Hmm,..errr... well I should explain that double knife edge at 1ms in the TDA. I am a sheepishly embarrassed about that one. 😱It's close at 1ms but to get a speaker to disappear more easily removing hints like that works really well. Not that a reflection this close would hint directly at the speaker position, it does influence the imaging. Especially early reflections.
It might not make a huge difference for the Omni, though it will make a difference for the 2.5 way for technical listening.
Everything after ~6-7 ms is way less detrimental from an imaging point of view. The rest of imaging quality is up to left and right speaker having very similar response to get the maximum out of that imaging. (FR and Phase)
It would even be interesting to try and clean it up for the Omni as well.
In trying to eliminate it. I put up absorbers (blankets), moved furniture, moved the TV, the speakers, everything I could think of. I chased that 1ms "white whale" all over the listening room and concluded it was not "physical". These "whales" are never real.
Your comments (sharp eye) and Hajj's comments (time alignment) were the clues. I realized (finally) that nothing in nature could cause a precise 1ms shift / reflection for all frequencies (different wavelengths). It lead me to conclude it was the AMP (oh no!). I use the AMP in "2ch direct sound" mode for 2ch measurements, which is suppose to bypass all DSP and tone controls, and it does, well mostly. However the time (distance) delays correction for speakers is still active. The default measure is "feet" and it rounded the distance to the nearest foot (30cm is 1ms) when it did it's auto EQ.
I have since corrected that (manually) and redid the TDA measurements and they are virtually the same. There is a little less HF noise but a smaller difference than when I quiet the room for testing. The graphs still look the same (sigh of relief). The FR plots are done with the AMP corrected.
The Omni has another problem I'm trying to fix, and its related to clarity and detail. I'm finding it hard to quantify it. More measurements.
Show the early waterfall plots 🙂. They are pretty honest for what hits your ears first.
If I'm searching for clarity I'll start there. Especially in the midrange.
Detail is a bit harder. A bright top end can give a sense of detail without it having to be real 🙂. Overall balance can also mess up what is perceived as detail. Especially if the area below 300 Hz is boosted compared to the higher frequencies it can sound warm, but it may cloud the detail in the higher range. There's a thin line where you can have both.
A gradual declining FR curve is usually good for that.
If I'm searching for clarity I'll start there. Especially in the midrange.
Detail is a bit harder. A bright top end can give a sense of detail without it having to be real 🙂. Overall balance can also mess up what is perceived as detail. Especially if the area below 300 Hz is boosted compared to the higher frequencies it can sound warm, but it may cloud the detail in the higher range. There's a thin line where you can have both.
A gradual declining FR curve is usually good for that.
Just some cleanup before moving forward.
Pic#1 - New waterfall for the Paradigm 7v5 with the AMP mis-delay fixed. This speaker (2ch) appear to be the most relatable system to use as a reference. The double knife edge @1m is removed. It improved the view of 7.5ms bounce, but the rest I believe are changes in background noise level, as the <50hz noise floor is up from my PC (running BEM sims and the room is warm now).
Pic#2 - I've added my standard measuring position FR @1m for the Paradigm 7v5 to view it with minimal room interaction. It will always have some, its just a lot less now. This can be directly compared to the Omni (other "in progress" thread). This can also be compared to results at the listening position (earlier pics).
Pic#3 - I've added the HT5.1 with the subwoofer @ listening position so you can see the bass extension. Its a dead give away, only a sub goes this low and its set for 60Hz as to not overlap the other speakers too much. The AMPs DSP is extracting bass from the LR and playing it, all speakers are set to "large" so they play bass as well. It appears I've been sitting in a good location with no serious "nulls".
Pic#1 - New waterfall for the Paradigm 7v5 with the AMP mis-delay fixed. This speaker (2ch) appear to be the most relatable system to use as a reference. The double knife edge @1m is removed. It improved the view of 7.5ms bounce, but the rest I believe are changes in background noise level, as the <50hz noise floor is up from my PC (running BEM sims and the room is warm now).
Pic#2 - I've added my standard measuring position FR @1m for the Paradigm 7v5 to view it with minimal room interaction. It will always have some, its just a lot less now. This can be directly compared to the Omni (other "in progress" thread). This can also be compared to results at the listening position (earlier pics).
Pic#3 - I've added the HT5.1 with the subwoofer @ listening position so you can see the bass extension. Its a dead give away, only a sub goes this low and its set for 60Hz as to not overlap the other speakers too much. The AMPs DSP is extracting bass from the LR and playing it, all speakers are set to "large" so they play bass as well. It appears I've been sitting in a good location with no serious "nulls".
