I've never seen such a step response (or group delay) and phase behaviour from a second order speaker before. If this were a commercial design it would cost mayor dollars I'm sure. More proof Troels knows what he's doing. It looks very well behaved in time domain and your room seems excellent!
2 ways to go forward. One would mean using active (FIR) filtering (I bet it would be awesome!) and the other would mean to pick speakers with even lower order filters (or no filters at all)
I'll let Troels explain the draw back from second order filters: Siri's Killer Note
So basically, there would/could still be small errors in harmonics. Troels seems like a very wise man indeed, knows what he is doing.
And you made an excellent room for these speakers. I bet even Troels would be proud to see these graph's.
Just to indulge me, could you show a Group Delay plot gated 20 ms? Just curious.
Too bad you wouldn't want to try active filters, just for fun!
Seems to me your speaker and room are up for the challenge. Any pictures to share of the room/speakers?
One more question: if you close your eyes while listening, is this a "They are here" or "You are there" kind of sound? In other words, do you get the impression the musicians are in the room with you or are you virtually transported to the actual event.
I dont mind taking other/more measurements. Give me some REW screen shots and specify how you want the configurations changed.
While this thread has mostly focused on GD, and many other audio enthusiasts focus on FR, Decay, and ETC, the single most important aspect in listening pleasure to me is imaging/soundstage.
When it comes to measurement data, imaging doesnt have a direct correlation. There is no single or combination of measurements that really directly tell you how well a system images. That said, focusing on imaging here for a moment, there are compromises.
A completely dead room properly employed with absorbent material seems to yield the best pinpoint imaging in terms of location. Yet, such a room is dry and sterile in terms of ambiance and envelopment. So comes the psycho-acoustic idea of allowing later (20-25ms delayed) reflections and room energy. But again, this is a compromise because it broadens the images and can lead to a diminished sense of air and space between images.
But to get to the second bolded part of your reply, yes, there is a strong sense of the speakers not being in the room at all. The soundstage is wide and deep with some vertical component as well. I do think there is room for improved envelopment and ambiance, but all the absorption for early reflection control diminishes the energy left over for later arrival.
I have toyed with the idea of adding rear full range speakers and delaying them 20-25ms with some sort of reverb to simulate later arriving diffused energy. But it would involve more speakers, more amps, and some sort of DSP for the delay and reverb. In other words, a fairly expensive venture. And one that may backfire.
Jim1961,
I have to say that if this is your phase plot from listening position - it is one of the best I have seen. This means you have a fantastic sounding system and room any recording studio would be proud to own. It also means you really know what you are doing and have eeked every little last bit of performance out of it. Can we get a photo of your setup so we can all admire what audio nirvana looks like?
Congratulations! I am not sure you can do much more other than go to DSP DRC and FIR!!!
All the measurements ive posted are at the LP. Now that you have looked over my "my room" thread over at Gearslutz, you can also see that ive been told before that ive eeked all the performance I can get. I am still wondering how deep the rabbit hole goes for I continue to find ways to make things better.
As far as knowing what I am doing, ive made nine mistakes for every one step forward. As they say, a broken clock is correct once a day
If your trying to catch a criminal, you follow the money. In audio, you follow the measurements. But in the case of audio, you have to interpret the measurement data correctly. To this point, I am still learning.
I live alone, so I am able to do whatever I want. I sometimes wonder why some/most audio enthusiasts put so much effort into a properly designed speaker and put none at all into the room. This is a bit like having a $2000 DAC with $200 speakers. But then I remember that aesthetics and WAF (wife approval factor) play a part for others and dont for me.
Hi Jim,
I recognise some parts of the journey you are on. Though I try to do it in a different way I'm after much of the same things you are. So I try to learn from you here, as you've been on quite the journey already.
If there is a correlation between imaging and measurements I'd look to FR, GD and phase to find it. Actually I am looking there.
Phase is one of the steering mechanisms used in the music.
I do get that late reflection you added, I've been into Car audio for quite a while and while it is impossible to control reflections there and keep a functional car I was impressed with what I got. I wanted to enlarge that space and have thought about adding the late ~25 ms reflection with speakers in the back. It never happened though, I ran out of channels, my setup is active in the car. As it is a very small car and I didn't want to add even more load on the battery I scrapped that option.
The reason I asked about the group delay at an earlier time (gated ~20 ms) was just curiosity. The button IR Windows controls those settings and all it would take is to set the right window from it's usual 500 to 20ms.
This picture has the IR window settings on top:
I don't expect to see much difference in your well behaved room though. But controlling the first few ms seems key to imaging. With FIR processing I can dail in a lot of different things and can control how wide a sweet spot is.
But the most ambience I got seems to come from no crossover (as I have in my setup) or first order ones. But it has to be in the recording for it to show.
