Much of that is due to the excess flat baffle area in your test baffle. This really hurts diffraction performance almost as much as the sharp edge does. It is easy to pick unfortunate dimensions that compound the problem. The radius or bevel is good but it isn't the only difference between the two measurements, it is actually almost impossible to make a straight comparison between baffle edges without secondary effects being part of the difference seen.
Maybe not. But enclosure resonance problems are very difficult to deal with after the box is built, so I feel it is better to err on the side of robust.
An unbraced small box like this would likely have its first mode resonance above 300 Hz, which is within the passband of this enclosure, so bracing and/or damping is necessary. I am using both.
The total wall thickness for all 6 walls will be 1.5", a 3/4" inner layer of birch plywood, and a 3/4" layer of cherry hardwood.
I suspect that the flat panel has "unfortunate dimensions". I wanted to see how much diffraction hump I could generate with a flat baffle, and whether it fell where I expected it to. It was close, but not exactly what I expected. I also wanted to test the common notion that a waveguide eliminates (or greatly reduces) the effect of baffle size/shape. I was skeptical of this notion, and this little experiment confirms to me that baffle shape is important, just as it is with flat faced tweeters.
This confirms to me that there is no shortcut to estimating waveguide performance... either I build and test a prototype, or I climb the learning curve and get on board with ATH4. I already spend too many hours a day sitting in front of a screen, so I will try the test prototype approach for now... The other advantage of building a prototype is that it allows me (or forces me) to develop the tooling and jigs I will need for construction.
Here is the design of my next prototype. I have not decided if I will use XPS foam board or plywood.
j.
Attachments
This reminds some other thread: https://www.diyaudio.com/community/threads/did-siegfried-linkwitz-miss-a-trick.361790/post-6383874
It is about OB but still, shape is nearly same.
There is also some information how complicated toroidal shape can be built. https://www.diyaudio.com/community/threads/did-siegfried-linkwitz-miss-a-trick.361790/post-6407781
It is about OB but still, shape is nearly same.
There is also some information how complicated toroidal shape can be built. https://www.diyaudio.com/community/threads/did-siegfried-linkwitz-miss-a-trick.361790/post-6407781
Finding fortunate ones is much harder
When the waveguide narrows the directivity, less sound is directed to the edge and this reduces the effects of edge diffraction. This truth is often mistaken as meaning the size and shape doesn't matter which is not true.
The only shortcut is to eliminate all excess flat panel area and start the edge rounding as soon as possible. In a basic box you can't do better but the size and shape of the baffle still has an effect as does the depth and rear edge treatment.
Nothing wrong with prototyping in the real world instead of the computer.
From some of the recent threads where low diffraction shapes have been discussed I think there is misunderstanding that diffraction and directivity are independent variables. If you remove all diffraction as far as possible beyond the waveguide there will be less directivity because there is nothing to interfere causing cancellations. The two need to be balanced together to achieve the intended design outcome, when it comes to optimizing that balance the computer really starts to pull into the lead (but only if you know how to drive it
)
Yeah diffraction provides quite some narrowing at the main hump, around baffle size wavelength. Easiest way to have this help for a woofer response narrowing, without all the ripple that would come above, is to use minimal baffle without any roundovers. In other words its a box with maximal roundovers, they are just very small If its a two way speaker minimal baffle might not be an option as you need some volume inside the box for bass alignment. Bass and mids have contradicting requitements for a box: to get bass the inside volume needs to be big, but mids would benefit outside being small (diffraction). What you can do is find suitable compromise, increase size until the maximal roundovers get too hard to manufacture. Or have the box as small as possible and use another box for bass, separating interests is way to vanish problems like this.
One can always add more bass boxes to a system but mids need to be good so always optimize for mid performance. One can sacrifice bass of a woofer box to get more mid performance and then just add the bass back in by introducing another box. Its obvious problem remover, unless one has limited the project to use certain amount of ways for some other, more important, reason.
For tweeters, freestanding axisymmetric waveguide is rather easy and fast to tweak with ATH script and ABEC. One can roll out tens of variations in few evenings manually. With some automation script tens in hour I guess. Sims are very fast in ABEC, it takes more time to edit the script, click the project open and copypaste results than wait it to solve. Its worth it to learn how to use it if you are after tweaking last minute details, "top performance".
