I'm on it
but ant seem to find much. is this what you are refering to https://www.jblpro.com/ProductAttachments/AES_Ureda_Analysis_of_Line_Arrays.pdf
yes, we definitely want reflection of over 30ms, its just the early reflection (within 20ms-25ms) that we dont want
I don't know where you get your numbers from but in my opinion it would be very bad to introduce reflections after 30 ms. I deliberately chose 20 ms. The direction of it matters too. Look up: "Haas Kicker" to find out why. Introducing later reflections than 30 ms would probably make you able to notice them.
You don't hear a 20 ms reflection coming from ~120 to 150 degree if straight ahead is considered 0 degree. But it does give you a sense of envelopment. Like being in a much bigger space. As my reflections are created virtually with ambient speakers that I aim at surfaces I can turn it on and off. I prefer to have them on. For an excellent example without using extra speakers, again look at Jim's room. His are set to 24 ms.
The reason for me to chose 20 ms is that it coincides with a faint reflection that was there already.
I tried to prevent all reflections in the first 20 ms. Get them down as far as I could within my space. My APL_TDA plot shows you that effort:
Run the demo version of it to check your results... All you need to do is make that measurement to get more info about the in room performance. I still see some reflections, and I know where they come from. That's the beauty if you spend the time learning to "read the room". Look at Jim's plots to see the difference. I convinced him to play along and make a TDA plot. You clearly see the effort he put into that room.
Last edited:
Thanks for chiming in Jim, your paper was of much help to me when I was trying to figure out if this could work at all. I know I chose a different set of compromises but as long as we realise it's always going to be some sort of compromise, you can make it work the best you can. I'm still shocked at what this set of compromises can do! I think that shows in my enthusiasm.
I'll give you a glance at what I collected over the years. I'm still collecting and reading though. It will probably never end. But this might give you an idea of my devotion:
It's only part of it as each folder is filled with papers as well and we only see up to "L". If I had scrolled down to "N" you would have seen Jim's paper.
Much more than this was read, but this is to give you a glance at what I studied and saved for future reference. You do see the "Line Arrays Part 1" etc. papers that I referenced in an earlier post, those are available right here on the forum by searching for the author: speaker dave (David L. Smith).
I'm probably not the only lunatic on here collecting and reading but I have to say it helps immensely to be able to make up your own mind. This thread is my way to give back what I've learned from it so far. Not to show off, just for sharing information. The power of this forum is to help each other advance.
At least, that's how I like to see it.
The diversity of papers shows my general interest in this subject.
I didn't mean to skip your question. The drivers in my array are 5 drivers wired in series and 5 groups like that in parallel. You end up with an impedance curve resembling that of a single driver.
Here's a graph of the impedance (and a corrected curve with a conjugate network) Part of that conjugate network is still in use, even though I haven't found prove as to why I like having it in there. Yes it measures slightly different. But after correction with FIR it gets hard to notice it in graphs. Yet it got me in listening. I dropped the correction for the climbing part toward high frequencies and kept the correction for the resonance peak:
Don't ask me why... I can't answer that yet.
The correction for the climbing curve toward the high end did not convince me at all. Smoother sound without it. I guess the copper rings in the Vifa do a good enough job without needing extra help.An in wall array should work, if you know where to put them to prevent early reflections. Look at ra7's corner arrays: http://www.diyaudio.com/forums/multi-way/284371-corner-floor-ceiling-line-array-using-vifa-tc9.html
Hardly take up any space and nearly invisible. Cover them with grille cloth and use separate subs for the bottom end. Just a suggestion
.Those speakers in your link look like a fashion statement. Fun, pretty and... are they any good? I don't know
.
Last edited:
What I meant is that EARLY reflections (in relation to the sweet spot) must be absorbed and ideally -10 to -20db within 20ms, but that you absolutely want to let secondary reflections. So you dont overly treat your room by placing too much panels which would result in a overly dead room ime. You only treat the early reflections and let the secondary reflection.
"You don't hear a 20 ms reflection coming from ~120 to 150 degree if straight ahead is considered 0 degree. "
Exactly, these are secondary reflections and are not detrimental to SQ. Only early reflections are detrimental.
That software seem very interesting for measurements. I dont have a good mic right now but will try it once I get one.
I'm
"You don't hear a 20 ms reflection coming from ~120 to 150 degree if straight ahead is considered 0 degree. "
Exactly, these are secondary reflections and are not detrimental to SQ. Only early reflections are detrimental.
I don't know where you get your numbers from but in my opinion it would be very bad to introduce reflections after 30 ms. I deliberately chose 20 ms. The direction of it matters too. Look up: "Haas Kicker" to find out why. Introducing later reflections than 30 ms would probably make you able to notice them.
You don't hear a 20 ms reflection coming from ~120 to 150 degree if straight ahead is considered 0 degree. But it does give you a sense of envelopment. Like being in a much bigger space. As my reflections are created virtually with ambient speakers that I aim at surfaces I can turn it on and off. I prefer to have them on. For an excellent example without using extra speakers, again look at Jim's room. His are set to 24 ms.
The reason for me to chose 20 ms is that it coincides with a faint reflection that was there already.