Attachments
Last edited:
Here's an idea, clean up the Paradigm as much as you can and use the Omni as diffused sources in the back 😀.
With a PC you could control the timing of each, clean up the response some with FIR and get an awesome result. The floor reflection is hard to battle though.
Are you willing to upload an IR for me to play with? The one at the listening position showing this APL plot?
With a PC you could control the timing of each, clean up the response some with FIR and get an awesome result. The floor reflection is hard to battle though.
Are you willing to upload an IR for me to play with? The one at the listening position showing this APL plot?
Last edited:
IR measurements are next, and yes I can send you some data if you want to try IRC. Curious as to how many taps would be required in the FIR. Doesn't matter really as computing power and memory are so cheap and generic now.
A diffuser? 😀 Well, they are the right height, and they have a flat top so I could make them into snack trays with cup holders 😉 I'm not too attached to their current form, the particle board cabinets cost me $12 CDN each, and they are just "sandboxes" for me to play in for now. So far I'm encouraged by what I hear, but its not good enough yet.
Just a caution. I'm focused on the Omni, so I'm not trying to make the Paradigms sound better. This thread has been great as it allows me to share (comparison, methods, analysis) and for more eyes to look at and comment. Is this Omni lacking, better, worse than a standard speaker in a room and can anyone tell the differences via the measurements. I'll stop the "match" game because it really doesn't apply to IR and Distortion measurement analysis. I have a difficult time matching the previous FR results to the speakers and I own them and know what they sound like.
A diffuser? 😀 Well, they are the right height, and they have a flat top so I could make them into snack trays with cup holders 😉 I'm not too attached to their current form, the particle board cabinets cost me $12 CDN each, and they are just "sandboxes" for me to play in for now. So far I'm encouraged by what I hear, but its not good enough yet.
Just a caution. I'm focused on the Omni, so I'm not trying to make the Paradigms sound better. This thread has been great as it allows me to share (comparison, methods, analysis) and for more eyes to look at and comment. Is this Omni lacking, better, worse than a standard speaker in a room and can anyone tell the differences via the measurements. I'll stop the "match" game because it really doesn't apply to IR and Distortion measurement analysis. I have a difficult time matching the previous FR results to the speakers and I own them and know what they sound like.
FR measurements matched to speaker
Repeated (pics left to right). See the next post for analysis. FR results are at the room listening position.
Pic#1 - photo of the speakers
Pic#2 - Speaker A - Omni (homebrew)
Pic#3 - Speaker B - Bookshelf Mission 701
Pic#4 - Speaker C - 2.5 way Paradigm 7v5
Pic#5 - Speaker D - 5.1 setup playing 5.0 (no sub)
Repeated (pics left to right). See the next post for analysis. FR results are at the room listening position.
Pic#1 - photo of the speakers
Pic#2 - Speaker A - Omni (homebrew)
Pic#3 - Speaker B - Bookshelf Mission 701
Pic#4 - Speaker C - 2.5 way Paradigm 7v5
Pic#5 - Speaker D - 5.1 setup playing 5.0 (no sub)
Attachments
Analysis FR measurement results
I found this nearly impossible after taking the measurement, and I own the speakers. I actually measured twice as I was expecting more vivid differences and didn't see them. I started out looking for technical analysis to match the speaker based on its size, nature and arrangement. The only thing I can say is a) none are broken, and b) since I picked them, they likely share the same preference bias. By broken, I mean they cover most of the audio range, are reasonably flat for a room measurement, and do not have glaring problems.
What is that "hump" around 120Hz. It appears in all measurements, and the common denominator is the room, as it can't be all the speakers. The listening position is 1.3m from back wall, round trip is 2*1.3m=2.6m, and its constructive interference (waves add). This puts the freq at 344/2.6 =132Hz +/-16Hz as up to 1/8 wavelength will still contribute. This agrees the the TDA done earlier in both time and frequency. Nice to have them consistent as they are different s/w tools and use different test methods.
Fluid had some good deductive reasoning (in PM) that I'll repeat. The seperate Omni thread lists the FR@1m and it mostly matches the room (SpeakerA done). The last 2 (C&D) share the same front LR (7v5) only the HT smoothes things and boosts bass (SpeakerC=7v5, SpeakerD=HT done). That leaves confirmation of SpeakerB which has earliest HF roll off and worst LF (SpeakerD done). That said, its hardly a technical argument and more of process of elimination. Fair enough.