If its a two way speaker minimal baffle might not be an option as you need some volume inside the box for bass alignment. Bass and mids have contradicting requitements for a box: to get bass the inside volume needs to be big, but mids would benefit outside being small (diffraction). What you can do is find suitable compromise, increase size until the maximal roundovers get too hard to manufacture. Or have the box as small as possible and use another box for bass, separating interests is way to vanish problems like this.One can always add more bass boxes to a system but mids need to be good so always optimize for mid performance. One can sacrifice bass of a woofer box to get more mid performance and then just add the bass back in by introducing another box. Its obvious problem remover, unless one has limited the project to use certain amount of ways for some other, more important, reason.
For tweeters, freestanding axisymmetric waveguide is rather easy and fast to tweak with ATH script and ABEC. One can roll out tens of variations in few evenings manually. With some automation script tens in hour I guess. Sims are very fast in ABEC, it takes more time to edit the script, click the project open and copypaste results than wait it to solve. Its worth it to learn how to use it if you are after tweaking last minute details, "top performance".
Last edited:
Completely off topic, but I signed into OpenAI ChatGPT to see what it might do...
I asked ChatGPT to "Discuss how directivity, sound power response, and early reflections influence the sound of a loudspeaker."
Here was the response, which took about 3 seconds
Directivity, sound power response, and early reflections are all important factors that influence the sound of a loudspeaker.
It is very clever, and the response certainly sounds correct with quick casual reading... however, read closely and carefully and some of it starts to sound like BS. The detrimental effect of strong early reflections is not comb filtering. And I thought that sound power was measured in a full anechoic environment, not a hemi-anechoic one...
- - - -
Here is another one... I asked "What is the audible threshold of distortion in music reproduction? "
The response took about 3 seconds
The audible threshold of distortion in music reproduction can vary depending on the individual listener and the specific type of distortion. Generally, humans are capable of perceiving distortion at relatively low levels. For example, distortion caused by clipping, which occurs when the amplifier reaches its maximum output and cuts off the waveform, can be audible at levels as low as 0.1% distortion. However, distortion caused by harmonic distortion, which is the addition of unwanted harmonics to the original sound, can be perceived at even lower levels. Some studies have suggested that distortion levels as low as 0.01% can be audible in music reproduction.
It's important to note that the perception of distortion can also be influenced by factors such as the overall sound quality of the playback system, the type of music being played, and the listening environment. Additionally, different individuals may have different levels of sensitivity to distortion. Therefore, it can be difficult to define an exact audible threshold of distortion in music reproduction that applies to all listeners and situations. In general, however, it is recommended that distortion levels be kept as low as possible to minimize any negative impact on the sound quality. [end]
Again, it seems like a plausible response, but I would sure like more information on the studies that indicate 0.01% distortion might be audible in music reproduction.
I asked ChatGPT to "Discuss how directivity, sound power response, and early reflections influence the sound of a loudspeaker."
Here was the response, which took about 3 seconds
Directivity, sound power response, and early reflections are all important factors that influence the sound of a loudspeaker.
- Directivity: Directivity refers to the pattern of sound radiation from a loudspeaker, which is a measure of how the sound energy is distributed in different directions. Directivity can be described in terms of a polar response, which is a graph that shows how the sound intensity varies as the angle of measurement changes around the loudspeaker. Directivity affects the sound quality of a loudspeaker in several ways. For example, a loudspeaker with a narrow directivity pattern will have a more focused sound that is ideal for use in a home theater or music system. On the other hand, a loudspeaker with a wider directivity pattern will be better suited for use in a large auditorium or concert hall, where a more diffuse sound is required.
- Sound Power Response: Sound power response refers to the overall acoustic output of a loudspeaker, which is a measure of how efficiently the loudspeaker converts electrical energy into sound energy. The sound power response is typically measured in a hemi-anechoic chamber, which is an environment that is designed to minimize the effects of room reflections. The sound power response is important because it determines the overall loudness and tonal balance of the loudspeaker. A loudspeaker with a smooth and consistent sound power response will produce a more natural and pleasing sound, while a loudspeaker with an uneven response will have peaks and dips in the frequency response that can lead to a harsh or boomy sound.
- Early Reflections: Early reflections refer to the sound waves that bounce off the walls, floor, and ceiling of a room before they reach the listener. Early reflections are important because they can significantly affect the sound quality of a loudspeaker. For example, if the early reflections are too strong or arrive too soon after the direct sound, they can cause a comb-filtering effect that creates peaks and dips in the frequency response of the sound. This can result in a muddled or unnatural sound. To minimize the effects of early reflections, loudspeakers are often placed in specific locations in a room and are sometimes mounted in a baffle or enclosure that helps to control the direction of the sound waves.
