That the response depends on amplitude is one factor which loudness compensation is supposed to try to address. But the point is that regardless of loudness you will have a significant drop above 5k. So if you want to get a visual idea of what you actually hear, as a rough approximation you can put a 10dB per octave slope from 5k, which would make the scope show very close to a perfect sine.
I built some more plates and is picking up subs this weekend so I can start doing some proper testing of the system!
But with my experience with previous versions of the system, since DMLs are phase incoherent and wide dispersion you don't have the same issues as with regular speakers where you have to make sure to get good coverage without overlap causing phase interference/comb filtering. That problem is specific to several point sources playing an identical signal.
As for testing that I have not done comparisons measuring one speaker alone and then a cluster of four, which is what I typically run per stack. But will try to do that when I get a chance to do some outdoor testing. Before I found that more plates seems to flatten the response, but now when they are more precisely made I suspect that the difference will be smaller since they will have more similar FR.
thank you Leob. I'm eager to hear your entirely personal, yet objective evaluation by ear. I don't need graphs. As a professional in the field for over 40 years, I've worked in the electroacoustic sector long enough to understand that what one seeks is not even remotely what one sees or would like to see on graphs and analyses. I believe the matter is far too complex to be examined in an engineering format, especially with this type of speaker, as there are completely lacking in-depth reference points. Thus, I believe it will lead us to reach a compromise, but with extreme limitations. Additionally, I think that while interesting in certain aspects, 600 pages on this topic are largely redundant and useless. However, DMLs have a different sound and in many ways are interesting. Returning to our topic, I trust your opinions, having carefully analyzed your words in previous posts.
Thanks @pixel1! Of course in the end it is the listening experience that counts, but I find measurements to be a nice point of reference that can help you understand what you are hearing....and with hearing being so subjective I think that can be really needed sometimes.
It becomes a feedback loop, so you learn to correlate what you hear with what the measurements show, which over time will help you to learn to identify the cause and affected frequency when you do hear an issue.
It becomes a feedback loop, so you learn to correlate what you hear with what the measurements show, which over time will help you to learn to identify the cause and affected frequency when you do hear an issue.
I believe from my experience that instrumental control is only useful for confirming (when it confirms) your empirical impressions, if they prove wrong, don't care and don't treat them. The only condition is that you have the right experience, even at an advanced age, to understand exactly what to feel and also what not to hear
@spedge Obviously the reason for listening with a filter is to actually learn about hearing and frequency response. It is not uncommon to hear people speak about the importance of frequencies they cannot even hear if you play them a test tone in that region. If instead of just sticking to ideas and biases do some testing, you would for example realise that distortion at 15k is not really something we can pick up on since we cannot hear overtones of a 15k signal.
No one is saying information above 5k is useless, I'm just saying to be realistic about the importance of that band in a typical music signal, and to understand the limitations both with our hearing and equipment in that region.
If you are hearing something going on with sibilance I'm sure you are on to something and I am curious what that could be, but I'm just not buying the explanation that it is distortion around 15k that is an issue.
No one is saying information above 5k is useless, I'm just saying to be realistic about the importance of that band in a typical music signal, and to understand the limitations both with our hearing and equipment in that region.
If you are hearing something going on with sibilance I'm sure you are on to something and I am curious what that could be, but I'm just not buying the explanation that it is distortion around 15k that is an issue.
I would say most speakers reproduce much less than that range, and so does most people's hearing and most music recordings as well. Perhaps for a speaker to be a good audiophile or PA speaker it should strive to reach 20k, but very few systems or recordings will have a decent response down to 20HZ.
That can be good for movies and such, but just like the top octave, the bottom (20-40Hz) is pretty pretty unimportant for most music. In my experience a HPF around 40Hz often sounds better because the information down there just contains crap that takes up headroom, and demands a well treated room to not be a total mess anyway. My subs are designed go down to 35Hz, which I think is perfect for most music.
My point is really that a lot of people tend to assign a lot of importance to those extreme ranges even though they are a very small part of the listening experience. You can make an really amazing sounding system that only covers 50-16k, and there are many bad systems that cover 20-20k.
If you go chasing rabbit holes related to things like "singing at 18kHz", or "distortion at 15k" I think you are more likely to end up with the latter.
