Very interesting. I have never seen such a 2Pi to 4Pi conversion effect. It should be a step of 6dB with a slope of around 3dB/Oct. Your measurement doesn't show that. Possibly because the baffle extension is too small but why would it reduce the peak at around 2kHz? I don't have an answer to that. I wish I had.
There's a lot going on here--not just the baffle step. And on a very narrow baffle such as this, the baffle step starts quite late and gets confounded with other response irregularities. There are a lot of trees, but the forest is that the baffle is largely accountable for the response irregularities you're seeing with the stock crossover and cabinet. The cabinet resonance may also be contributing and might be discernible with the high-resolution Klippel machine, but just damping the cabinet won't get rid of most of the problem.
I have superimposed your modified baffle FR (red) with the original one. It is now easier to see how your baffles worked. I still cannot explain why there is reduction in the peak but the difference is to small to explain by a simple chart comparison.
That does make the baffle action easier to see. Thanks. I can't resist posting one more plot--the system response with my revised crossover without the front baffle wings. I took this at about 15 degrees off axis--close to a normal listening angle with no toe-in--to show how much the high frequency response improves when you get rid of the diffraction peaks and dips caused by the center mounting of the drivers. The speaker sounds great with the new crossover, but access to the inside of the cabinet is too difficult and treacherous to make it practical.
That is my guess too. Resistance in pipe is not significant compare to head loss, which means kinetic energy drop due to height. Do you have a measurement for with the straws and without the starws? That should be pretty clear what the difference is.
Isn't that a beauty?
As I always say, often the issues are not due to cost cutting, it is instead caused by cutting off a competent designer from the design/manufacturing process.
It looks like unmodified, the peaking from baffle step happens to coincide with the resonant peak. The latter is obviously still there even after modification, it's just no longer riding on top of a bigger wave so to speak. Ideally you'll want to address both.
Either the "it" means something else or you are failing to see the difference between the two curves.
I just edited my last post to make it clear that the plot was taken with my Xover mod but no front baffle wings. I've already posted the plot with both mods in place. If you squint, you can see the stock response without the super tweeter in the system plot I posted earlier. The super tweeter makes a very small contribution--it just adds a little output above 14 kHz. However, without any filter it has strong output down to 3 kHz. I tried to blend it in lower at around 8 kHz with a proper low pass filter on the main tweeter, but I always ended up with a peak. I'm sure I could have figured it out eventually, but it's academic since no one in their right mind would go through what I went through to get the vanity woofer ring off and dislodge the old crossover. Sony really didn't want people messing with their speaker.Isn't that a beauty?
Can you share us how the baffle extensions look?
I know I have been asking too many favours but can you try two more plots, please. One with the with just your crossover changes without your baffles in place. And, the same but this time the super-tweeter disconnected (or covered).
I value our correspondence very much. Thank you.
Thank you for the responses, Sir. Much obliged. You have proved once again that it is often not the cost cutting but the lack of design acumen that limits a product's quality.