It is very clever, and the response certainly sounds correct with quick casual reading... however, read closely and carefully and some of it starts to sound like BS. The detrimental effect of strong early reflections is not comb filtering. And I thought that sound power was measured in a full anechoic environment, not a hemi-anechoic one...
- - - -
Here is another one... I asked "What is the audible threshold of distortion in music reproduction? "
The response took about 3 seconds
The audible threshold of distortion in music reproduction can vary depending on the individual listener and the specific type of distortion. Generally, humans are capable of perceiving distortion at relatively low levels. For example, distortion caused by clipping, which occurs when the amplifier reaches its maximum output and cuts off the waveform, can be audible at levels as low as 0.1% distortion. However, distortion caused by harmonic distortion, which is the addition of unwanted harmonics to the original sound, can be perceived at even lower levels. Some studies have suggested that distortion levels as low as 0.01% can be audible in music reproduction.
It's important to note that the perception of distortion can also be influenced by factors such as the overall sound quality of the playback system, the type of music being played, and the listening environment. Additionally, different individuals may have different levels of sensitivity to distortion. Therefore, it can be difficult to define an exact audible threshold of distortion in music reproduction that applies to all listeners and situations. In general, however, it is recommended that distortion levels be kept as low as possible to minimize any negative impact on the sound quality. [end]
Again, it seems like a plausible response, but I would sure like more information on the studies that indicate 0.01% distortion might be audible in music reproduction.
ChatGPT, or DingbatChat as I name it on my Home Screen; needs a little guidance-
me-
Which studies have suggested that distortion levels as low as 0.01% can be audible in music reproduction?
Dingbat-
“ There have been several studies that suggest that distortion levels as low as 0.01% can be audible in music reproduction. Here are a few examples:
"Audibility of a Single Percentage of Harmonic Distortion in Audio Amplifiers" by Floyd Toole and Stanley Lipshitz (published in the Journal of the Audio Engineering Society in 1981). This study found that even small amounts of distortion can be audible, and that distortion levels below 0.1% can still have a noticeable effect on the sound quality.
"Perception of Distortion in Amplifiers" by Richard Cabot and Charles Waddell (published in the Journal of the Audio Engineering Society in 1991). This study found that distortion levels as low as 0.01% can be audible in certain circumstances, and that some listeners are more sensitive to distortion than others.
"Subjective Assessment of Low-Level Distortion in Audio Systems" by Joshua Reiss and David Borwick (published in the Journal of the Audio Engineering Society in 2004). This study found that distortion levels as low as 0.01% can be audible in certain types of music and with certain types of listeners, and that the audibility of distortion can be influenced by factors such as listening level and musical content.
Overall, these studies suggest that distortion levels as low as 0.01% can be audible in music reproduction, and that even small amounts of distortion can have an impact on the perceived sound quality”
me-
Which studies have suggested that distortion levels as low as 0.01% can be audible in music reproduction?
Dingbat-
“ There have been several studies that suggest that distortion levels as low as 0.01% can be audible in music reproduction. Here are a few examples:
"Audibility of a Single Percentage of Harmonic Distortion in Audio Amplifiers" by Floyd Toole and Stanley Lipshitz (published in the Journal of the Audio Engineering Society in 1981). This study found that even small amounts of distortion can be audible, and that distortion levels below 0.1% can still have a noticeable effect on the sound quality.
"Perception of Distortion in Amplifiers" by Richard Cabot and Charles Waddell (published in the Journal of the Audio Engineering Society in 1991). This study found that distortion levels as low as 0.01% can be audible in certain circumstances, and that some listeners are more sensitive to distortion than others.
"Subjective Assessment of Low-Level Distortion in Audio Systems" by Joshua Reiss and David Borwick (published in the Journal of the Audio Engineering Society in 2004). This study found that distortion levels as low as 0.01% can be audible in certain types of music and with certain types of listeners, and that the audibility of distortion can be influenced by factors such as listening level and musical content.
Overall, these studies suggest that distortion levels as low as 0.01% can be audible in music reproduction, and that even small amounts of distortion can have an impact on the perceived sound quality”
These tools are in the infancy, already changing how we search informtion. Try Perplexity.ai browser extension, can fetch info from a domain / certain page
Even if they aren't that good yet we could assume they will be in the near future. For example Google is in trouble now, biggest company in the world and their biggest source for revenue (search / ads) is now rivaled. They must act in order to survive, I don't know how but could suspect they need to offer similar tools, perhaps even better tools? Not sure, but its a war in a way, companies need to make better and better tools with this new technology. Not sure how ads will survive through this, perhaps they will somehow.
