^ What is the goal, when the problems are solved?
I see from the augerpro tests that the best results have achieved ~15db lower relative level of box sound than the worst? How do we know what is the target, how much is needed for it not being audible?
Toole's paper (Loudspeakers and Rooms for Sound Reproduction—A Scientific Review, found by Google) shows some graphs (around page 460) of audibility of lateral echoes in rooms, decibels vs delay. Close to 0ms delay the graphs show audibility threshold of around -20db of relative level (testing single reflection, human speech). It is not the same thing here, but might provide some ballpark that -15db is pretty close to being inaudible? Certainly a lot better than the worst case already!
I see from the augerpro tests that the best results have achieved ~15db lower relative level of box sound than the worst? How do we know what is the target, how much is needed for it not being audible? Toole's paper (Loudspeakers and Rooms for Sound Reproduction—A Scientific Review, found by Google) shows some graphs (around page 460) of audibility of lateral echoes in rooms, decibels vs delay. Close to 0ms delay the graphs show audibility threshold of around -20db of relative level (testing single reflection, human speech). It is not the same thing here, but might provide some ballpark that -15db is pretty close to being inaudible?
Certainly a lot better than the worst case already!
Last edited:
I have just spent a couple of hours writing an overly long response talking around the B&W paper, experiences simulating speaker cabinets and a few other things. I hit the preview post to check the quotes and unquotes. Dismissed the preview thinking about a correction and somehow ended up with a reload page. Everything I had written was gone leaving just the original quote. What did I do?
I cannot face rewriting it and it was too long for forum post anyway. Does anyone ever look at the articles, blogs or wiki?
PS Browsing elsewhere on the web I have just seen the same reload box for about 2 seconds which then cleared by itself. It would seem the issue is likely with my browser or network connection leading to web pages not loading properly and retrying.
Last edited:
This is an interesting thought. Intuitively it seems right. Toole, Olive, and friends showed that the audible effect of cabinet resonance was due to the effect on frequency response. Whatever effects were measurable in the time domain (like energy decay) were not audible. In short, when we hear a cabinet resonance, we hear the frequency response peak. We do not hear the ringing.
They also showed that high Q cabinet resonances were less audible than low Q resonances, and theorized that this is because the high Q resonances are much harder to energize than the low Q ones... so viewed from a probabilistic standpoint, the high Q resonances occur much less frequently.
We also know that the energy available in music/video soundtrack varies dramatically with frequency. The peak is between 200 - 300 Hz, and it drops down slowly below 200 Hz, but drops down sharply above 300 Hz. So resonances between 200 and 300 Hz are probably worst case.
But do you know when you start typing if it is going to be a long post? In this case the paper sparked a few things as I was writing and it just grew. Lesson learnt hopefully.
I hope that you don't mind if I disagree with that, I think it depends on what you are listening to, percussion and in particular hand clapping sounds quite different when substantial steps are taken with box noise, but to make the issue more complicated I think if I was still listening to my Nad 3020 and Q dac ( instead of a class d and Sanskrit 10th dac ) the difference would be much more difficult to detect. Ringing blurs very short sounds, but this is an extreme example, how much it would effect our enjoyment of a whole piece of music is something different .
Probably true. Toole's listening tests were blind tests, and it is my opinion that some subtle aspects of musical enjoyment are sometimes not revealed by blind testing.
Although I have stated that "Cabinet structural resonances are only a problem if they are audible, and the evidence seems to indicate that they are not audible unless they are pretty severe"... When it came time for me to make my own cabinets, I made them very stiff and highly braced. Under the "knuckle test", they sound like solid concrete blocks. So I did not take the chance that structural resonances are probably not audible...
I agree that DAC quality and source material can reveal previously hid detail, and previously masked speaker problems. My current DAC uses the ESS Sabre ESS9038Q2M, and with 24 bit 96 kHz naturally recorded music, I hear things I never heard before, for better or worse (!)