I tried to prevent all reflections in the first 20 ms. Get them down as far as I could within my space. My APL_TDA plot shows you that effort:
Run the demo version of it to check your results... All you need to do is make that measurement to get more info about the in room performance. I still see some reflections, and I know where they come from. That's the beauty if you spend the time learning to "read the room". Look at Jim's plots to see the difference. I convinced him to play along and make a TDA plot. You clearly see the effort he put into that room.
That software seem very interesting for measurements. I dont have a good mic right now but will try it once I get one.
I'm
Your images explain much. I was wondering about the impedence of 25 speakers in parallel!
From what I have read, line arrays produce a great stereo footprint.
How well does in produce live band sounds? Utlimately that is what we want isn't it?
Something I have always noticed about live music and reproduced music is the difference in the power of the sound coming from the live instruments, and sometimes the frequencies.
Let me explain: Listening to a live band I hear the following:
1. Bass - usually a low rumble no problem here, no definition.
2. Cymbals. Cymbals always sound wrong. I have actualy been given the opportunity to try out playing a set of drums. The cymbals, high hat, etc have a thick, full sound to them not the drum-machine type tinkle that is heard on cheaper systems.
3. Drums: this is the acid test: drums have, in live music, a sharp, powerful attack stage that I have never heard reproduced by a system. Expecially the kick drum
4. Guitar - a full blooded guitar sound is I think one of the least challenging to reproduce.
5. Vocals - are fed through a mike and speakers, so this presents the least of a challenge, but vocals should be clear and smooth without any raspiness or crackle.
6. Violin - the lower end of the violin seems to be missing in much of the music I listen to. Same for the string section, where well defined powerful low ends are needed.
These are just my impressions - if correct, how does your system produce these sounds?
I listened to a live band recently playing at 90db at 15 metres away and I despaired that I would ever be able to bring that sound home.
From what I have read, line arrays produce a great stereo footprint.
How well does in produce live band sounds? Utlimately that is what we want isn't it?
Something I have always noticed about live music and reproduced music is the difference in the power of the sound coming from the live instruments, and sometimes the frequencies.
Let me explain: Listening to a live band I hear the following:
1. Bass - usually a low rumble no problem here, no definition.
2. Cymbals. Cymbals always sound wrong. I have actualy been given the opportunity to try out playing a set of drums. The cymbals, high hat, etc have a thick, full sound to them not the drum-machine type tinkle that is heard on cheaper systems.
3. Drums: this is the acid test: drums have, in live music, a sharp, powerful attack stage that I have never heard reproduced by a system. Expecially the kick drum
4. Guitar - a full blooded guitar sound is I think one of the least challenging to reproduce.
5. Vocals - are fed through a mike and speakers, so this presents the least of a challenge, but vocals should be clear and smooth without any raspiness or crackle.
6. Violin - the lower end of the violin seems to be missing in much of the music I listen to. Same for the string section, where well defined powerful low ends are needed.
These are just my impressions - if correct, how does your system produce these sounds?
I listened to a live band recently playing at 90db at 15 metres away and I despaired that I would ever be able to bring that sound home.
Yes that would be the ultimate goal...
My goal was to get "live" like sound. But as correct as possible. Hard to explain in words but judging from Jan Fekkes review I'm well on my way.
Live recordings do differ though in quality. But even some CD studio material can have "live like" qualities. I was talking to Jan about that. Some songs like "Ashes to Ashes" from Bowie (RIP) have a huge studio presence in how they build up that song. Impossible to bring that to a stage.
Other bands like Opeth, but even more Rival Sons feel like they recorded it in one take playing "live" and adding a few bits and pieces later on perhaps. But it still has that "live" feel. That is so thrilling to hear you get the sense of being there.
To try and answer your questions;
For bass, the DSP processing helps not to let bass dominate the room. You can still follow the notes down low. But they are there, you feel them too.
That made a clear difference in a lot of songs.. being able to hear the bass groove under the songs surprised quite a lot of people listening here.
Cymbals, I've heard studio monitors recently and I couldn't believe the amount of sizzle... that was just too much for me and not the sound I remember from being close to drums. Some people may find my rendition dull in comparison with the missing sizzle while I experience this as more correct, full bodied. It can even make you flinch your eyes with a sharp attack.
One thing I do try to avoid is the Tizzz and Boom sound. That's what a lesser system sounds to me (even the high dollar audiophile systems often suffer from that type of sound). It should be a more natural presentation than that.
Drums, I worked hard on that part actually. One of the goals to get right. Most people think the attack from the drum is down low, it isn't. That defined punch from a kick drum comes from higher up in frequency but the part down low is definitely important too, the timing is key here. One of the reasons for me to strive for time coherency. I do think I've got it right. I was missing out for a long time until I managed to get my 100-200 Hz arrive in time. Without correcting the timing the kick just wasn't there even if the SPL level was correct.
Missing floor and ceiling reflections do help to get that right i.m.h.o.
Guitar, you say easy to get right... I'm less easy to please as I'm a big guitar fan. Playing some guitar myself, though I'm not actually any good at it as I do have high standards. I love the sound of guitar, and it's quite more demanding to get right than one might think. I actually learned something from this post: http://www.diyaudio.com/forums/everything-else/284131-how-listen-7.html#post4561570.