Even if they aren't that good yet we could assume they will be in the near future. For example Google is in trouble now, biggest company in the world and their biggest source for revenue (search / ads) is now rivaled. They must act in order to survive, I don't know how but could suspect they need to offer similar tools, perhaps even better tools? Not sure, but its a war in a way, companies need to make better and better tools with this new technology. Not sure how ads will survive through this, perhaps they will somehow.
Last edited:
This thread on ChatGPT shows that the issue of the truthfulness of its statements does not stem from incomplete information or would be solved by giving proper source. The neural network is currently, indeed, hallucinating and suffering from a sort of confabulation. This seems to be the side effect of its chat-ability.
Well, the model is just calculating what word / letters would most probably come next. It really doesn't do reasoning or fact checking or anything like that, just gives some text that has high probability, what ever that is. Try the perplexity.ai, or new Bing, who give references to sites where the info is from so you can fact check.
For now, its good tool for some use cases like as text processor, and perhaps on some cases good to do reflection, bounce ideas with as long as one knows the answer might be completely wrong. Writing and responses, be what ever, can boost your own thought processes and lead to tracks. Tasks in programming, text formatting, summarization and stuff like that, a godsend Try ask a slogan for your company, or turn your emails into more formal one, or how 15yo would write them and so on
edit.
See the text above from this very post prompted for sentiment:
There is million use cases that can boost productivity and give insight. Searches / facts stuff will improve upon time but its amazing just as text / data processor alone. I should start correcting my posts for grammar through it. Exciting times
edit.
Asked to refactor my post with better grammar, without changing the sentiment:
For now, its good tool for some use cases like as text processor, and perhaps on some cases good to do reflection, bounce ideas with as long as one knows the answer might be completely wrong. Writing and responses, be what ever, can boost your own thought processes and lead to tracks. Tasks in programming, text formatting, summarization and stuff like that, a godsend
Try ask a slogan for your company, or turn your emails into more formal one, or how 15yo would write them and so on edit.
See the text above from this very post prompted for sentiment:
There is million use cases that can boost productivity and give insight. Searches / facts stuff will improve upon time but its amazing just as text / data processor alone. I should start correcting my posts for grammar through it. Exciting times
edit.
Asked to refactor my post with better grammar, without changing the sentiment:
Last edited:
You are absolutely correct, and thats what the new Bing is doing, perplexity as well, perhaps others I don't know of. Bing uses ChatGPT in the search engine in this way so you can do the fact checking, there is links in the answer to the sources which the answer is based on. There could be an AI agent that does the fact-checking for you, but what if its biased for Microsofts interests, or advertizers interest, or your governments interests, or biased by your own prompts before?
Also people, literature, science, can provide "false" info so an AI is not alone, fact-checking and personal research is key what ever source you have. It would be best to have multiple sources. Do a google search and the first page is filled with results that they got money from, paid results. You still need to weed out the answer you are looking for. Usually this means checking out multiple "sources", links, and filter the content to find what you are looking for and this is Googles business model, billions of dollars per year for people trying to find stuff out between the ads.
Perhaps there are range of "fact-check-agents" in the future, one for each religion, belief and policy and you employ the whole army to report you what they think is right and choose to trust the "right one" Or use one agent that you feel is the correct one, that aligns with your belief system.
While its very interesting topic, and times, its wrong thread for this discussion Also easily gets into religion and politics which is forbidden, so I'll stop here as well.
Also people, literature, science, can provide "false" info so an AI is not alone, fact-checking and personal research is key what ever source you have. It would be best to have multiple sources. Do a google search and the first page is filled with results that they got money from, paid results. You still need to weed out the answer you are looking for. Usually this means checking out multiple "sources", links, and filter the content to find what you are looking for and this is Googles business model, billions of dollars per year for people trying to find stuff out between the ads.
Perhaps there are range of "fact-check-agents" in the future, one for each religion, belief and policy and you employ the whole army to report you what they think is right and choose to trust the "right one"
Or use one agent that you feel is the correct one, that aligns with your belief system.While its very interesting topic, and times, its wrong thread for this discussion
Also easily gets into religion and politics which is forbidden, so I'll stop here as well.
Last edited:
