and R stuff that comes out of the recording studio has any phase "truth"? But let's say you really love listening to mono sources, the sound bounces around, as the video notes, and scrambles the phase in a manner that can't be reverse engineered even if you had a good technique to capture it.
B.
Last edited:
Even if L-R are phase matched, the actual listening position (path difference) could still introduce errors, as the speed of sound is approximately 1 foot/ms. The same would also apply to LF/HF sections.
Well... since our hearing is EXTREMELY sensitive to phase and timing, I know that this more important that you may think at first.
Imagine yourself sitting in a chair, and someone stands behind you and talking. Just moving some inch or centimeters left or right is easy noticed by the listener.
What this guy is doing is basically making sure that what can be done in time resolution is done correctly. I see no negative things by that.
Actually the opposite.
I have access to a demo room with very advanced software and hardware where people much more skilled than me is working with this daily, but their tools and equipment is just not accessible for you and me.... at least not now or in coming 5 years. Since I KNOW myself what can be done and have heard it myself I am curious to find something close. And this is the most serious guy I have seen soo far. If you find it suspicius... fine with me.
I ask the question if someone else in here is doing something similar, without just using DIRAC, and let the SW doing it for you.
I am not against DIRAC, actually it is fantastic! And maybe the best SW there is! But this is DIY.. so we want to do stuff on our own.. don´t we? 😎
Imagine yourself sitting in a chair, and someone stands behind you and talking. Just moving some inch or centimeters left or right is easy noticed by the listener.
What this guy is doing is basically making sure that what can be done in time resolution is done correctly. I see no negative things by that.
Actually the opposite.
I have access to a demo room with very advanced software and hardware where people much more skilled than me is working with this daily, but their tools and equipment is just not accessible for you and me.... at least not now or in coming 5 years. Since I KNOW myself what can be done and have heard it myself I am curious to find something close. And this is the most serious guy I have seen soo far. If you find it suspicius... fine with me.
I ask the question if someone else in here is doing something similar, without just using DIRAC, and let the SW doing it for you.
I am not against DIRAC, actually it is fantastic! And maybe the best SW there is! But this is DIY.. so we want to do stuff on our own.. don´t we? 😎
A good way to proceed with this thread starts with a wiki summary of how sound localization works:
https://en.wikipedia.org/wiki/Sound_localization
For vision and hearing, the brain is really good at taking a whole mess of different and varying cues and producing a stable image (OK, it is only observable as "stable" when you turn your attention to it). With the image formed, the clarinet player will always seem to keep the same seat once your brain has accumulated enough cues to locate the player.
For a fun example to confound the wannabee-engineers, there is evidence that the difference in shape between your own two pinnas constitutes a spectral localization cue.... and helps with the paradox of mid-line localization.
BTW esl 63, apropos your behind the back gedank experiment, did you know that there is localization, albeit crude, using just one ear? No kidding. Personal note: I was in just such an experiment when in college in 1960.
BTW, not sure how many recordings or movies have any stereo information. Localization is just simulated by the recording engineer using a pan pot to produce loudness differences between L and R (and maybe somebody has a recording board that also "invents" timing differences these days).
https://en.wikipedia.org/wiki/Sound_localization
For vision and hearing, the brain is really good at taking a whole mess of different and varying cues and producing a stable image (OK, it is only observable as "stable" when you turn your attention to it). With the image formed, the clarinet player will always seem to keep the same seat once your brain has accumulated enough cues to locate the player.
For a fun example to confound the wannabee-engineers, there is evidence that the difference in shape between your own two pinnas constitutes a spectral localization cue.... and helps with the paradox of mid-line localization.
BTW esl 63, apropos your behind the back gedank experiment, did you know that there is localization, albeit crude, using just one ear? No kidding. Personal note: I was in just such an experiment when in college in 1960.
BTW, not sure how many recordings or movies have any stereo information. Localization is just simulated by the recording engineer using a pan pot to produce loudness differences between L and R (and maybe somebody has a recording board that also "invents" timing differences these days).
I have experienced the effects of combined phase flattening, full-range controlled directivity (down to the room's Schroeder frequency) and nearfield midrange/low-treble absorption (as opposed to moving the loudspeakers out into the room away from the walls and/or putting them on sticks above the floor), and I can say the effects are spectacular: Subconscious Auditory Effects of Quasi-Linear Phase Loudspeakers (ff).
In fact, I have learned a lot more about why certain loudspeakers sound the way they do by simply correcting their excess phase response.
However, I do have some comments about the video, above, that I think need to be said:
In fact, I have learned a lot more about why certain loudspeakers sound the way they do by simply correcting their excess phase response.
However, I do have some comments about the video, above, that I think need to be said:
- Much of what the video's author presented is related to the all-pass injection of the crossover filters used (I think he stated LR24 IIR filters), and that he is simply taking out that all-pass behavior using RePhase convolution filters in his player or DSP crossover (FIR-type). If you are using steep crossover filters between the ways of your loudspeakers (and LR24 filters inject at least 360 degrees of phase lag of the lower frequency drivers relative to the higher frequency drivers). That's a lot. My experience has been that the phase swings need to be within ±90 degrees of the mean phase (i.e., zero degrees) to reach inaudibility of phase shifts. YMMV.
- The author is using loudspeakers having two sets of reflex ports (if I understand his comments correctly). I avoid loudspeakers having reflex ports due to the their effects on phase response at low frequencies, which are hard to correct and lead to excessive audio system time delays relative to the source, and potential video synchronization issues. Additionally, the author here is spending time trying to undo the effects of his loudspeakers' reflex ports, and that can simply be avoided by not using them in your loudspeaker designs.
- The negative effects on the sound quality presented by reflex ported loudspeakers extends to beyond merely phase growth issues, i.e., there are other noted dynamic effects in the time domain that are not pleasing to my ear. I recommend infinite baffle/horn loading instead, and perhaps the introduction of full-time subwoofers at 40 Hz and below to avoid the effects of modulation distortion seen in bass reflex designs.
- The effects of measuring the loudspeakers at the listening position on the phase response is overwhelming to the eye on the phase response plots. If you simply move the measurement microphone to be 1m in front of the front baffle of the loudspeaker under test, then place plenty of absorption material on the floor between the loudspeaker and the microphone (about a metre wide on the floor), all the noisy phase response plots will immediately clean themselves up to something that is easily seen and comprehended, without leading the RePhase user in circles trying to correct for non-minimum-phase room behavior.
- If you are in a relatively small listening room <2m from your loudspeakers, then the techniques shown by the author are usually valid. If your listening position(s) are 3 m or more from the loudspeakers (like most of the audio enthusiasts I know in my geographic area), then measuring at the listening position introduces too many issues, IMHO. If you wish to phase correct your loudspeaker itself, use 1 m microphone measurement distance, then if you are trying to correct the time delays (and therefore phase) of the loudspeakers in a stereo or multichannel 5.1, etc. array, then you can measure at the listening position, but I find that trying to EQ amplitude or phase response at the listening position is limited to only attenuating peaks in response due to room resonances.
- If you are using direct radiating loudspeakers, and trying to measure at the listening position(s) more than 1m away from each loudspeaker's front baffle, you are kidding yourself. You will have to either move into the loudspeaker with the measurement microphone to exclude the non-minimum-phase room effects of early reflections, use gobs of broad-band absorption in a live-end/dead-end listening room arrangement, or use loudspeakers having much better full-range directivity control. Otherwise, you'll never actually hear the effects of phase flattening in-room.
- In general, the video simply moves too fast for those trying to learn the techniques of DSP phase correction using both REW and RePhase, and is in my estimation turning off more people than enticing them into the fold. I know that the video can be played over and over until the student viewer begins to get what the author is doing, but the fact of the matter is that this puts a big load on the learner to pick it up, something that can be avoided if the instructor simply slows down a bit and is more verbose about switching windows and filling in numeric field values using the two applications.
In my engineering career, I've had to learn the hard way (i.e., through reading and trial-and-error, without videos). This was time consuming and tough-that ends up weeding out the casual learners from those that have to learn the material the hard way to do their jobs. When I started teaching graduate engineering classes, I immediately learned to slow down and explain the window switches and application input windows much more carefully, even though I was simultaneously being videoed for distance learners.
- In general, I think the techniques used by the author are actually going to become the mainstream of 21st century home audio enthusiasts, but it will likely alienate the older "mossback" audiophiles that fail to educate themselves on the theory and use the new tools to eek out that last 5% of loudspeaker/room performance out of their setups. As such, I think that--even in this thread--you can see the effects of the introduction of DSP/signal processing techniques, combined with in-room acoustic measurements on the viewing audience, separable into age buckets for the students. Some reject the concepts outright, even before understanding what they are watching, and experiencing the effects subjectively in-room.
Last edited:
Anyone ever test a log vs linear sweep and try out different phase rotations for the start of the sweep?
I often wonder if flattening the phase for one kind sweep is enough. It is so easy to be myopic - even if you use lots of tools.
I attached a 600hz rotating linear sweep (it tilts up if your system is flat for a log sweep).
Aside from "flattening", all are within the widely held catechism of Toole, show up on ordinary Spinorama freq responses, and accepted by everybody... not that everybody goes to the trouble of putting thick carpets on their floors and walls as they ought to.
A lot of the "experiments" remind me of epidemiological health research where "meat eaters" die sooner than "vegetarians". But being "vegetarian" comes along with a whole bunch of other life-style differences. Say, as Chris does, that ported boxes are deleterious for phase. Well, ported boxes are terrible in other ways too; so you can't credit just the phase improvement. Ditto for XOs.
B.
Thanks Chris for constructive information.
1. Agree, and i have heard the effect of DIRAC demo on a small DYNAUDIO special 25 that only has 6dB slope filters. And it was an improvement.
2. Using electrostatic speakers and a crossower at approx 100 Hz to dipole woofers 3 x 15" MFB drivers (24dB) and i have an analog all pass filter that i use to do phase shift ( 0-170 deg) for integration at crossover frequency. Below 37 Hz there is another 24dB crossover to the "garage subwoofer" consisting of 4 x15" drivers that pressurizes the room between 15Hz and 37Hz the room is 32 feet or 10m long... so no need for dipole below 37Hz. Also this gives space behind the speaker approx 6,5 meter. The garage volume is big enough to avoid ports 🙂 But with those filters and the distance to the garage woofer introduces phase and time shifts that i want to get rid of.
3. Exactly! Once you heard dipole bass there is no going back to reflex.
4. The first reflex floor roof sidewall is objects for diffusion and/or damping. I use BAD (binary array diffusers) that is a type of a combination of both. On the floor I place absorbing devices. Behind the speakers there is so late reflections that I do NOT damp them. so late reflections contributes to the blending between acoustics and music source information. There is quite much psyco acoustics to read about that.
5. Agree, and nr one is to do the correct timing first. HE (in the link) has videos about how to smooth the low frequency response using several subs and use delay and phase corrections. The result is really good without need to turning the EQ knobs... at the final stage some peaks is attenuated. It´s not magic it´s physics.
6. Yepp
7. Thats the reason i asked the question... It seems that he knows John Mulcahy (REW) and in one video from last year he showed functions that was not released... or maybe released the same week. And he referred to John as he knew him... so I agree, this is not for beginners. But if you have read some books or have a grade in acoustics then you are better off.