That explains that even simple acoustic guitar is harder to reproduce right than one may think. For me it all falls back to time coherency to get it to sound true to life.
Vocals, I hear a lot of different vocal qualities in songs, even from the same album if we talk about recorded music, not a live performance. Some are crystal clear, easy to follow, others are more mumbled, hard to follow, even on the same album. I spend a lot of time to figure out what was causing that. Probably different techniques in recording, mixing and mastering.
The thread: http://www.diyaudio.com/forums/multi-way/277519-fixing-stereo-phantom-center.html shed some light on that. I spend some time figuring out how cross talk could indeed reduce focal clarity. After playing with cross talk techniques and seeing the effect it had in frequency response I tried 2 simple EQ cuts that were evident in the crosstalk results. One centred at 3700 Hz and one more, a little less deep at 7400 Hz.(*)
Won't work universally, works for me figuring out the distance to the speakers and the size of an average head. These are not deep cuts, but they did clear up the more mumbled vocals. The clear vocals remained clear as before. So in my opinion the cross talk in Stereo is an issue. I also adopted an S-curve from that same thread. Meaning a bit of mid-side processing related to head shading. The combination with the dips I mentioned made me more pleased with vocal clarity and kept it consistent from left to centre to right.
Violin, I'll admit knowing to little about the true sound to say anything useful about it. But I can think of reasons why you feel that way. I bet if you solve early reflections from floor and ceiling (as well as the others) things will fall more into place.
Now it may sound I have an answer for everything here. And I'm convinced I'm on my way to better sound with the things I have done so far. But there's always the matter of taste coming in. I adjust my sound to what pleases me. Yes I made sure the timing is right. But the tonal balance is adjusted to taste. Oddly enough I ended up with a curve in my room within a few dB or less from Mitchba, with him using completely different speakers, though the same kind of recipe as far as processing (for time coherency) and battling first reflections go.
(I do add some other things like ambient reflections for my musical pleasure and my mid/side processing is also a deviation compare to Mitch's approach)
Dark red is Mitch, light red is me... We did not copy each other here. We each on our own looked for the best tonal balance working on a lot of different source material. It's remarkable how close our curves ended up though. But still I do think the room might be of influence to find your right curve as well as the distance to your speakers.
(*) = the cuts are mostly done in the mid channel of mid-side processing. The side channels do not have the same problem to that extend.
The mid-side processing I do might discomfort others. To me it makes perfect sense to play with it after reading up on the theory. But even though it makes perfect sense to me doesn't mean it will work equally well for others.
That's what I mean by the taste part. The above room curve also has an SPL range for which it works best. I think I play at 85 to 90 dB average when I'm alone. That's pretty loud in a room but not concert levels. The SPL does make a huge difference in the presentation of the tonal balance and placement of the image. Playing louder is like a zoom function. The centre vocal comes forward towards you. Again: absence of early reflections is key here.
These line arrays may not be for everybody. But for me they are the best thing I've heard so far in music reproduction. They are pushing all of my buttons. Especially after all the work I put into it to get it right for me.
The amount of different sounds I've heard my speakers make is staggering. I learned a lot just by analysing all of the differences. What I have now is a sort of "Best off" of those different sounds.
I do not regret one minute of that journey.
This post needs a disclaimer: we are deeply into subjective performance here. This is how I "feel" it must perform for me. I have done my best to stay objective in my posts, though I have swayed from that many times due to sheer enthusiasm. The above post explains what I was after. But I cannot determine what you guys would want a system to do. Our own interpretation of sound may differ from one another. Me, I'm passionate about music. Convinced the true glory or magic is in the songs, I saw it as a challenge to bring out that magic with the help of science. I managed to please myself and quite a few listeners so far. But that still doesn't make it the universal solution.
But it sure is satisfying to see people connect with the music on an emotional level. To me it means I'm doing something right!
(sorry for the long post)
My goal was to get "live" like sound. But as correct as possible. Hard to explain in words but judging from Jan Fekkes review I'm well on my way.
Live recordings do differ though in quality. But even some CD studio material can have "live like" qualities. I was talking to Jan about that. Some songs like "Ashes to Ashes" from Bowie (RIP) have a huge studio presence in how they build up that song. Impossible to bring that to a stage.
Other bands like Opeth, but even more Rival Sons feel like they recorded it in one take playing "live" and adding a few bits and pieces later on perhaps. But it still has that "live" feel. That is so thrilling to hear you get the sense of being there.
To try and answer your questions;
For bass, the DSP processing helps not to let bass dominate the room. You can still follow the notes down low. But they are there, you feel them too.
That made a clear difference in a lot of songs.. being able to hear the bass groove under the songs surprised quite a lot of people listening here.
Cymbals, I've heard studio monitors recently and I couldn't believe the amount of sizzle... that was just too much for me and not the sound I remember from being close to drums. Some people may find my rendition dull in comparison with the missing sizzle while I experience this as more correct, full bodied. It can even make you flinch your eyes with a sharp attack.