Comments about the speakers. The Mission uses a repaired (by me) AL tweeter, a mishap projected the tweeters so far forward it snapped the VC wires and sprayed ferrofluid onto the back on the diaphragms. Amazing it still works. The Paradigm uses a Ti tweeter and has a pronounced peak near 20Khz, which I can't hear, but I think causes some modulation ringing at other HF's occasionally. The Omni uses a waveguide and compression driver thats too bright in MF but its EQ down when listening but it behaves well otherwise. The Omni has the largest woofer (8inch vs 5inch) that projects into the room better but it doesn't show in the graphs as their may be a volume (level) dependancy when I listen to them.
I found this nearly impossible after taking the measurement, and I own the speakers. I actually measured twice as I was expecting more vivid differences and didn't see them. I started out looking for technical analysis to match the speaker based on its size, nature and arrangement. The only thing I can say is a) none are broken, and b) since I picked them, they likely share the same preference bias. By broken, I mean they cover most of the audio range, are reasonably flat for a room measurement, and do not have glaring problems.
What is that "hump" around 120Hz. It appears in all measurements, and the common denominator is the room, as it can't be all the speakers. The listening position is 1.3m from back wall, round trip is 2*1.3m=2.6m, and its constructive interference (waves add). This puts the freq at 344/2.6 =132Hz +/-16Hz as up to 1/8 wavelength will still contribute. This agrees the the TDA done earlier in both time and frequency. Nice to have them consistent as they are different s/w tools and use different test methods.
Fluid had some good deductive reasoning (in PM) that I'll repeat. The seperate Omni thread lists the FR@1m and it mostly matches the room (SpeakerA done). The last 2 (C&D) share the same front LR (7v5) only the HT smoothes things and boosts bass (SpeakerC=7v5, SpeakerD=HT done). That leaves confirmation of SpeakerB which has earliest HF roll off and worst LF (SpeakerD done). That said, its hardly a technical argument and more of process of elimination. Fair enough.
Comments about the speakers. The Mission uses a repaired (by me) AL tweeter, a mishap projected the tweeters so far forward it snapped the VC wires and sprayed ferrofluid onto the back on the diaphragms. Amazing it still works. The Paradigm uses a Ti tweeter and has a pronounced peak near 20Khz, which I can't hear, but I think causes some modulation ringing at other HF's occasionally. The Omni uses a waveguide and compression driver thats too bright in MF but its EQ down when listening but it behaves well otherwise. The Omni has the largest woofer (8inch vs 5inch) that projects into the room better but it doesn't show in the graphs as their may be a volume (level) dependancy when I listen to them.
Picking a bad spot - on purpose
The previous listening position had a "hump" at 132Hz. That was not chosen, it just happened because that was a good sounding spot, and it suited the furniture arrangement.
What if you're not lucky? What does a bad spot look like. I used REW room sim tab to setup my room, speakers and mic. Then I moved the REW mic locator around until I hit a predicted front-back axial mode that had destructive interference causing a bad dip @120Hz (a hole). It's the exact opposite of the constructive interference that made a 130Hz bump earlier. The scales are adjusted so both graphs kinda line up. BTW the absorption and reflection in a room are never exact you may have to move abit (cm's and the predicted frequency is never exact.
This -15dB@120Hz dip cannot be equalized away in a 2 speaker system. The waves always cancel regardless of amplitude adjustments. Something has to move (you or the speakers) or "maybe" you can treat the room, or add more speakers.
The previous listening position had a "hump" at 132Hz. That was not chosen, it just happened because that was a good sounding spot, and it suited the furniture arrangement.
What if you're not lucky? What does a bad spot look like. I used REW room sim tab to setup my room, speakers and mic. Then I moved the REW mic locator around until I hit a predicted front-back axial mode that had destructive interference causing a bad dip @120Hz (a hole). It's the exact opposite of the constructive interference that made a 130Hz bump earlier. The scales are adjusted so both graphs kinda line up. BTW the absorption and reflection in a room are never exact you may have to move abit (cm's and the predicted frequency is never exact.
This -15dB@120Hz dip cannot be equalized away in a 2 speaker system. The waves always cancel regardless of amplitude adjustments. Something has to move (you or the speakers) or "maybe" you can treat the room, or add more speakers.
Attachments
Last edited:
If you really want to analyse you've got to look at differences.
The room will be the same for all speakers, if you measure them in the same position they will have similar room effects, aside from the Omni that hits a few more spots.
Look at the first wave front. A waterfall plot shows that (as does APL). A frequency dependant window can also show that.
You can't boost a null. That's true, but you can find what creates the null.
If we look at the Omni APL plot:
We can see the early wave front above about 300 Hz isn't forming a straight clear line. This will also show up in the IR. You'll see lots of early reflections if measured at the listening position.