You need to understand that in the octave between 500 and 1000Hz, a 5 dB dip, or 1% difference in THD, will certainly have more than 10 times as much impact on the listening experience than the same magnitude in the octave between 5-10kHz. And between 10 and 20k we are talking 100s of times less effect, if any at all for the top part unless you are a dog or teenager.
That can be good for movies and such, but just like the top octave, the bottom (20-40Hz) is pretty pretty unimportant for most music. In my experience a HPF around 40Hz often sounds better because the information down there just contains crap that takes up headroom, and demands a well treated room to not be a total mess anyway. My subs are designed go down to 35Hz, which I think is perfect for most music.
My point is really that a lot of people tend to assign a lot of importance to those extreme ranges even though they are a very small part of the listening experience. You can make an really amazing sounding system that only covers 50-16k, and there are many bad systems that cover 20-20k.
If you go chasing rabbit holes related to things like "singing at 18kHz", or "distortion at 15k" I think you are more likely to end up with the latter.
You need to understand that in the octave between 500 and 1000Hz, a 5 dB dip, or 1% difference in THD, will certainly have more than 10 times as much impact on the listening experience than the same magnitude in the octave between 5-10kHz. And between 10 and 20k we are talking 100s of times less effect, if any at all for the top part unless you are a dog or teenager.
Hello SapphireSloth,
Have you made (or could you make) test with small pieces of foam at 1/3 and 2/3 of each 4 sides? In a test panel I have, it was the best compromise between bandwidth, FR smoothness and energy storage.
Would you share the spectrogram of your measurements to see when the peak of energy appears? To help, below is the setting I use for the spectrogram in REW..
I remember a patent suggesting the suspension you tested... stored somewhere into my hard drive... I should find it.
@homeswinghome how small should the foam pieces be? I'm clamping them front-to-back in this frame and it is really hard to get gauge the pressure being applied. Previously I tested full Top Bottom Left Right surround at different compression levels, 4 corners, full TB, 1 small piece center TB. I didn't bother supporting the sides because any restriction on their movement significantly reduce LR and even brought HF down, while leaving mids loud.
I am going to bring the frame out soon and test some different panels in it and I could start with some more mount arrangements. I'm thinking the 1/3 2/3 on TB might work better than the full TB clamp, but I think the LR sides won't want any mounting on my panels. I'm testing ~10.75" x ~22" panels mounted in this frame with a 9"x19" opening. If the panel was twice as wide the LR mounts probably make a lot more sense.
After the testing I did before it feels like this frame I built is really only good for solid clamping, and the best solid clamping arrangement every time I tested was full top and bottom (short sides) with long LR sides completely free.
I feel like the best mounts may be a moderate full top and bottom clamp with a a couple very light mounts on LR long sides. With my current frame idk how I can do that, and I think it will need the light LR foam to be glued to both panel and frame. That makes testing harder as well. All the weatherstripping I bought is too firm and only has adhesive on one side. I do have some Super 77 I could see if it stays tacky after drying on the other side, but in order to get the right amount of pressure on the mounts the dimensions have to be just perfect and that is really hard to do with this frame.
Another important factor for the sides is that baffles bring a pretty big improvement and it's even more tricky to get delicate and precise LR mounts while keeping the baffle extremely close without the height of the mount causing edge interference. Generally I think the stiffer the panel the less the long LR mounts matter, as long as the panel has enough damping anyway.
I am going to bring the frame out soon and test some different panels in it and I could start with some more mount arrangements. I'm thinking the 1/3 2/3 on TB might work better than the full TB clamp, but I think the LR sides won't want any mounting on my panels. I'm testing ~10.75" x ~22" panels mounted in this frame with a 9"x19" opening. If the panel was twice as wide the LR mounts probably make a lot more sense.
After the testing I did before it feels like this frame I built is really only good for solid clamping, and the best solid clamping arrangement every time I tested was full top and bottom (short sides) with long LR sides completely free.
I feel like the best mounts may be a moderate full top and bottom clamp with a a couple very light mounts on LR long sides. With my current frame idk how I can do that, and I think it will need the light LR foam to be glued to both panel and frame. That makes testing harder as well. All the weatherstripping I bought is too firm and only has adhesive on one side. I do have some Super 77 I could see if it stays tacky after drying on the other side, but in order to get the right amount of pressure on the mounts the dimensions have to be just perfect and that is really hard to do with this frame.