One thing I do try to avoid is the Tizzz and Boom sound. That's what a lesser system sounds to me (even the high dollar audiophile systems often suffer from that type of sound). It should be a more natural presentation than that.
Drums, I worked hard on that part actually. One of the goals to get right. Most people think the attack from the drum is down low, it isn't. That defined punch from a kick drum comes from higher up in frequency but the part down low is definitely important too, the timing is key here. One of the reasons for me to strive for time coherency. I do think I've got it right. I was missing out for a long time until I managed to get my 100-200 Hz arrive in time. Without correcting the timing the kick just wasn't there even if the SPL level was correct.
Missing floor and ceiling reflections do help to get that right i.m.h.o.
Guitar, you say easy to get right... I'm less easy to please as I'm a big guitar fan. Playing some guitar myself, though I'm not actually any good at it as I do have high standards. I love the sound of guitar, and it's quite more demanding to get right than one might think. I actually learned something from this post: http://www.diyaudio.com/forums/everything-else/284131-how-listen-7.html#post4561570.
That explains that even simple acoustic guitar is harder to reproduce right than one may think. For me it all falls back to time coherency to get it to sound true to life.
Vocals, I hear a lot of different vocal qualities in songs, even from the same album if we talk about recorded music, not a live performance. Some are crystal clear, easy to follow, others are more mumbled, hard to follow, even on the same album. I spend a lot of time to figure out what was causing that. Probably different techniques in recording, mixing and mastering.
The thread: http://www.diyaudio.com/forums/multi-way/277519-fixing-stereo-phantom-center.html shed some light on that. I spend some time figuring out how cross talk could indeed reduce focal clarity. After playing with cross talk techniques and seeing the effect it had in frequency response I tried 2 simple EQ cuts that were evident in the crosstalk results. One centred at 3700 Hz and one more, a little less deep at 7400 Hz.(*)
Won't work universally, works for me figuring out the distance to the speakers and the size of an average head. These are not deep cuts, but they did clear up the more mumbled vocals. The clear vocals remained clear as before. So in my opinion the cross talk in Stereo is an issue. I also adopted an S-curve from that same thread. Meaning a bit of mid-side processing related to head shading. The combination with the dips I mentioned made me more pleased with vocal clarity and kept it consistent from left to centre to right.
Violin, I'll admit knowing to little about the true sound to say anything useful about it. But I can think of reasons why you feel that way. I bet if you solve early reflections from floor and ceiling (as well as the others) things will fall more into place.
Now it may sound I have an answer for everything here. And I'm convinced I'm on my way to better sound with the things I have done so far. But there's always the matter of taste coming in. I adjust my sound to what pleases me. Yes I made sure the timing is right. But the tonal balance is adjusted to taste. Oddly enough I ended up with a curve in my room within a few dB or less from Mitchba, with him using completely different speakers, though the same kind of recipe as far as processing (for time coherency) and battling first reflections go.
(I do add some other things like ambient reflections for my musical pleasure and my mid/side processing is also a deviation compare to Mitch's approach)
Dark red is Mitch, light red is me... We did not copy each other here. We each on our own looked for the best tonal balance working on a lot of different source material. It's remarkable how close our curves ended up though. But still I do think the room might be of influence to find your right curve as well as the distance to your speakers.
(*) = the cuts are mostly done in the mid channel of mid-side processing. The side channels do not have the same problem to that extend.
The mid-side processing I do might discomfort others. To me it makes perfect sense to play with it after reading up on the theory. But even though it makes perfect sense to me doesn't mean it will work equally well for others.
That's what I mean by the taste part. The above room curve also has an SPL range for which it works best. I think I play at 85 to 90 dB average when I'm alone. That's pretty loud in a room but not concert levels. The SPL does make a huge difference in the presentation of the tonal balance and placement of the image. Playing louder is like a zoom function. The centre vocal comes forward towards you. Again: absence of early reflections is key here.
These line arrays may not be for everybody. But for me they are the best thing I've heard so far in music reproduction. They are pushing all of my buttons. Especially after all the work I put into it to get it right for me.
The amount of different sounds I've heard my speakers make is staggering. I learned a lot just by analysing all of the differences. What I have now is a sort of "Best off" of those different sounds.
I do not regret one minute of that journey.
This post needs a disclaimer: we are deeply into subjective performance here. This is how I "feel" it must perform for me. I have done my best to stay objective in my posts, though I have swayed from that many times due to sheer enthusiasm. The above post explains what I was after. But I cannot determine what you guys would want a system to do. Our own interpretation of sound may differ from one another. Me, I'm passionate about music. Convinced the true glory or magic is in the songs, I saw it as a challenge to bring out that magic with the help of science. I managed to please myself and quite a few listeners so far. But that still doesn't make it the universal solution.
But it sure is satisfying to see people connect with the music on an emotional level. To me it means I'm doing something right!
(sorry for the long post)
Last edited:
I have a couple of tracks to test speakers that have predominant violins on them.
Violins are one of the hardest instruments to reproduce well, I believe.
The combination of horse's hair rubbing against tightly wound up strings on a very small and resonant frame that project sounds like nothing else...
One of the description that I remember reading about bad violins' reproduction is that they sounded like... syrup.