However, in your Omni thread you've shown a closer measurement:
Here that ridge is present. So the Omni is perfectly capable of getting the first wave front to you in perfect condition. (reason why I picked the wrong graph when searching for the Omni in that first round)
Can the Omni do better at the listening position? You bet! Just make sure it's positioned more freely away from close boundaries.
If we don't get that clear ridge above 300 Hz at the listening spot it has to be something near the speakers that disrupts the wave front before it hits the microphone.
Your pictures show the Omni and all other speakers beside it are positioned very close to it. If it's measured like that the first wave front is reflected and diffracted off of all those speakers. The Omni would need space around it to be able to get everything in line at the microphone.
Do we hear the ragged response like the Omni shows in the APL listening spot plot? That would depend on the overall room response over a longer time frame. But what it does do is being less clear, less detail projected to the listener. Even the FR response of that first wave front plays a role. As balance between the different frequency areas plays a huge role in how we interpret the little things. Even a 0.5 dB difference can totally change the perception if it's over a broad frequency range.
So would it be worth it to get everything above ~300 Hz more reflection free? It would be if you like detail and clarity. Get that right for the Omni and it will be able to do better in both clarity and detail.
The room will be the same for all speakers, if you measure them in the same position they will have similar room effects, aside from the Omni that hits a few more spots.
Look at the first wave front. A waterfall plot shows that (as does APL). A frequency dependant window can also show that.
You can't boost a null. That's true, but you can find what creates the null.
If we look at the Omni APL plot:

We can see the early wave front above about 300 Hz isn't forming a straight clear line. This will also show up in the IR. You'll see lots of early reflections if measured at the listening position.
However, in your Omni thread you've shown a closer measurement:

Here that ridge is present. So the Omni is perfectly capable of getting the first wave front to you in perfect condition. (reason why I picked the wrong graph when searching for the Omni in that first round)
Can the Omni do better at the listening position? You bet! Just make sure it's positioned more freely away from close boundaries.
If we don't get that clear ridge above 300 Hz at the listening spot it has to be something near the speakers that disrupts the wave front before it hits the microphone.
Your pictures show the Omni and all other speakers beside it are positioned very close to it. If it's measured like that the first wave front is reflected and diffracted off of all those speakers. The Omni would need space around it to be able to get everything in line at the microphone.
Do we hear the ragged response like the Omni shows in the APL listening spot plot? That would depend on the overall room response over a longer time frame. But what it does do is being less clear, less detail projected to the listener. Even the FR response of that first wave front plays a role. As balance between the different frequency areas plays a huge role in how we interpret the little things. Even a 0.5 dB difference can totally change the perception if it's over a broad frequency range.
So would it be worth it to get everything above ~300 Hz more reflection free? It would be if you like detail and clarity. Get that right for the Omni and it will be able to do better in both clarity and detail.
IR measurements are next, and yes I can send you some data if you want to try IRC. Curious as to how many taps would be required in the FIR. Doesn't matter really as computing power and memory are so cheap and generic now.
Don't focus on number of tabs as DSP can't fix everything. I only use DSP to fix the small stuff and due to using arrays to get a reasonable frequency response. But DSP can't fix all room issues. It can, but only at that single spot and it would make other spots worse. So if we get the speakers to behave over a wide area DSP can help make it better. So why does DSP room correction work at lower frequencies? Look at the wave lengths. But DSP can't fix position dependant reflections.
So we need to do all we can with placement and room integration first. Then use some DSP to get the rest in line. I'd rather fix the speakers output with DSP than try and fix the room.
The number of tabs would only determine the resolution down in the low frequencies.
I'll get back to this topic in a little while... still have some other work I need to do first.
Last edited:
Lets look at room interaction and our perception of that for a bit.
Let someone close to you talk in your living room and stand about one meter away from you. Do you consciously hear the room you are in? I bet if you try you can hear it.
Next move to a way smaller place, like a bathroom or even a toilet. Repeat the talk, do you hear the room now? What changed in your perception?
Next move outside and repeat the conversation. Does it sound different to you again?
Our brain filters all kinds of things without us making conscious decisions about it all the time. But it does help you out doing this. It actually measures the (size of the) room for you, helped out by using your eyes too. This way we adapt to the room and how it sounds without thinking much about it all. If we place a speaker in the room it will still do the same. You might enjoy the speakers while your brain adapts to what is heard. If you record it at the listening position, and play it back on headphones you'll become much more aware of what the room is doing to the sound at that spot. Suddenly you do hear more room. Way more than what was perceived in that listening spot.