Another important factor for the sides is that baffles bring a pretty big improvement and it's even more tricky to get delicate and precise LR mounts while keeping the baffle extremely close without the height of the mount causing edge interference. Generally I think the stiffer the panel the less the long LR mounts matter, as long as the panel has enough damping anyway.
Adding powered subwoofers is part of the speaker system with most subs reaching down to 20hz. Majority of all speakers reproduce high frequencies up to 20khz. since most tweeters are rated to play that high with some even higher then 20khz.
Its good for both movies and music. Most audiophiles feel 20-20khz is very important for music. those extreme ranges are important it what separates the average joe from audiophiles. Most wont hear a 20khz test tone but when applied to music that 20khz will add that air or spaciousness in the music. You can make an even better system that covers the full frequency spectrum which is 20-20khz.. The 20-20khz does not make a system sound bad. The standard goal is to have the least amount of distortion in all frequencies.
Hello,
You could try in the range of few centimeters (1, 2, 3cm...) depending on what is the minimum needed to keep the panel in position with no interference with the frame.
In the tests I recently did, my opinion was 1/3, 2/3 TB (Top Bottom) and LR (Left Right) was better than only TB but it is not the same panel...
From those tests (almost only measurements), I found difficult to say which arrangement is the best. They change a lot the bandwidth. Some peaks disappear, other appear, same for dips.
The reason I suggested to have a look to the spectrogram is I wonder if a minimum delayed energy is a better criteria than the FR.
Christian
I spent many hours testing some things today and I started with some mounting tests. Here's this test frame:
Next time I certainly will not use a router to make those insets because they are 1" wide and about 0.8" deep. Pic is of the back side, the front side has a very thin (~1/8" thick) face of that 1" wide inset remaining.
None of these mounts are independently adjustable, but there's two screws on each side which I thought could help with adjustments. They don't work very well, but I tried.
All the tests are done with this same panel and the exciter and frame were never moved, except one test where I moved the frame forward 1ft just to see how it changes things. When adjusting foam the panel had to be removed, foam moved, then put back in so there might be some small changes to the effective position of the exciter relative to the front frame opening. This was tested in my living room which is completely untreated, but it is medium size and has carpet and a couch at least. Frame is in center of room but ~3ft in front of my 85" tv with mic 6ft away. The exciter is a rattling 'thruster' mounted at approx 3/5 position.
I've attached the mdat. I found later after a lot of other tests when I remounted this panel and moved it back next to the tv there was a big dip in 150hz - 225hz and I discovered that shimming in a piece of foam in the center of LR between front of frame and panel made a significant improvement.
Next time I certainly will not use a router to make those insets because they are 1" wide and about 0.8" deep. Pic is of the back side, the front side has a very thin (~1/8" thick) face of that 1" wide inset remaining.
None of these mounts are independently adjustable, but there's two screws on each side which I thought could help with adjustments. They don't work very well, but I tried.
All the tests are done with this same panel and the exciter and frame were never moved, except one test where I moved the frame forward 1ft just to see how it changes things. When adjusting foam the panel had to be removed, foam moved, then put back in so there might be some small changes to the effective position of the exciter relative to the front frame opening. This was tested in my living room which is completely untreated, but it is medium size and has carpet and a couch at least. Frame is in center of room but ~3ft in front of my 85" tv with mic 6ft away. The exciter is a rattling 'thruster' mounted at approx 3/5 position.
I've attached the mdat. I found later after a lot of other tests when I remounted this panel and moved it back next to the tv there was a big dip in 150hz - 225hz and I discovered that shimming in a piece of foam in the center of LR between front of frame and panel made a significant improvement.
Attachments
Today was the first day that I've ever setup two panels for a stereo test. I've spent months testing just one panel at a time, completely overlooking the stereo experience. One thing I've always known about these DML panels is that 1 decent panel sounds much better than 1 tower speaker in a few ways, but what about 2v2?
Well, the center image is way more focused and sharp with my tower speakers. The DMLs are very 'open' which is pretty amazing for some things, but it kind of detracts from the experience for a lot of tracks. After lots of A/B'ing and some adjustments to improve the DML center focus I decided that the 'open' sound is definitely worth losing a little center focus for, but not too much.
For testing the center image focus I used CamillaDSP to convert the input to mono and send it to both L and R and then I sat in the middle and listened for how tight that center sounds. I also listened in stereo to various songs to confirm too. I played with the setup to get that center more focused for the DMLs, but the tower speakers were no effort at all - easily very focused and sharp. The center image is pretty important for the experience of many tracks so this is going to need more attention than I expected.
The first thing I tried for improving the image is how I thought I was going to build DMLs last year, which is also a suggestion I've seen here: put some absorptive material behind the panels to reduce rear reflections. I used folded microfiber blankets right behind (not touching) the panels for this but the panels really don't sound as good that way. They sound more "in a box", closed, muffled. It can measure better this way but it definitely does not sound as good. Maybe some large absorption panels several feet away would be better.
I moved the DMLs around the room to try to focus the center but it took me a while to find something that works. First I tried moving them closer together (blocking the view of tv) but that didn't really improve center focus, it only shrunk the entire soundstage a little bit. Then I tried moving them further out into the room, that didn't work. Toeing in and out didn't work. Toeing in and sitting close didn't work. What did work was moving them back towards the wall, about 2ft off the back wall. My tv is pretty big and this wall is the short one so I've only got 1ft or so from panel to side walls. This was the spot that my towers were always in but I had to move them out to be able to put panels there. The result was a big improvement in center focus. Still not as good as the towers, but good enough to trade that loss for the openess and midrange clarity of DMLs.
I know from previous tests that baffles make a big impact on the soundstage but I didn't experiment with any today. Trying things on 2 panels is twice as hard as 1 and I don't have enough scrap wood and clamps to mess with them both :| I am pretty confident that baffles will improve the center image along with the FR.
Well, the center image is way more focused and sharp with my tower speakers. The DMLs are very 'open' which is pretty amazing for some things, but it kind of detracts from the experience for a lot of tracks. After lots of A/B'ing and some adjustments to improve the DML center focus I decided that the 'open' sound is definitely worth losing a little center focus for, but not too much.
For testing the center image focus I used CamillaDSP to convert the input to mono and send it to both L and R and then I sat in the middle and listened for how tight that center sounds. I also listened in stereo to various songs to confirm too. I played with the setup to get that center more focused for the DMLs, but the tower speakers were no effort at all - easily very focused and sharp. The center image is pretty important for the experience of many tracks so this is going to need more attention than I expected.
The first thing I tried for improving the image is how I thought I was going to build DMLs last year, which is also a suggestion I've seen here: put some absorptive material behind the panels to reduce rear reflections. I used folded microfiber blankets right behind (not touching) the panels for this but the panels really don't sound as good that way. They sound more "in a box", closed, muffled. It can measure better this way but it definitely does not sound as good. Maybe some large absorption panels several feet away would be better.
I moved the DMLs around the room to try to focus the center but it took me a while to find something that works. First I tried moving them closer together (blocking the view of tv) but that didn't really improve center focus, it only shrunk the entire soundstage a little bit. Then I tried moving them further out into the room, that didn't work. Toeing in and out didn't work. Toeing in and sitting close didn't work. What did work was moving them back towards the wall, about 2ft off the back wall. My tv is pretty big and this wall is the short one so I've only got 1ft or so from panel to side walls. This was the spot that my towers were always in but I had to move them out to be able to put panels there. The result was a big improvement in center focus. Still not as good as the towers, but good enough to trade that loss for the openess and midrange clarity of DMLs.
I know from previous tests that baffles make a big impact on the soundstage but I didn't experiment with any today. Trying things on 2 panels is twice as hard as 1 and I don't have enough scrap wood and clamps to mess with them both :| I am pretty confident that baffles will improve the center image along with the FR.
Agreed.
A professional recording engineer will usually master a track with an HPF around 60Hz and an LPF around12kHz.
It's rather pointless to have a system that produces nothing more than rumble at sub frequencies, and cannot be heard above very high frequencies despite what the specs claim.
My partner, who builds magnificent open-baffle systems, is a 76year-old ex-drummer, (who used to play very very loudly indeed), and his hearing is more shredded than mine is not only because of his age, but also because of the common problem that musos often have from long-term high SPLs on stage. But yet both of us can both pick up deficiencies in high frequencies (that is, between 5khz and 10khz) when listening to high-quality recordings. We can both listen to a recording, and without taking response curves, identify peaks or dips, and their associated frequencies fairly accurately. Measurements simply confirm this.
There's very little important information above the 12kHz range (take a spectrogram of a good recording!) and it is pointless to chase after it. Again, as has been stated many times here, any male who is a few years older than his teens or twenties, simply cannot hear up to 20kHz. Yes, of course you might only-just be able to pick up a single tone at 15kHz. With headphones on. In a dead-quiet environment. But your ears are probably 20db down at that frequency anyway, and your actual range is nowhere close to that.
On the low end, maybe some movie theatre systems might have some vague benefit from a sub that extends to 20Hz. But it's pretty pointless if you're not driving at least 1kW RMS into it. You can't hear it anyway. Only feel it. And it will be rumble. Nothing more, I don't think anybody is able to identify what the notes are between 20Hz and 40Hz, so it will be for special effects, not for music.
Chest thump and gut-wrenching bass comes from the 50 - 100hz range, and 300W will do fine if that's what you're looking for.
I play electric bass. The lowest open-E is 41Hz. Even with the biggest PA systems, I don't think I've ever heard a decent, portable bass sub that goes much below 50Hz or so. It's just pointless rumble below that. Even a 5-string bass that goes down to 31Hz rarely plays the fundamental at 31Hz. A good example: look for Mad Men doing 'Ain't No Sunshine.' The bassist plays a beautiful open B in the intro. I have only ever heard one system reproducing that fundamental note, and it was a 4 x 15" speaker system. Most other systems will play the harmonics above that note, and it still sounds very OK to the ear. It's a psycho-acoustic thing.
I agree that many of the best systems rarely play 20hz - 20kHz, and that some of the worst systems actually do.
You might also want to try damping the panels directly. If the edges are not correctly terminated, then the panels will have uncontrolled resonance, and will smear the image.
Measure a single test panel and use REW's spectrogram or waterfall facilities. Coat the panel with layers of clear lacquer spray-paint. Start one layer, and increase until you get tight resonance. You might also want to put more layers in different positions to see what happens. Or use your fingers and gently press in various places around the panel to find where the smearing comes from.
LOL, yes the audiophiles who never spend on things that are completely meaningless based on misinformed biases, you mean those?
How can something you do not hear be important for your listening experience?
Do yourself a favour. Get the ABX testing plugin for Foobar2000, make versions of tracks with a HPF filter at 35Hz and an LPF at 16kHz and check if you can spot which is which in a blind test.
Strictly speaking, "most speakers" would include laptops and phones which you lucky if you get response below 200Hz and above 15k. But in a proper speaker covering to 20kHz is quite easy...that even really cheap systems can manage. I would sure expect a high grade system to do that with ease just based on using decent quality components.
Covering 20Hz is hard if you want any kind of amplitude. I been looking a lot at sub response graphs recently to find some suitable for my needs, and I have not seen any that covers 20Hz well. There are some subs that do go down to 20Hz, but they are typically labeled as "cinema" subs, and means major sacrifices in size and output level. For PA use there are some massive infra subs that are used in some of the biggest most expensive stadium systems, but most PA's will drop of quickly below 40Hz. And most PAs will be adjusted to drop off rapidly in the top octave even if they might be able to produce a flat response to 20k, to get a more pleasant sound at loud levels.
Yes. I've been building Hi Fi systems and professional PA systems and mixing desks, and have spent thousands of hours recording and engineering and mixing music both live and in the studio. And I'm afraid I'm not nearly an audiophile in the same league as those who listen to their music for an hour or two over weekends as a hobby, or those who can apparently tell the difference whether their speaker cables are directional and/or whether they're lifted off the floor or not.
Just like the engineer in the link (David Mellor, recording engineer and Audio Masterclass lecturer) I have always had ears "of the coarsest flannel cloth."
Brilliant follow-up