For most speakers, it's easy for the sound of violins to get melted into a viscous blob of smeared sounding "syrup".
Very few speakers will reproduce all the nuances of a well performed violin.
I have no doubt that 50 little Vifa drivers are capable of doing exactly just that. I was almost there with my 32 cheaper line arrays...
Come to think of it, I didn't play any violin tracks with the pair that Byrtt sent me... will do that!
Violins are one of the hardest instruments to reproduce well, I believe.
The combination of horse's hair rubbing against tightly wound up strings on a very small and resonant frame that project sounds like nothing else...
One of the description that I remember reading about bad violins' reproduction is that they sounded like... syrup.
For most speakers, it's easy for the sound of violins to get melted into a viscous blob of smeared sounding "syrup".
Very few speakers will reproduce all the nuances of a well performed violin.
I have no doubt that 50 little Vifa drivers are capable of doing exactly just that. I was almost there with my 32 cheaper line arrays...
Come to think of it, I didn't play any violin tracks with the pair that Byrtt sent me... will do that!
Great long post. Somethings I do not understand, but I was really pleased to find agreement on come points.
What I have read about arrays is that they present a higher SPL throughout the room. A greater distance. And a kick drum puts out 105 dB at 1 foot (can't find the reference, it was in a book ).
This is really difficult to achieve at lower sound levels then, my goal would be a system that can play really well at low sound levels, say 75 dB SPL . BTW the live band was about 90dB where I stood, and I could hear everything.
I hear what you say about voice. What is really annoying me now is the difference in recording quality of some CD's - even with my incredibly basic set up.
I totally agree on the cymbals. It is when I hear a live band playing cymbals that I realize that silly artificial sizzle is really out of place.
How - through DSP then?
What I have read about arrays is that they present a higher SPL throughout the room. A greater distance. And a kick drum puts out 105 dB at 1 foot (can't find the reference, it was in a book ).
This is really difficult to achieve at lower sound levels then, my goal would be a system that can play really well at low sound levels, say 75 dB SPL . BTW the live band was about 90dB where I stood, and I could hear everything.
I hear what you say about voice. What is really annoying me now is the difference in recording quality of some CD's - even with my incredibly basic set up.
Sample sound tracks - guitar, violins, etc
When I finally get my speakers done, there are some sample tracks that I plan to use to test them. Does it sound like a ... is the question.
Will post link when I can.
When I finally get my speakers done, there are some sample tracks that I plan to use to test them. Does it sound like a ... is the question.
Will post link when I can.
Glad it helped...
No argument there
Yes, this issue I solved with DSP. You cannot force such a correction and expect great things. You have to look at what is happening out in the room on more than one spot. Else the DSP solution might correct it at the listening position but make things worse elsewhere in the room.
To make it visible, this is the result without any correction:
It's a combined response of left and right and not representative of the separate left and right problems, but it does show the timing being off in the low registers. This still gave the drums their sound, as the SPL was there, but the kick was missing.
I've spend quite a bit of time, thinking about the problem and coming up with a solution, watching what happens over time in measurements.
After implementing my "correction" the kick was there again, as it should be.
There's still influences of the room in both left and right channel separately, but the combined result looks excellent.
This isn't a substitute for room treatment though. You simply cannot fix everything with DSP. Due to the way line arrays interact with the room there's less that needs fixing. The floor and ceiling bounces/reflections are quite a big problem in more traditional speakers and would need to be addressed. A big plus for the way line arrays, with a floor to ceiling size, interact with the room.
This is really difficult to achieve at lower sound levels then, my goal would be a system that can play really well at low sound levels, say 75 dB SPL . BTW the live band was about 90dB where I stood, and I could hear everything.
Like I said in my reply, the room curve and SPL level kind of fit together. Volume then becomes a sort of zoom dial.
At lower average SPL levels you might want to change the room target curve. That's where the Fletcher Munson curve comes in. It's not that you can't enjoy music played back at a lower level. At least that's where I would look to get a balanced sound even at lower levels.
To listen to this system as background music I have to turn it down by quite a bit. I do enjoy listening at the louder level. I kept an SPL meter close for a long while as you only realise how loud you are playing by trying to start a conversation. The seemingly absence of distortion even at louder levels lets you turn it up without the sound falling apart. It makes it easy to listen too loud.
Yes, recordings can be a mixed bag. But in general most music is actually very pleasing. I also want to note the difference between mixing and mastering.
Mastering happens at a certain SPL level, which isn't exactly written in stone. So to hear that music as it was meant to sound you should know that intended level of SPL to get the true picture. I've noticed different mastering's from the same album/CD for different area's like Japan, Europe or the US, and noticing quite big level changes in bass between them. Those tiny or sometimes even bigger differences can really make or break the sound.
I wasn't satisfied with any Led Zeppelin album in digital form until the recent 2015 remasters. I still had the vinyl version "ringing" in my head that felt different to me. Luckily that is remedied. For me, the same goes for Van Halen. I only have one CD that feels true to the Vinyl, that's the Steve Hoffman remastered debut CD. The rest all comes up short on bass for me. That's a mastering choice, nothing wrong with the mix. Although having Ed on the left all the time does get old. The instrumental intro's and pieces are way more enjoyable where more than one guitar is used. Just my subjective observations of 2 bands I love for entirely different reasons.