So whenever we listen to speakers we listen to the room too (obviously). Part of that will be filtered out by our brains ability to adapt to spaces (without thinking).
This is the reason we don't always judge a (gradual or even sharper) dip in the frequency response as offensive right away. But it does alter what we think we hear. Music producers and mixers use this all the time.
Our brain picks up the first sound but also picks up the reflections and based on it's content it lumps together the parts that make up the whole sound and uses that to determine the total.
A cool description of that process can be found in this piece: Moulton Laboratories :: The Brave New World: Loudspeakers to the Left of Us! Loudspeakers to the Right of Us!
So while listening to speakers our brain is at work to figure out which parts belong together all the time. That actually determines the positional queues in imaging. However, in Stereo we listen with two ears to two speakers. This further complicates our decision making.
Because the left speakers wave front also reaches the right ear and vice versa. This creates a combing pattern that introduces holes in the direct frequency response at our ears. Reflections soften those holes over a little longer time.
So the harder we work on a clean reflection free area to listen in we change that part of Stereo perception too! We get to hear more obvious combing dips right in that sweet spot if all reflections within the room are absorbed. Moving just a tiny bit changes where those dips in the FR are going to be.
This is not nearly as obvious in a room that still has some reflections. Yet those reflections do trigger our brain where the sound is coming from.
We can decipher the location or origin of sounds that way trough picking up subtle hints in the form of those reflections... The first arriving reflections within ~3ms completely change that image perception. Later reflections do still count to alter tonal perception, but have less influence on the direction of sounds as interpreted by our brain.
So no graph by itself can tell you exactly what you hear, you need to dig way deeper to determine what your brain gets to process. The measurements don't lie. The better we get at understanding of how our brain interprets that measured signal can make us relate those measurements back to what it is we perceive.
All of that work our brain does without us really working for it will determine how the overall sound is perceived. The more we use what we know or have learned about those learned behaviours, the more we can play with getting that sound field to our liking.
An overall FR response like we looked at in this thread tells us nothing. At least not anything useful about how we perceive it.
Let someone close to you talk in your living room and stand about one meter away from you. Do you consciously hear the room you are in? I bet if you try you can hear it.
Next move to a way smaller place, like a bathroom or even a toilet. Repeat the talk, do you hear the room now? What changed in your perception?
Next move outside and repeat the conversation. Does it sound different to you again?
Our brain filters all kinds of things without us making conscious decisions about it all the time. But it does help you out doing this. It actually measures the (size of the) room for you, helped out by using your eyes too. This way we adapt to the room and how it sounds without thinking much about it all. If we place a speaker in the room it will still do the same. You might enjoy the speakers while your brain adapts to what is heard. If you record it at the listening position, and play it back on headphones you'll become much more aware of what the room is doing to the sound at that spot. Suddenly you do hear more room. Way more than what was perceived in that listening spot.
So whenever we listen to speakers we listen to the room too (obviously). Part of that will be filtered out by our brains ability to adapt to spaces (without thinking).
This is the reason we don't always judge a (gradual or even sharper) dip in the frequency response as offensive right away. But it does alter what we think we hear. Music producers and mixers use this all the time.
Our brain picks up the first sound but also picks up the reflections and based on it's content it lumps together the parts that make up the whole sound and uses that to determine the total.
A cool description of that process can be found in this piece: Moulton Laboratories :: The Brave New World: Loudspeakers to the Left of Us! Loudspeakers to the Right of Us!
So while listening to speakers our brain is at work to figure out which parts belong together all the time. That actually determines the positional queues in imaging. However, in Stereo we listen with two ears to two speakers. This further complicates our decision making.
Because the left speakers wave front also reaches the right ear and vice versa. This creates a combing pattern that introduces holes in the direct frequency response at our ears. Reflections soften those holes over a little longer time.
So the harder we work on a clean reflection free area to listen in we change that part of Stereo perception too! We get to hear more obvious combing dips right in that sweet spot if all reflections within the room are absorbed. Moving just a tiny bit changes where those dips in the FR are going to be.
This is not nearly as obvious in a room that still has some reflections. Yet those reflections do trigger our brain where the sound is coming from.
We can decipher the location or origin of sounds that way trough picking up subtle hints in the form of those reflections... The first arriving reflections within ~3ms completely change that image perception. Later reflections do still count to alter tonal perception, but have less influence on the direction of sounds as interpreted by our brain.
So no graph by itself can tell you exactly what you hear, you need to dig way deeper to determine what your brain gets to process. The measurements don't lie. The better we get at understanding of how our brain interprets that measured signal can make us relate those measurements back to what it is we perceive.