Your report says the walls were measured with the mic at a half-inch and the speakers/drivers mic at an unspecified distance further away.
Can you please explain exactly what you did "correcting for the differing conditions." In a room, you can't apply some theoretical math "distance" correction that might apply out of doors or in an anechoic chamber.
I am puzzled by your reluctance to answer that question despite being asked a few times before.
B.
Ben> quite simply your ask is in bad faith. It is all about finding a chink in my method so you can blow it completely out of proportion and thereby attempt to convince us all that you're not really a jacka$$, but just a humble truth-seeker asking questions.
For all the honest readers out there, here is a simple explanation. I went from near field on the bare driver (1/2") and pulled the mic out to various distances, stopping around 30-something inches. At first I was looking for any significant change in the response character so I could decide on a distance to measure the cabinet for the actual testing. In doing that test I also adjusted the signal level and SPL adjustment within Soundeasy to keep the overall level of response about the same. I then used this correction to apply to the cabinet measurements going forward so the difference in levels you see between the bare driver and the cabinet are the actual differences if you were to measure everything from one point in space. Now that correction might be a bit off, I had to kind of eyeball it. Luckily the responses at the different distances were close enough that I could overlay the responses pretty well just tweaking the SPL adjustment in SE. I would be pretty confident that they are within 5dB of absolutely correct.
So for the accuracy of the absolute level of the cabinet testing responses vs the bare driver baseline response, I am reasonably confident. For the relative accuracy of the differences in level of the actual cabinet measurements to each other, I have perfect confidence since the same correction was applied to all. Hope that last part makes sense.
For all the honest readers out there, here is a simple explanation. I went from near field on the bare driver (1/2") and pulled the mic out to various distances, stopping around 30-something inches. At first I was looking for any significant change in the response character so I could decide on a distance to measure the cabinet for the actual testing. In doing that test I also adjusted the signal level and SPL adjustment within Soundeasy to keep the overall level of response about the same. I then used this correction to apply to the cabinet measurements going forward so the difference in levels you see between the bare driver and the cabinet are the actual differences if you were to measure everything from one point in space. Now that correction might be a bit off, I had to kind of eyeball it. Luckily the responses at the different distances were close enough that I could overlay the responses pretty well just tweaking the SPL adjustment in SE. I would be pretty confident that they are within 5dB of absolutely correct.
So for the accuracy of the absolute level of the cabinet testing responses vs the bare driver baseline response, I am reasonably confident. For the relative accuracy of the differences in level of the actual cabinet measurements to each other, I have perfect confidence since the same correction was applied to all. Hope that last part makes sense.
Last edited:
Took quite a while to get an answer from you: as far as I can follow your creative writing, your "adjustment" is sort of a personal estimate and your plots comparing cab and speaker are interesting but don't provide an idea of how loud is the cab sound really is compared to the speaker. Not to mention the unsolved issue of how much driver sound is leaking around to the cab mic.
It may be possible to reconstruct your report using a legitimate "correction" that eliminates the shortcomings now. Hope so.
BTW, remarks like "Ben> quite simply your ask is in bad faith" are plain hostile and against the rules of the forum. An apology is in order.
B.
It may be possible to reconstruct your report using a legitimate "correction" that eliminates the shortcomings now. Hope so.
BTW, remarks like "Ben> quite simply your ask is in bad faith" are plain hostile and against the rules of the forum. An apology is in order.
B.
Last edited:
Here is the post where I discuss what I did: https://www.diyaudio.com/forums/mul...on-methods-shootout-thread-5.html#post6299618
Is the correction perfect? No. Is there a better correction value? Maybe. I'm open to anyone looking at my measurements and giving a well-reasoned correction that is better.
Is the correction perfect? No. Is there a better correction value? Maybe. I'm open to anyone looking at my measurements and giving a well-reasoned correction that is better.