To come back to the subject of violins, I've heard them on my system. I just meant I lack the true reference of an up close violin being played. I've heard them in a concert hall, that sound seems to translate quite well in my room. But to really judge I'd have to have a way better reference. When I actually was in that concert hall with the orchestra playing (a number of times actually) I was paying more attention to Steve Vai, who was conducting it all to later on join them on guitar. That gives away my preference in music again. I lack a decent classical music upbringing. I tried it, but it didn't stick. The closest I come to liking big orchestra's is works like John Barry and movie scores from people like Hans Zimmer
. Especially when they involve Rodrigo & Gabriella.
Last edited:
Wesayso,
I know this question really belongs over in a dsp thread but I have been reading and following along and have a question. I see in your last post the before and after graphs you just put up. I know you didn't change the room or the speakers so I keep asking myself how you clean up the long decay times I see in the first graph that becomes the second much cleaner decay times? Are you doing a windowing function that changes the actual digital output signal in the lower octaves, how is that accomplished? Is that what you talk about when you talk about the cycle count, I feel so uninformed on how all the dsp functions are working, (bowed Head) so unworthy!
I know this question really belongs over in a dsp thread but I have been reading and following along and have a question. I see in your last post the before and after graphs you just put up. I know you didn't change the room or the speakers so I keep asking myself how you clean up the long decay times I see in the first graph that becomes the second much cleaner decay times? Are you doing a windowing function that changes the actual digital output signal in the lower octaves, how is that accomplished? Is that what you talk about when you talk about the cycle count, I feel so uninformed on how all the dsp functions are working, (bowed Head) so unworthy!
I figured this question might come up after seeing the graphs .
The cycles you always see me talk about is the windowing used to base the correction upon. So to EQ I use a frequency dependant window with certain setting that I deem necessary.
The correction in time is phase manipulation. Much like the one you can do with RePhase. For this correction you can set different parameters within DRC to control it's behaviour(*). If I show REW measurements from the left side it might show this more clearly:
In this plot you can see the delay at ~140 Hz. First I had to make sure I was seeing similar delays at different seating positions. Meaning it must be something that's the same at these different positions to be able to correct for it. For that 140 Hz delay this was true. But not for the second delay at ~72 Hz.
So I chose settings with it's own frequency dependant window to repair the 140 Hz delay, but not the ~72 Hz one.
There are programs that can do the puzzling for you, based on multiple measurements to find the same spots to fix and spots to ignore like I mentioned above. I did it by hand, by looking at what happened over time in that 140 Hz dip, knowing it was a common issue over a wider area.
After the fix the graph looked like this:
You still see the problem area at ~72 Hz, but it did move in time. There's still a dip there that I don't fix. But the first arriving sound wave has been moved to minimum phase behaviour. The right side is a bit different, and has problem areas in different FR ranges due the asymmetric room. But having a dip in the left response made me put a small rise in the right to make up for that as it's below 80 Hz and shouldn't be noticeable in Stereo. It's only a few dB's anyway.
Can't do tricks like that higher up in FR but in this part it still works quite well.
Here's the FR of left/right and stereo to show what I mean.
Hope this helps to shed some light on what is happening over time. The short answer would be: phase manipulation in the FIR correction. I'm kind of nibbling time off of the excess phase as measured to get it back to minimum phase and thus changing the input signal to do that. Similar things can be done with RePhase but the same warnings do apply. You can't fix everything with DSP or EQ as you know. But it's still pretty remarkable what can be done.
Of coarse I also tried fixing that 72 Hz hole, knowing I really shouldn't. Moving around the listening spot left to right while listening to music made it very clear and obvious that was a very bad idea. Looked great in plots though.
Still do a couple of other tricks but this should hopefully paint a clear picture of what is happening. The people using RePhase kind of do the same. If they are careful enough to know what part to correct for and what to leave alone. Can't stress that part hard enough.
Doing it the wrong way or making the wrong choices might actually give this kind of DSP manipulation a bad name. It's much easier to mess it up than to get right. As with all things I guess... you need to know what to look at and where to start looking.
(*) = I use settings outside of / and different from the intended scope of DRC, with a custom template tailored to my choices. I love flexible programs like that.
.The cycles you always see me talk about is the windowing used to base the correction upon. So to EQ I use a frequency dependant window with certain setting that I deem necessary.
The correction in time is phase manipulation. Much like the one you can do with RePhase. For this correction you can set different parameters within DRC to control it's behaviour(*). If I show REW measurements from the left side it might show this more clearly:
In this plot you can see the delay at ~140 Hz. First I had to make sure I was seeing similar delays at different seating positions. Meaning it must be something that's the same at these different positions to be able to correct for it. For that 140 Hz delay this was true. But not for the second delay at ~72 Hz.
So I chose settings with it's own frequency dependant window to repair the 140 Hz delay, but not the ~72 Hz one.