That's why a lot of people chase the audio with full range drivers only, like the Lowther crowd etc.
I try to do it with huge speakers with 25 full range speakers per side in a line array. But I'm not allowed to clutter the room with damping, nor am I free to place the speakers where I want. So I try to make the most out of what I have with some damping, and a lot of processing.
Here's the impulse of my line array:
As you see it is going more horizontal before crossing the zero line compared to yours.
That is the difference between no crossover and a second order crossover without processing. My approach has other flaws though, You'll notice the step starts to rise first before it descents again. That's due to different time arrivals of all those separate sources. But it does bring a lot of ambience within a recording that has it.
So in a way, I'm not convinced we can't point out what's responsible for what. Maybe we just don't know yet where to look. Of coarse we can't measure the amount of imaging one gets...
Even though I still have many reflections to deal with, my lines disappear completely. If I close my eyes I can't point them out. Step by step I'm taking this a bit further, much like you are.
I recognise some parts of the journey you are on. Though I try to do it in a different way I'm after much of the same things you are. So I try to learn from you here, as you've been on quite the journey already.
If there is a correlation between imaging and measurements I'd look to FR, GD and phase to find it. Actually I am looking there
.Phase is one of the steering mechanisms used in the music.
I do get that late reflection you added, I've been into Car audio for quite a while and while it is impossible to control reflections there and keep a functional car I was impressed with what I got. I wanted to enlarge that space and have thought about adding the late ~25 ms reflection with speakers in the back. It never happened though, I ran out of channels, my setup is active in the car. As it is a very small car and I didn't want to add even more load on the battery I scrapped that option.
The reason I asked about the group delay at an earlier time (gated ~20 ms) was just curiosity. The button IR Windows controls those settings and all it would take is to set the right window from it's usual 500 to 20ms.
This picture has the IR window settings on top:
I don't expect to see much difference in your well behaved room though. But controlling the first few ms seems key to imaging. With FIR processing I can dail in a lot of different things and can control how wide a sweet spot is.
But the most ambience I got seems to come from no crossover (as I have in my setup) or first order ones. But it has to be in the recording for it to show.
That's why a lot of people chase the audio with full range drivers only, like the Lowther crowd etc.
I try to do it with huge speakers with 25 full range speakers per side in a line array. But I'm not allowed to clutter the room with damping, nor am I free to place the speakers where I want. So I try to make the most out of what I have with some damping, and a lot of processing.
Here's the impulse of my line array:
As you see it is going more horizontal before crossing the zero line compared to yours.
That is the difference between no crossover and a second order crossover without processing. My approach has other flaws though, You'll notice the step starts to rise first before it descents again. That's due to different time arrivals of all those separate sources. But it does bring a lot of ambience within a recording that has it.
So in a way, I'm not convinced we can't point out what's responsible for what. Maybe we just don't know yet where to look. Of coarse we can't measure the amount of imaging one gets...
Even though I still have many reflections to deal with, my lines disappear completely. If I close my eyes I can't point them out. Step by step I'm taking this a bit further, much like you are.
Rather than piecemeal this, you can go here and open the mdat and look at whatever you want:
https://www.gearslutz.com/board/11060591-post473.html
https://www.gearslutz.com/board/11060591-post473.html
I suppose you could build your speakers into a crescent shape so all the time arrivals are the same.
Good point concerning the inherent ambiance in a recording. I suppose there are two schools of thought here.
1) The more purest approach whereby your only concerned with preserving the original recording.
2) Where your using room acoustics to add ambiance to a recording that it may not otherwise have.
In a sense, 2) is always the case in ones typical listening space given the room cannot be totally eliminated from what you hear, which makes 1) in absolute terms impossible.
But depending on room criteria, how much you add and in what way is something you can either design into it or eliminate as much as possible.
I basically feel everything we hear is measurable. But I agree that interpretation and knowing where to look is key.
I honestly just want it to sound good! But I also like to learn as much from it as I can.
I chose this type of speakers for several reasons, space being one of them, high efficiency another and less floor/ceiling reflection as a reason to have to fight it less.
An arced array would have different problems, and not give the wide sweet spot I want (for my family for movies etc.). But you'd also have trouble again with reflections.
Lots more into that story but not that important to share here.
Look at: http://www.diyaudio.com/forums/full-range/242171-making-two-towers-25-driver-full-range-line-array.html if you want to know more.
I agree with you totally on the measurements. Enjoyed the read on Gearslutz and recognise the rationale and process.
I don't think I'll ever be really done either.
I chose this type of speakers for several reasons, space being one of them, high efficiency another and less floor/ceiling reflection as a reason to have to fight it less.
An arced array would have different problems, and not give the wide sweet spot I want (for my family for movies etc.). But you'd also have trouble again with reflections.
Lots more into that story but not that important to share here.