All of that work our brain does without us really working for it will determine how the overall sound is perceived. The more we use what we know or have learned about those learned behaviours, the more we can play with getting that sound field to our liking.
An overall FR response like we looked at in this thread tells us nothing. At least not anything useful about how we perceive it.
Last edited:
If you really want to analyse you've got to look at differences.
The room will be the same for all speakers, if you measure them in the same position they will have similar room effects, aside from the Omni that hits a few more spots.
Agreed, that was the reason for testing more than one speaker.
We can see the early wave front above about 300 Hz isn't forming a straight clear line. This will also show up in the IR. You'll see lots of early reflections if measured at the listening position.
However, in your Omni thread you've shown a closer measurement:
Here that ridge is present. So the Omni is perfectly capable of getting the first wave front to you in perfect condition.
That's a fair comment. The traditional speakers fire forward in a "cone" pattern so they are not as affected as much by an adjacent boundary. The Omni fires in a torroidal pattern so everything near it gets sprayed with sound.
Can the Omni do better at the listening position? You bet! Just make sure it's positioned more freely away from close boundaries.
If we don't get that clear ridge above 300 Hz at the listening spot it has to be something near the speakers that disrupts the wave front before it hits the microphone.
OK, that's a good observation. I'll move some stuff and remeasure. The best position is probably right beside (L&R side) the flat TV such that the TV side is aligned to the Omni disc centers. All other stuff will be removed or adjusted to see if I can repeat the single knife edge TDA in a room.
So would it be worth it to get everything above ~300 Hz more reflection free? It would be if you like detail and clarity. Get that right for the Omni and it will be able to do better in both clarity and detail.
Test results tomorrow.
Lets look at room interaction and our perception of that for a bit.
Floyd Toole's "Sound Reproduction, Loudspeakers and Rooms" has been a great reference. He presents a lot of double bind test result for psychoacoustics as well as corresponding test measurements. I'm about 1/2 way through it now (Part 2).
Let someone close to you talk in your living room and stand about one meter away from you. Do you consciously hear the room you are in? I bet if you try you can hear it.
Next move to a way smaller place, like a bathroom or even a toilet. Repeat the talk, do you hear the room now? What changed in your perception?
Next move outside and repeat the conversation. Does it sound different to you again?
Yes, in many ways the "liveliness" of the environment is determined by the reflections. My preference is for more lively rooms (not echo chambers) for enjoyment. This room we're testing in is "neutral" in that it is neither lively nor dead. BTW, my least favourite acoustic environments are anachoic chambers and open fields because they appear "flat" and 2D to me. But that's just my preference.
So while listening to speakers our brain is at work to figure out which parts belong together all the time. That actually determines the positional queues in imaging.
Agreed, in many ways I've read comparisons to other senses we have. Like your eyes will adjust for ambient light and color in an autonomous manner. You cannot prevent it from happening, so you may as well understand it and possibly take advantage of it.
So no graph by itself can tell you exactly what you hear, you need to dig way deeper to determine what your brain gets to process. The measurements don't lie. The better we get at understanding of how our brain interprets that measured signal can make us relate those measurements back to what it is we perceive.
All of that work our brain does without us really working for it will determine how the overall sound is perceived. The more we use what we know or have learned about those learned behaviours, the more we can play with getting that sound field to our liking.
An overall FR response like we looked at in this thread tells us nothing. At least not anything useful about how we perceive it.
Although it is interesting that a relatively "flat" response in desired as a starting point so the reproduction is not too tonally colored. Also, its interesting because a lot of audio gear is sold based on the anachoic FR measurements.
Tom Hidley, the world most famous acoustician proposed “non-environment” for the music studio control room. It is a semi-anechoic conditions where the only room surfaces left were the front wall and the floor, acoustically speaking. The remaining walls and the ceiling were made (nearly) totally sound absorbent.
One of my room built in the garage is treated somewhat like this, ceiling and back wall is very deeply covered with truckload of Owens Corning 704, and rest of the wall is all diffused at ear height. I also use a 2.5 feet height small buffer wall just in front of the listing position to kill the direct reflection from the floor. (The buffer also works to mimic the concert hall experience where chairs are always in front of me). This room was originally built for music mixing purpose. Very flat RT60 above 200Hz, and ver low early reflection.
So far so good, but there is a problem with this room as a music “listening” room. Sound stage is small compared to my living room that has zero treatment with horrible reflections, although the music in this heavily treated studio room is much more accurate (both objective and subjective) and has a super precise imaging.