There are programs that can do the puzzling for you, based on multiple measurements to find the same spots to fix and spots to ignore like I mentioned above. I did it by hand, by looking at what happened over time in that 140 Hz dip, knowing it was a common issue over a wider area.
After the fix the graph looked like this:
You still see the problem area at ~72 Hz, but it did move in time. There's still a dip there that I don't fix. But the first arriving sound wave has been moved to minimum phase behaviour. The right side is a bit different, and has problem areas in different FR ranges due the asymmetric room. But having a dip in the left response made me put a small rise in the right to make up for that as it's below 80 Hz and shouldn't be noticeable in Stereo. It's only a few dB's anyway.
Can't do tricks like that higher up in FR but in this part it still works quite well.
Here's the FR of left/right and stereo to show what I mean.
Hope this helps to shed some light on what is happening over time. The short answer would be: phase manipulation in the FIR correction. I'm kind of nibbling time off of the excess phase as measured to get it back to minimum phase and thus changing the input signal to do that. Similar things can be done with RePhase but the same warnings do apply. You can't fix everything with DSP or EQ as you know. But it's still pretty remarkable what can be done.
Of coarse I also tried fixing that 72 Hz hole, knowing I really shouldn't. Moving around the listening spot left to right while listening to music made it very clear and obvious that was a very bad idea. Looked great in plots though
.Still do a couple of other tricks but this should hopefully paint a clear picture of what is happening. The people using RePhase kind of do the same. If they are careful enough to know what part to correct for and what to leave alone. Can't stress that part hard enough.
Doing it the wrong way or making the wrong choices might actually give this kind of DSP manipulation a bad name. It's much easier to mess it up than to get right. As with all things I guess... you need to know what to look at and where to start looking
.(*) = I use settings outside of / and different from the intended scope of DRC, with a custom template tailored to my choices. I love flexible programs like that.
Thanks for the explanation Wesayso. I do understand what you are saying. I guess I would be looking at this more in the sense of the speaker in isolation not knowing what room or placement someone would put my speakers. Two very different situations and needed solutions. I really only have to look at the crossover region to correct any phase anomalies and not what the room is adding with room modes. Auto DRC as I have seen in Pro audio can make as much a mess as make things better, so I follow how you have to be careful with the use of such software control.
Yes indeed. With your speaker the best you can do is make the hand over from one driver to the next as seamless as can be, both in time and frequency. Much like Grim Audio did with their LS1.
A great description of that can be found in their published paper: Grimm Audio paper
I guess you already have that, but for those following this discussion.
A great description of that can be found in their published paper: Grimm Audio paper
I guess you already have that, but for those following this discussion
.How about temperature effects?
http://www.harbeth.co.uk/usergroup/...istening-room-and-why-it-should-be-controlled
Now I need an air-conditioner as well. .
http://www.harbeth.co.uk/usergroup/...istening-room-and-why-it-should-be-controlled
Now I need an air-conditioner as well. .
Last edited:
I guess temperature will have it's effect. Not that much of a problem over here in a living room where the temperature gets adjusted to ~ 20 degree over much of the year. We only have a few weeks in the summer where it is hot. As that doesn't happen all that often we'd be outside way more than usual .
I've noticed changes in sound in my car though, that environment has way more temperature changes throughout the year. My 1982 car does not have air-conditioning.
.I've noticed changes in sound in my car though, that environment has way more temperature changes throughout the year. My 1982 car does not have air-conditioning.
That's something I have often wondered about myself. I definitely get a 3D "picture" with depth. Placement of instruments behind the vocals, in some songs the other way around actually. Some pop songs seem more flat though.
I have listened to a lot of music, some go wide and deep, some songs more in front of the speakers, other material is small and almost completely confined in the centre. Some pop songs have everything in a wide line from left to right. In short it's different on each track I play. To me that's a very good sign. It means I'm not introducing a one trick widening or depth enhancement that's the same on every track you play. Something like that could happen with early reflections for instance.
But I don't know if depth perception would/could be maximised by placing damping panels behind the speakers. I've heard the depth come and go in different settings/processing I have played with. Focussing strongly on having and keeping it of course.
I cannot bring myself to try damping panels behind the speakers as I know I won't be able to keep them if I like them.
In an ideal setup I'd want them there. But do I miss it? Not right now, no. The ambient channels help to strengthen the sense of 3D and depth (as well as give a sense of envelopment). So I do not feel I miss out.
At one point I noticed more depth on higher resolution recordings compared to CD quality. Checked time and again and this result was consistent. I could not figure out why. That is until I noticed part of my processing chain caused a delay on my main signal but not on my ambient channels. When I remedied that the imaging and depth was pretty much comparable again between both formats. What happened (at least I think so) is that high res material caused less delay (faster processing time) on the mains than CD quality, making the time gap between mains and the ambient channels longer than when playing a CD track. I have since introduced the same processing chain on the ambient channels to keep all channels delayed equally. Problem solved. But it shows a difference in timing from the ambient channels does have a pretty big effect on depth perception.
I haven't tried to optimise that perception yet. I can choose the balance and signal that is played trough those speakers and although I have played with it somewhat, the sky is the limit....