8. DIRAC has a module called "Unison" that is used by some car companies. Since all speaker placements is well known... and the shape of the listening room, they can do some magic tricks which I do not see that we (audio enthusiasts) will have for many years due to the complexity of the tuning. My collegues is spending weeks/months for every car model. Using special written tuning software. They also have array microphones with many many microphones to calculate the direction, phase, amplitude from each driver (25 driver units, incl Atmos) . What UNISON does is that it uses all drivers to cooperate and control the reverberation and acoustics. Some elements acts as absorbers as they actively controls the reverberation pattern.
https://www.dirac.com/2016-2-11-int...-entirely-new-approach-to-sound-optimization/
Hearing it is the only way to get an idea what is happening. The listening room gets BIG!! Depending on what listening room you enter in the infotainment menu the car is transformed to Gothenborg concert hall as an example. This can only be done when you know exactly where all drivers is located and what behaviour they have. This is not the case for home cinema owners yet. But maybe in the future the you get an helmet with ear plugs and additional 25 microphones, and you are instructed to move around in the sofa with the helmet listening to chirp signals until the speakers tells you to take off the helmet and buckle up, accelerate in to the audio Nirvana and never use a SONOS mono speaker again. But until then I will try myself and see how deep this rabbit hole goes. If you have a friend With the Volvo XC 90 or a dealer next by. Take the opportunity to listen. And one advice is to set ALL parameters to default factory settings....
1. Agree, and i have heard the effect of DIRAC demo on a small DYNAUDIO special 25 that only has 6dB slope filters. And it was an improvement.
2. Using electrostatic speakers and a crossower at approx 100 Hz to dipole woofers 3 x 15" MFB drivers (24dB) and i have an analog all pass filter that i use to do phase shift ( 0-170 deg) for integration at crossover frequency. Below 37 Hz there is another 24dB crossover to the "garage subwoofer" consisting of 4 x15" drivers that pressurizes the room between 15Hz and 37Hz the room is 32 feet or 10m long... so no need for dipole below 37Hz. Also this gives space behind the speaker approx 6,5 meter. The garage volume is big enough to avoid ports 🙂 But with those filters and the distance to the garage woofer introduces phase and time shifts that i want to get rid of.
3. Exactly! Once you heard dipole bass there is no going back to reflex.
4. The first reflex floor roof sidewall is objects for diffusion and/or damping. I use BAD (binary array diffusers) that is a type of a combination of both. On the floor I place absorbing devices. Behind the speakers there is so late reflections that I do NOT damp them. so late reflections contributes to the blending between acoustics and music source information. There is quite much psyco acoustics to read about that.
5. Agree, and nr one is to do the correct timing first. HE (in the link) has videos about how to smooth the low frequency response using several subs and use delay and phase corrections. The result is really good without need to turning the EQ knobs... at the final stage some peaks is attenuated. It´s not magic it´s physics.
6. Yepp
7. Thats the reason i asked the question... It seems that he knows John Mulcahy (REW) and in one video from last year he showed functions that was not released... or maybe released the same week. And he referred to John as he knew him... so I agree, this is not for beginners. But if you have read some books or have a grade in acoustics then you are better off.
8. DIRAC has a module called "Unison" that is used by some car companies. Since all speaker placements is well known... and the shape of the listening room, they can do some magic tricks which I do not see that we (audio enthusiasts) will have for many years due to the complexity of the tuning. My collegues is spending weeks/months for every car model. Using special written tuning software. They also have array microphones with many many microphones to calculate the direction, phase, amplitude from each driver (25 driver units, incl Atmos) . What UNISON does is that it uses all drivers to cooperate and control the reverberation and acoustics. Some elements acts as absorbers as they actively controls the reverberation pattern.
https://www.dirac.com/2016-2-11-int...-entirely-new-approach-to-sound-optimization/
Hearing it is the only way to get an idea what is happening. The listening room gets BIG!! Depending on what listening room you enter in the infotainment menu the car is transformed to Gothenborg concert hall as an example. This can only be done when you know exactly where all drivers is located and what behaviour they have. This is not the case for home cinema owners yet. But maybe in the future the you get an helmet with ear plugs and additional 25 microphones, and you are instructed to move around in the sofa with the helmet listening to chirp signals until the speakers tells you to take off the helmet and buckle up, accelerate in to the audio Nirvana and never use a SONOS mono speaker again. But until then I will try myself and see how deep this rabbit hole goes. If you have a friend With the Volvo XC 90 or a dealer next by. Take the opportunity to listen. And one advice is to set ALL parameters to default factory settings....