My current conclusion is a neutral room might not be an ideal room for ordinary music listening purpose. I think we want exaggerated sound stage environment.
For example, many audiophiles select the speakers with exaggerated sound stage rather than neutral ones, hoping he can have better sound stage. B&W speakers which high frequency is intentionally boosted to give the illusion of larger sound stage. It is not only B&W. 99% of the commercial speakers are not properly time aligned, and the phase shift makes human feel the speaker has larger sound stage. Many seasoned audiophiles loves back loaded horn and OB, but their phase response is horrible = better sound stage feel. Rear firing tweeters, omni directional, back loaded horn, OB, IR, all of them are for exaggerated sound stage. This is probably why Danlavy had to close his business. Almost of all Danlavy customers were mastering engineers who wanted to have accurate speakers.
Then you need a neutral room? I’m skeptical about it.
One of my room built in the garage is treated somewhat like this, ceiling and back wall is very deeply covered with truckload of Owens Corning 704, and rest of the wall is all diffused at ear height. I also use a 2.5 feet height small buffer wall just in front of the listing position to kill the direct reflection from the floor. (The buffer also works to mimic the concert hall experience where chairs are always in front of me). This room was originally built for music mixing purpose. Very flat RT60 above 200Hz, and ver low early reflection.
So far so good, but there is a problem with this room as a music “listening” room. Sound stage is small compared to my living room that has zero treatment with horrible reflections, although the music in this heavily treated studio room is much more accurate (both objective and subjective) and has a super precise imaging.
My current conclusion is a neutral room might not be an ideal room for ordinary music listening purpose. I think we want exaggerated sound stage environment.
For example, many audiophiles select the speakers with exaggerated sound stage rather than neutral ones, hoping he can have better sound stage. B&W speakers which high frequency is intentionally boosted to give the illusion of larger sound stage. It is not only B&W. 99% of the commercial speakers are not properly time aligned, and the phase shift makes human feel the speaker has larger sound stage. Many seasoned audiophiles loves back loaded horn and OB, but their phase response is horrible = better sound stage feel. Rear firing tweeters, omni directional, back loaded horn, OB, IR, all of them are for exaggerated sound stage. This is probably why Danlavy had to close his business. Almost of all Danlavy customers were mastering engineers who wanted to have accurate speakers.
Then you need a neutral room? I’m skeptical about it.
I agree with what you're saying w.r.t requirements for a "technical listener" vs a "recreational listener".
The garage you described, I would call it "dead" sounding as you've removed too many reflections. My use of "neutral" meaning that it contains moderate reflections (see TDA plots). It has furniture and carpet but makes no control attempts. A "lively" room would have many bare surfaces, hardwood floors and minimal furniture. So I use "neutral" in a different tense, I do not mean the room effect has been "removed".
The garage you described, I would call it "dead" sounding as you've removed too many reflections. My use of "neutral" meaning that it contains moderate reflections (see TDA plots). It has furniture and carpet but makes no control attempts. A "lively" room would have many bare surfaces, hardwood floors and minimal furniture. So I use "neutral" in a different tense, I do not mean the room effect has been "removed".
I agree with what you're saying w.r.t requirements for a "technical listener" vs a "recreational listener".
The garage you described, I would call it "dead" sounding as you've removed too many reflections. My use of "neutral" meaning that it contains moderate reflections (see TDA plots). It has furniture and carpet but makes no control attempts. A "lively" room would have many bare surfaces, hardwood floors and minimal furniture. So I use "neutral" in a different tense, I do not mean the room effect has been "removed".
Yes, it's hard to say how much reflection would be "neutral", and obviously Tom Hidley and you have different opinion about it. George Massenburg should also have another opinion. 🙂
Although it is interesting that a relatively "flat" response in desired as a starting point so the reproduction is not too tonally colored. Also, its interesting because a lot of audio gear is sold based on the anachoic FR measurements.
You've got to figure out when you want the wave front to be right. 🙂
If I had an anechoic room I'd shoot for flat too. In a room everything changes.
For example, many audiophiles select the speakers with exaggerated sound stage rather than neutral ones, hoping he can have better sound stage. B&W speakers which high frequency is intentionally boosted to give the illusion of larger sound stage. It is not only B&W. 99% of the commercial speakers are not properly time aligned, and the phase shift makes human feel the speaker has larger sound stage. Many seasoned audiophiles loves back loaded horn and OB, but their phase response is horrible = better sound stage feel. Rear firing tweeters, omni directional, back loaded horn, OB, IR, all of them are for exaggerated sound stage. This is probably why Danlavy had to close his business. Almost of all Danlavy customers were mastering engineers who wanted to have accurate speakers.