I bet I could also alter the ambient signal to include some sort of reverb which would help perceive even more depth. But that's more trickery again. I'm not against it though, if it enhances the musical engagement. This is a (living/) listening room, I don't need to make mixing decisions and can go for the most pleasant rendition. I know that's a possible way some people actually use to create the (fake) impression of being in a big hall to listen to classical orchestra's. I like the ambient channels, but think adding the reverb would be pushing things to far for me. But it isn't that hard to try and I just might.
The reason for having the (fake) ambient signals is to fake listening in a larger space. Done! It works, walls do disappear. So where does it end?
When listening to music as background, the ambient amplifier is simply turned off. You never hear it is on at the sweet spot until someone turns it off while listening. Then you notice what's missing.
Let me give an example of a pop song that actually does very well in my room. I do not think highly of the musical content and it really isn't my normal listening genre but I do try a lot of different things and listening on headphones to that track I noticed something that made me curious.
On Facebook I came across: "Selena Gomez - Hands to Myself". Judging the clip, I do think it was posted on FB for entirely different reasons. Worked for me to get me to listen to it. But I just had to try it on the arrays. That pop song is actually very well produced. This has it all. The placement of effects is mind boggling.
It has depth, stuff happening way out in front of speakers and a very louring beat underneath it all. The room just starts dancing along. I almost forgot about the clip! The beat dances at ~40 to 50 Hz and hits hard. So be sure your system can do that. This might become a new reference for me in the fun pop genre. Replacing or beside the entertaining "Sugababes - Push the Button" track. I do have many more POP songs I play, but for completely other reasons. Anyone that's reading this and tries that track, let me know what you think. Not of the vocal qualities and certainly not the clip (lol) but of the production and placement of sounds throughout the song. It's not turning me around to become a fan of the genre just yet, but I can enjoy this...
Long post to say: I don't know, isn't it?
I have listened to a lot of music, some go wide and deep, some songs more in front of the speakers, other material is small and almost completely confined in the centre. Some pop songs have everything in a wide line from left to right. In short it's different on each track I play. To me that's a very good sign. It means I'm not introducing a one trick widening or depth enhancement that's the same on every track you play. Something like that could happen with early reflections for instance.
But I don't know if depth perception would/could be maximised by placing damping panels behind the speakers. I've heard the depth come and go in different settings/processing I have played with. Focussing strongly on having and keeping it of course.
I cannot bring myself to try damping panels behind the speakers as I know I won't be able to keep them if I like them.
In an ideal setup I'd want them there. But do I miss it? Not right now, no. The ambient channels help to strengthen the sense of 3D and depth (as well as give a sense of envelopment). So I do not feel I miss out.
At one point I noticed more depth on higher resolution recordings compared to CD quality. Checked time and again and this result was consistent. I could not figure out why. That is until I noticed part of my processing chain caused a delay on my main signal but not on my ambient channels. When I remedied that the imaging and depth was pretty much comparable again between both formats. What happened (at least I think so) is that high res material caused less delay (faster processing time) on the mains than CD quality, making the time gap between mains and the ambient channels longer than when playing a CD track. I have since introduced the same processing chain on the ambient channels to keep all channels delayed equally. Problem solved. But it shows a difference in timing from the ambient channels does have a pretty big effect on depth perception.
I haven't tried to optimise that perception yet. I can choose the balance and signal that is played trough those speakers and although I have played with it somewhat, the sky is the limit....
I bet I could also alter the ambient signal to include some sort of reverb which would help perceive even more depth. But that's more trickery again
. I'm not against it though, if it enhances the musical engagement. This is a (living/) listening room, I don't need to make mixing decisions and can go for the most pleasant rendition. I know that's a possible way some people actually use to create the (fake) impression of being in a big hall to listen to classical orchestra's. I like the ambient channels, but think adding the reverb would be pushing things to far for me. But it isn't that hard to try and I just might.The reason for having the (fake) ambient signals is to fake listening in a larger space. Done! It works, walls do disappear. So where does it end?
When listening to music as background, the ambient amplifier is simply turned off. You never hear it is on at the sweet spot until someone turns it off while listening. Then you notice what's missing.
Let me give an example of a pop song that actually does very well in my room. I do not think highly of the musical content and it really isn't my normal listening genre but I do try a lot of different things and listening on headphones to that track I noticed something that made me curious.
On Facebook I came across: "Selena Gomez - Hands to Myself". Judging the clip, I do think it was posted on FB for entirely different reasons
. Worked for me to get me to listen to it. But I just had to try it on the arrays. That pop song is actually very well produced. This has it all. The placement of effects is mind boggling.It has depth, stuff happening way out in front of speakers and a very louring beat underneath it all. The room just starts dancing along. I almost forgot about the clip!
The beat dances at ~40 to 50 Hz and hits hard. So be sure your system can do that. This might become a new reference for me in the fun pop genre. Replacing or beside the entertaining "Sugababes - Push the Button" track. I do have many more POP songs I play, but for completely other reasons. Anyone that's reading this and tries that track, let me know what you think. Not of the vocal qualities and certainly not the clip (lol) but of the production and placement of sounds throughout the song. It's not turning me around to become a fan of the genre just yet, but I can enjoy this...Long post to say: I don't know, isn't it?