Yes, room reflections and diffraction from a sharp edge box, so not smooth at all. This is the SoundEasy MLS that is in theory, gated. You might read their users guide to find out. I have a small cluttered room. No choice about that. I don't have a great big basement and piles of old cotton pillows nor is my system portable enough to go outside where the ambient noise is so high I can't measure anything.
Just for a sanity check, I measured my Fostex FE125's and they were just about ruler flat to 5K, where they go all to pieces.
For what it's worth I found ARTA so straight forwards that I didn't need to look at the manual except for the hardware configuration with LIMP. I am sure if you spend some time looking at all the menus and sub menus that you will see how things are supposed to go.
Gating isn't something that a program does automatically you have to set the start and end points yourself.
http://www.mediafire.com/?7dsikmz44lnf7br
Here is the link to a short guide I wrote a few months back in getting started with making measurements with ARTA.
Gating isn't something that a program does automatically you have to set the start and end points yourself.
http://www.mediafire.com/?7dsikmz44lnf7br
Here is the link to a short guide I wrote a few months back in getting started with making measurements with ARTA.
I don't think you're gating the measurements properly and this is letting in too much of the rooms influence, if you gate properly then you don't need to apply any smoothing at all. This is detailed in the guide I posted above. Don't worry if you're only able to get accurate data down to like 800Hz or so, this is normal in a typical listening environment/room, but it is enough to show if the integration between the tweeter/woofer is good.
The sharp edge box diffraction still made it look ugly, so the bit of smoothing just made the dip in response obvious, which was the question at hand, and the amount of droop I put in the treble to get it to balance clear. But yes, this was a quick first attempt to understand how to pick the window totally new tool for me, lots to learn. I do get similar results to my other two methods.
It is easier than SE. Your guide is most helpful as a quick start. I am sure those with LspCad will appreciate it too. Picking the window yourself jives you a sense of confidence in the reading.
So, build a better box. I am working up tape-on box edges to compare small & large radius with large facets to compare with felt blocks and a few other tricks I have used. Then I will bring the slump up and readjust the top end.
It is easier than SE. Your guide is most helpful as a quick start. I am sure those with LspCad will appreciate it too. Picking the window yourself jives you a sense of confidence in the reading.
So, build a better box. I am working up tape-on box edges to compare small & large radius with large facets to compare with felt blocks and a few other tricks I have used. Then I will bring the slump up and readjust the top end.
Well with 1/3 octave smoothing you can't really tell anything for sure.
When you don't gate measurements properly you get the room and any reflections contaminating the results. Now if the amount of reflective sound is low you can use a small amount of smoothing (1/12th octave) to get rid of it. However when things start getting worse, peaks and dips that aren't supposed to be there, start creeping in and no amount of smoothing with get rid of it, not to mention the smoothing gets rid of any dips or peaks that are supposed to be there.
To be sure of exactly what's going on you need to gate out all the serious reflections. Reducing the gate length reduces the low frequency accuracy, this makes getting measurements down low really hard, but in most cases it doesn't matter. Most crossovers occur around 2kHz and a short gate is more then enough to accurately capture what's going on.
Any MLS recording software will have options for setting the start and end points of the gate, ARTA's just happens to be very user friendly.
When you don't gate measurements properly you get the room and any reflections contaminating the results. Now if the amount of reflective sound is low you can use a small amount of smoothing (1/12th octave) to get rid of it. However when things start getting worse, peaks and dips that aren't supposed to be there, start creeping in and no amount of smoothing with get rid of it, not to mention the smoothing gets rid of any dips or peaks that are supposed to be there.
To be sure of exactly what's going on you need to gate out all the serious reflections. Reducing the gate length reduces the low frequency accuracy, this makes getting measurements down low really hard, but in most cases it doesn't matter. Most crossovers occur around 2kHz and a short gate is more then enough to accurately capture what's going on.
Any MLS recording software will have options for setting the start and end points of the gate, ARTA's just happens to be very user friendly.
Agree. Played with it some more. The result is the same basic profile, less low end as expected, and a lot of the expected ripple from diffraction. Shorter gating helps show the diffraction problems that one may otherwise blame on the room. Even with no smoothing, the same mid-range dip. By smoothing, you can get a better idea of where you need to put the new poles. I am doing a lot of comparisons between the three tools. That is how you learn which tells you what. You still have to link that to what you hear. When all is said and done, it is my wife's and my ears that I have to please. Really, I just need to decide which of the issues I will take on first. I have some ideas about the diffraction I have not tried that could help with setting the offset. I have some ideas on internal cabinet design to reduce reflections that need testing. Of course, lots of crossover work. The success so far is even in this state, it does not have the edginess with horns that bothers my wife. I can no longer attribute it to the tweeter breakup, but on basic driver distortion. It is quite listenable.
Considerably more progress. I was playing around with different edge radius profiles while comparing ARTA to TrueRTA quick sweep and SE MLS. ARTA does seem to be a lot better at showing issues, where the gating in Quick Sweep is too long averaging things out. In a small room, through the crossover region, I think I trust ARTA for flexibility. It looks like another tool in the kit. It is much friendlier than SoundEasy within the limits of what it does. If paying the license allows me to save plots and overlay them, it is worth the cost. Beats screen print and holding the pages to the light. I see issues in the same places with all three tools. Changes are reflected proportionally in all three tools. I doubt any of them are correct, but between them you can build an understanding. TrueRTA has an advantage that you can choose how much smoothing in a display and it holds 20 plots to compare. Sound easy has the advantage of plugging the acoustic and impedance/phase data into its design simulator.
Simple conclusions: Bigger the radius, the better. We knew that. Bevels not as good as radius. We knew that. The better managed the diffraction, the less over-hot the tweeter sounds. I did not know that. The last test was 2 inch thick foam with a cone cut out of it centered over the tweeter. On axis, very nice. Returned R5 to original and it balances better. Off axis it is a bit dull. I still think it has too much BSC, but I am not unwinding expensive coils until I build the next prototype box. Of course, changing the edge radius changes the baffle step too. Again, we knew that.
I am getting a better feel for how much high end roll off sounds about right compared to the 1M measurement. More than I thought. I still have some slight miss-givings about my mic calibration, as it seemed to be more severe than a lot of samples posted on the WEB. It does not really matter as long as I can correlate what I hear with what I see. The M-Audion Profire is far less of a pain than the e-mu 1616m I had, plus it does not crash Windows.
Simple conclusions: Bigger the radius, the better. We knew that. Bevels not as good as radius. We knew that. The better managed the diffraction, the less over-hot the tweeter sounds. I did not know that. The last test was 2 inch thick foam with a cone cut out of it centered over the tweeter. On axis, very nice. Returned R5 to original and it balances better. Off axis it is a bit dull. I still think it has too much BSC, but I am not unwinding expensive coils until I build the next prototype box. Of course, changing the edge radius changes the baffle step too. Again, we knew that.
I am getting a better feel for how much high end roll off sounds about right compared to the 1M measurement. More than I thought. I still have some slight miss-givings about my mic calibration, as it seemed to be more severe than a lot of samples posted on the WEB. It does not really matter as long as I can correlate what I hear with what I see. The M-Audion Profire is far less of a pain than the e-mu 1616m I had, plus it does not crash Windows.
ARTA allows you to use from zero to 1/3 octave smoothing depending on the analysis you're using. It also allows you to save overlays in the graphs too so you can keep several sweeps saved for comparison.
I will admit though that being able to save the impulse responses is might useful.
I will admit though that being able to save the impulse responses is might useful.
Just working my way through the book. Lots of options. As I had no power all weekend, no progress. ( Read the autobiography of Chris Kraft. Learned a lot about the astronauts and the early Mercury and Gemini programs. Confirmed what I thought I knew about Apollo. Good read.)
A little more time and I will probably pay for the licensed version. Saving is handy, and if it is part of my standard kit, it is only fair. I want to be able to compare measurements I make outside with those in my office. Different computers. All in all, a pretty friendly tool. To use it out of the box without reading the manual, may cause you to be misled a bit. 5th Element's quick start is a great place to start, but then RTFM (read the freaking manual). Theory matters.
A little more time and I will probably pay for the licensed version. Saving is handy, and if it is part of my standard kit, it is only fair. I want to be able to compare measurements I make outside with those in my office. Different computers. All in all, a pretty friendly tool. To use it out of the box without reading the manual, may cause you to be misled a bit. 5th Element's quick start is a great place to start, but then RTFM (read the freaking manual). Theory matters.
Completed my prototypes. With the 3/4 radius on the baffle, all dimensions as Zaph specified, foam lined and verified the tuning at 45 Hz. This is with the METAL dome tweeter, not the cloth as was specified. I used about twice as much stuffing as was supplied and it took care of several peaks in the 4K to 5K range. The great reduction in edge diffraction changes the character massively. I do mean massively. I went back to the stock crossover. One trick, this is with my subs. It takes the lowest octave to balance the tweeter. Unlike a lot of off the shelf speakers, it does not have a big mid bass peak to provide that false balance. Conclusion: The BOX is what makes this design work as intended. How about within 2 dB 500 to where the tweeter rolls off? That's not smoothed! Without the subs, it sounds bright. Simple lack of balance. If one wanted to run them without the subs, I would tune the box higher to get that big hump.
I have to thank 5thElement as switching to ARTA made it much easier to see the changes I heard.
The box is 3/4 MDF, oak bracing, dynamat with 1 inch foam over it, fully stuffed. Front baffle is reinforced with 1/4 hardboard. All glued. Every time I try removable panels, I don't like the results. I also put a couple layers of dynamat around the port tube as the thin plastic could get excited. I did not use the big plastic binding post cups either. I find they get excited. I use through bolt posts.
I am looking for ideas on how to do a standoff grill cloth that won't cause any diffraction. In the mean time, I need to do about 20 coats of paint.
Next box will be non parallel sides. This project had to stay regular for WAF.
I have to thank 5thElement as switching to ARTA made it much easier to see the changes I heard.
The box is 3/4 MDF, oak bracing, dynamat with 1 inch foam over it, fully stuffed. Front baffle is reinforced with 1/4 hardboard. All glued. Every time I try removable panels, I don't like the results. I also put a couple layers of dynamat around the port tube as the thin plastic could get excited. I did not use the big plastic binding post cups either. I find they get excited. I use through bolt posts.
I am looking for ideas on how to do a standoff grill cloth that won't cause any diffraction. In the mean time, I need to do about 20 coats of paint.
Next box will be non parallel sides. This project had to stay regular for WAF.
It's nice to see you've managed to resolve the original issues that you had and in a scientifically satisfactory way too. I know Zaph listens with subs, so perhaps he balances his 2 ways with them also. At a guess he'd probably listen with and without the sub though just to make sure.
I do know that lots of 6.5" drivers are far happier in cabinets larger then the usual compact two way would allow for. This would naturally allow for more of a bass hump then a larger cabinet would, so it could just be that small humps are more a product of the compact design.
I do know that lots of 6.5" drivers are far happier in cabinets larger then the usual compact two way would allow for. This would naturally allow for more of a bass hump then a larger cabinet would, so it could just be that small humps are more a product of the compact design.
Yea, ScanSpeaks want 30L or more. Commented over in the box character thread, after re-reading Olson decided to ease the rear edges. A bit better measurable through the mids. Three alternative stuffings on order and I am going to swap the smooth foam for dimples on the rear at least. I have a bit of a 500 Hz artifact from the box somewhere.
I figured as much. Automotive paint has totally changed since my last set. Now it is Dupont nason. Different primers, different thinners, sealers and procedures. You can't sand the sealer any more. Going to be expensive, again.
Nice work and a bold choice of color! 😉
Interesting that you choose the 27TBC/G over the 27TDFC tweeter. I personally think the 27TDFC is the stronger performer.
What’s the enclosure volume and port tuning you settled for?
Regards
/Göran
The tweeters were a surprise. I would have guessed the TDFC in a heartbeat. But by listening from both myself and my wife, we went with the metal. It has quite a bit more detail and I guess a bit lower distortion even though it has nasty breakup at 26K.
The box was also a surprise, being tuned just as John designed. I normally cross over at 60 Hz, so it is not terrible important to me.
Things that are different is how I do my bracing, it is fully glued and the front has an extra hardboard layer for rigidity. I don't use those stupid vibration prone terminal cups but through panel binding posts. I also wrap the port tube with dynamat. Foam lined and moderately stuffed. I'm going to let them work awhile before I try to make any in-room measurements. So far, it looks like I don't need the Rotels to mask the uglies, as I think they work fine on the Parasound.No complaints from the wife yet.
Again, I can't express how important the large edge radius is on all 12 edges. The felt ring less so, but it helps. It could be having a tweeter that actually puts out to the upper octave makes the diffraction that much important.
The box was also a surprise, being tuned just as John designed. I normally cross over at 60 Hz, so it is not terrible important to me.
Things that are different is how I do my bracing, it is fully glued and the front has an extra hardboard layer for rigidity. I don't use those stupid vibration prone terminal cups but through panel binding posts. I also wrap the port tube with dynamat. Foam lined and moderately stuffed. I'm going to let them work awhile before I try to make any in-room measurements. So far, it looks like I don't need the Rotels to mask the uglies, as I think they work fine on the Parasound.No complaints from the wife yet.
Again, I can't express how important the large edge radius is on all 12 edges. The felt ring less so, but it helps. It could be having a tweeter that actually puts out to the upper octave makes the diffraction that much important.
They PASS my wife's critical hearing test. Harry James horns. "The piano sounds like a piano". Only two speakers have ever met that approval. 2Ce and Sequels. I have just a bit of upper mid voicing to do to get Joni Mitchel right. Small, and her voice is really tough.
- Status
- Not open for further replies.