Then you need a neutral room? I’m skeptical about it.
I'd say I have a time response which is comparable to Dunlavy's speakers. Yet with only a few changes this gets me width and depth without sacrificing the time response.
Fix the cross talk dip and it will change your perception (largely improving the sense of depth and more 3D overall). Stereo isn't a bad concept. Starting with a controlled room isn't a bad idea. Leaves less to chance.
Yes, it's hard to say how much reflection would be "neutral", and obviously Tom Hidley and you have different opinion about it. George Massenburg should also have another opinion. 🙂
I'll read with interest what all of these guys like Tom and George, Floyd, Earl, John, another Tom and many more have to say, but in the end I'm out to please Ronald!
It does help to read, read some more and try to connect the dots. Get the general consensus of all these guys together, put it all in a big bowl and stir it. After that, experiment for yourself.
Rather than take someone's word for it I've tried to find my own truth, with experiments heavily influenced by the published work from many, not limiting myself to one guru's opinion. If you ask me, every single one of them will be right about something. Pretty hard to know who will be right on everything.
Last edited:
TDA at various distances from speakers
These are the Omni speakers beside a large flat TV. This position should cause minimize near reflections as the angles are oblique and the side profile of the TV is thin. They are still 1.3m from a back wall as before, all other equipment has been removed.
All measurements are along the speaker midpoint line, using the disc center as d=0. So a d=0 measurement puts the mic midway between the speakers, and a d=1m moves it perpendicular out but along that midline.
Pics below are [0m, 0.5m, 1m, 1.5m, 2m, 2.5m, 3m, 3.5m] the last measurement is effectively against the wall. There is a sofa in the path at 2-2.9m. You can see that moving out shows more room interaction which means the ratio of direct / indirect sound is changing.
These are the Omni speakers beside a large flat TV. This position should cause minimize near reflections as the angles are oblique and the side profile of the TV is thin. They are still 1.3m from a back wall as before, all other equipment has been removed.
All measurements are along the speaker midpoint line, using the disc center as d=0. So a d=0 measurement puts the mic midway between the speakers, and a d=1m moves it perpendicular out but along that midline.
Pics below are [0m, 0.5m, 1m, 1.5m, 2m, 2.5m, 3m, 3.5m] the last measurement is effectively against the wall. There is a sofa in the path at 2-2.9m. You can see that moving out shows more room interaction which means the ratio of direct / indirect sound is changing.
Attachments
-
Centerline @0m.jpg252.5 KB · Views: 113
-
Centerline @0.5m.jpg262.4 KB · Views: 109
-
Centerline @1m.jpg265.8 KB · Views: 118
-
Centerline @1.5m.jpg294.2 KB · Views: 111
-
Centerline @2m.jpg299.6 KB · Views: 113
-
Centerline @2.5m.jpg306.8 KB · Views: 84
-
Centerline @3m.jpg317.8 KB · Views: 78
-
Centerline @3.5m.jpg323.9 KB · Views: 72
-snip ---
I'll read with interest what all of these guys like Tom and George, Floyd, Earl, John, another Tom and many more have to say, but in the end I'm out to please Ronald!
It does help to read, read some more and try to connect the dots. Get the general consensus of all these guys together, put it all in a big bowl and stir it. After that, experiment for yourself.
Rather than take someone's word for it I've tried to find my own truth, with experiments heavily influenced by the published work from many, not limiting myself to one guru's opinion. If you ask me, every single one of them will be right about something. Pretty hard to know who will be right on everything.
That last part is the most important issue I think. When you read that written description or opinion you still need to interpret what that means w.r.t what you personally want. As you say, this is all done to meet our personal preferences. Otherwise we'd all agree on one design, one flavour, one arrangement and one type of music 🙂
The ridge is back, which is a good thing. Now check the FR response, not of a ~500 ms window, instead use a frequency dependant window at the spot just in front of that couch.
Get that first wave front that hits your ears right. At least that's what I would do. There's a nice gap between the first wave front (now forming a ridge again) and the reflections. That's usually quite a pleasant sound. Make that first wave front as good as possible for left and right and try that. All the other graphs will start to look better if you do.
Get that first wave front that hits your ears right. At least that's what I would do. There's a nice gap between the first wave front (now forming a ridge again) and the reflections. That's usually quite a pleasant sound. Make that first wave front as good as possible for left and right and try that. All the other graphs will start to look better if you do.
- Status
- Not open for further replies.
- Home
- General Interest
- Room Acoustics & Mods
- Speaker and room interaction (pics) - your opinion