Look at: http://www.diyaudio.com/forums/full-range/242171-making-two-towers-25-driver-full-range-line-array.html if you want to know more.
I agree with you totally on the measurements. Enjoyed the read on Gearslutz and recognise the rationale and process.
I don't think I'll ever be really done either
.
Thanks for that, not often that I can take a peek in a room like that
Link to "stereo" mdats:
https://www.gearslutz.com/board/11060860-post474.html
Cool, that'll give us the opertunity to listen to your room
And provide a DRC corrected test track.
That's a nice test case for your 4 cycle window if you ask me Greg!
It did cross my mind though, but we'd have to supply a flac back so Jim can burn it to a CD and play.
So Jim, do you have a favourite track you would like to listen too with FIR correction applied?
Should be a fun test! I'd say that room wouldn't need anything longer than that.
Greg, in the past I've seen you do a predicted IR based on the IR response from a measurement. Could you do that with this?
Or tell us how you did it?
It did cross my mind though, but we'd have to supply a flac back so Jim can burn it to a CD and play.
So Jim, do you have a favourite track you would like to listen too with FIR correction applied?
Should be a fun test! I'd say that room wouldn't need anything longer than that.
Greg, in the past I've seen you do a predicted IR based on the IR response from a measurement. Could you do that with this?
Or tell us how you did it?
Last edited:
Well X, I did take a look at the FR and decay... the decay is awesome, something I'll never get in my room. The FR could use a bit of FIR help. But I did notice flat FR isn't everything. As long as the general trend is OK, that's why I can listen with pleasure a bit of center of my listening sweet spot without being annoyed. FR is a bit worse there but in listening it isn't an immediate problem. It could help imaging though, if there is unbalance between left and right. But seeing the prior work Jim did I think it will be quite close already. We'll see if Jim is up for it .
.
Finite impulse response
as opposed to
Infinite impulse response (IIR) which is usual filters used
Digital EQ really, but one that doesn't mess up your phase but actually can correct part of it.
You can create FIR filters with a left and right IR taken at the listening position.
This should explain it a bit:
The upper FR is a raw measurement in my room of the array.
The second is the FIR correction, based on measurement data
The third is a prediction of the resultant FR
The last one an actual measurement with FIR correction applied. I play digital from a PC so I convolve (read correct with FIR filter) each song that I play on the fly. It creates a very small delay though. Basically it is similar to what those guys used to listen to your room. One can also add data from another room and play back the characteristics of that room in your own space. It's a brave new world out there
.Besides the FR correction it also tries to correct time behaviour. With your supplied measurements we should be able to "correct" any song you like and get it to you to play on your system. A sort of off line convolution.
Last edited:
OK
Someone map out for me what you want me to do, and what we hope to accomplish by doing it.
Edit: ok, see you did that.
But what is the end game? I cant take my 1200 CD's and reburn them all. But I am up for a couple songs for experimentation sake. I guess I send/post some flac files to you guys, you do your FIR thing to them, and you send them back to me?
Edit2: Lets do this. You guys listen both ways, with and without FIR. If you think the difference is noteworthy, ill get you some flac files.
Someone map out for me what you want me to do, and what we hope to accomplish by doing it.
Edit: ok, see you did that.
But what is the end game? I cant take my 1200 CD's and reburn them all. But I am up for a couple songs for experimentation sake. I guess I send/post some flac files to you guys, you do your FIR thing to them, and you send them back to me?
Edit2: Lets do this. You guys listen both ways, with and without FIR. If you think the difference is noteworthy, ill get you some flac files.
Last edited:
That would be a way of doing it. Just as a test. If you like it you get to decide what you want to do next (lol). You might actually really like it. I know I do.
I play everything trough the computer. But hey, I'm a ICT manager by career. no wonder I like to play around with that. But slowly more and more people look into this and find out what it can do.
Greg here has been doing it for a while and is coaching us to get the results going. He's by far the most experienced on this thread. I know just enough to be dangerous.
There's even a miniDSP that can do the FIR correction process for you as a stand alone device. Play the music trough it from for example CD, let miniDSP correct it and send to DAC etc.
But that would have to go in the chain before the amp and control your mains. But we're not even close to that yet. Just see if you like it by doing a little experimenting.
I play everything trough the computer. But hey, I'm a ICT manager by career. no wonder I like to play around with that. But slowly more and more people look into this and find out what it can do.
Greg here has been doing it for a while and is coaching us to get the results going. He's by far the most experienced on this thread. I know just enough to be dangerous
.There's even a miniDSP that can do the FIR correction process for you as a stand alone device. Play the music trough it from for example CD, let miniDSP correct it and send to DAC etc.
But that would have to go in the chain before the amp and control your mains. But we're not even close to that yet. Just see if you like it by doing a little experimenting.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.