I thought I answered this before, but there must still be a misunderstanding with one of us. Do you think the source driver is radiating outside, like a normal speaker? Then I could understand your concern, but as I pointed out before this is the setup:
Now clearly there will be some sound from the little backside box. But I think we can agree it will be much less than a driver directly radiating to the surrounding? If so the signal (cabinet radiation) should be reasonably high compared to the noise (small backside cabinet of the source driver). What little amount of noise is contaminating the signal at this point is at least contaminating all other measurements equally, so there is still a lot of value in relative differences.
Look buddy, your total lack of self-awareness will never be cause for me to give you an apology.
Two points I agree with: 1) Ben is a naysayer who discounts anything that he does not want to believe, regardless of evidence. 2) your "comparative" accuracy is indeed going to be very good as you state.
But I am very sceptical of the absolute levels being accurate. A box radiates in a very complex pattern and its sound radiation is anything but simple. This means that any linear frequency scaling of a single mic position is not going to be very accurate since, in reality, the sound radiates in a very complex manor and that a single point is not going to be very representative of the real far field level..
Thanks Earl. Can you follow my link in post 133 and tell me what you think of the plot where I have responses at 1/2", 6", and 30" and tell me what you think? For the reasons you bring up I was expecting much larger differences than I actually ended up seeing. If, in looking at those plots you could suggest a better correction to the 1/2" vs 30" responses, I'm all ears.
I still think that it is more complicated than that. What if you moved the mic angularly around the source. I doubt that they would track that well in that case.
I wish that there were an easy way to do this, but I can't think of one. (Use a reverb chamber and compare the power response with and without the driver radiation would be the best that I can think of, but reverberation chambers are very rare.)
My point would simply be that your comparative results are good and indicate how one should build an enclosure. But finding out how audible this all is would be a mammoth undertaking. My advice would be to just be happy with the comparative results, build your cabinets that way and move on to more important problems.
I wish that there were an easy way to do this, but I can't think of one. (Use a reverb chamber and compare the power response with and without the driver radiation would be the best that I can think of, but reverberation chambers are very rare.)
My point would simply be that your comparative results are good and indicate how one should build an enclosure. But finding out how audible this all is would be a mammoth undertaking. My advice would be to just be happy with the comparative results, build your cabinets that way and move on to more important problems.
Glad you praise the comparative value of the tests. But, it was nice to learn what you mean by "comparative".
Earl: "My point would simply be that your comparative results are good and indicate how one should build an enclosure. But finding out how audible this all is would be a mammoth undertaking. "
It looks like you mean comparing different kinds of panel materials, not comparison of the driver and the cab wall... which was the core purpose being defended by augerpro. Yes, that's useful. But we still are no closer to figuring out how audible is cab wall vibration.
B.
Earl: "My point would simply be that your comparative results are good and indicate how one should build an enclosure. But finding out how audible this all is would be a mammoth undertaking. "
It looks like you mean comparing different kinds of panel materials, not comparison of the driver and the cab wall... which was the core purpose being defended by augerpro. Yes, that's useful. But we still are no closer to figuring out how audible is cab wall vibration.
B.
Last edited:
I had the same concerns and had outlined a post describing them regarding converting an accelerometer reading into absolute radiation.
Then I realised you were using a mic for this part, a helpful step maybe. Now, for comparison I'm reasonably comfortable especially as Earl has also seen value in them... for absolutes I feel that finding power would give a result that I'd be more inclined to sink my teeth into. I have doubts about the response measurement for that.
So on a rare occasion I select the text, right click and choose copy.. just in case. I sometimes also construct difficult posts in a text editor on my second screen.
Then I realised you were using a mic for this part, a helpful step maybe. Now, for comparison I'm reasonably comfortable especially as Earl has also seen value in them... for absolutes I feel that finding power would give a result that I'd be more inclined to sink my teeth into. I have doubts about the response measurement for that.
Sorry to hear that. I recall at one time in the past I would see timeouts, and yes, accidentally hitting the back button.
So on a rare occasion I select the text, right click and choose copy.. just in case. I sometimes also construct difficult posts in a text editor on my second screen.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.