Hi! I posted earlier for help with an MLTL that I built for the Alpair 10.2. I'm very happy with the sound, but have been using DSP to tame a peak at 8khz. I've done lots of measurements and am confident that the peak is in the driver, not the cabinet. Does any one know of a physical way to reduce that peak? Thanks in advance!
The peak is audible, even surviving thru a full EnABL, but compared tot he original it was so much better that i didn’t really notice it until i compared to A10.3 which brought things much closer to the A7.3.
Off-axis and the acrylic gloss trick should bring it down a bit (so you don’t need so much DSP) https://www.diyaudio.com/community/...-tweak-and-other-metal-cone-markaudio.409632/
dave
Off-axis and the acrylic gloss trick should bring it down a bit (so you don’t need so much DSP) https://www.diyaudio.com/community/...-tweak-and-other-metal-cone-markaudio.409632/
dave
I got the Alpair 10.3 as a challenge after becoming proficient with notch filters. But it took me hours (without first consulting the FR chart) to settle on slightly bright/airy 0.4mH || 0.6uF (or 0.36mH || 0.68uF more common values) notch centered at 10khz (formula 5khz/sqrt(mH*uF)). Jantzen aircore and zcaps strongly recommended. The chart from taobao store showed a monster plateau 7-14khz (center 10khz), a bigger problem than the 7.8khz 12dB spike.
Alternatively, I had them "spirit of minimalism" in desktop bowls firing nearly straight up, each supported by 12"-over-a-can underdesk (pic in recent Full Range Speaker Photo Gallery #5249).
I do intend to coat/plast my 10.3 but a notch filter using good parts is more elegant (definition: simple and surprisingly effective).
Alternatively, I had them "spirit of minimalism" in desktop bowls firing nearly straight up, each supported by 12"-over-a-can underdesk (pic in recent Full Range Speaker Photo Gallery #5249).
I do intend to coat/plast my 10.3 but a notch filter using good parts is more elegant (definition: simple and surprisingly effective).
Last edited:
Hmm, ~34400/pi/8000 = ~1.369 cm diameter, so what's the dustcap diameter? Its measured response implies that this and all the BW above it are 'on/in' the dustcap, so can be damped numerous ways to tame all of its BW. FWIW I used star shaped stickers like Kindergarten teachers use to 'grade' student worksheets.
Felt, a couple inches thick maybe.
The main challenge with notch filter or other kinds of EQ is to not over-damp critical frequencies like 1-2khz intelligibility, 5khz dynamics, 9khz metal string etc. The Alpair 10.3 (a Mark Audio trait?) has no 5khz (dynamics) to spare at all. So any irreversible mod runs a significant risk of tired sleepy sound.
The main challenge with notch filter or other kinds of EQ is to not over-damp critical frequencies like 1-2khz intelligibility, 5khz dynamics, 9khz metal string etc. The Alpair 10.3 (a Mark Audio trait?) has no 5khz (dynamics) to spare at all. So any irreversible mod runs a significant risk of tired sleepy sound.
Last edited:
10.3 from taobao store for comparison (broad notch filter 0.4mH || 0.6uF centered at 10khz starts working on the HF plateau from 5khz and 16khz, leaving just a little brightness).
Attachments
Last edited:
wchang: "Alternatively, I had them "spirit of minimalism" in desktop bowls firing nearly straight up, each supported by 12"-over-a-can underdesk (pic in recent Full Range Speaker Photo Gallery #5249)."
It seems like a lot of fullrange drivers would have a pretty flat frequency response in a cabinet that had a baffle that angled the driver toward the ceiling.
It seems like a lot of fullrange drivers would have a pretty flat frequency response in a cabinet that had a baffle that angled the driver toward the ceiling.
Well the three charts (two from #6) are remarkably consistent modulo smoothing; all showed the plateau 7-14khz. Raw on-axis, I can confirm the 10.3 highest spike was near 8khz but hearing sensitivity falls off above 10khz.
I've praised Mark Audio taobao store before. Not every driver has a soundclip; those that do, the musical instrument tonality therein agrees with the accompanying FR chart.
I've praised Mark Audio taobao store before. Not every driver has a soundclip; those that do, the musical instrument tonality therein agrees with the accompanying FR chart.
Yeah, but... If seated far, the very high frequency "air and sheen" cannot be heard, and if 5khz is attenuated the music loses dynamics -- both macro (impact) and micro (articulation).
That said, the way an up-firing coherent driver(s) project a "bodily" image midair is fascinating. Ghostly suspended between driver and ceiling....
AFAIK this 10.3 chart is the only one with off-axis curves.
Do these links and included audioclips work?
https://item.taobao.com/item.htm?id=639405250410 (MAOP7)
https://item.taobao.com/item.htm?id=571656336184 (MAOP11)
https://item.taobao.com/item.htm?id=8294210293 (Alpair 10.3 disappeared from official store)
https://evanyu.taobao.com/search.htm?spm=a1z10.3-c-s.w4002-22432692471.1.3987199eFzfeYe&search=y (official store inventory)
The MAOP7 audioclip wooden recorder has a hint of steeliness to it. I just bought them, in offical store slot-loaded cabs, off of idlefish market for the price of a 11 driver.... Going to try a bigger Maevan.
https://www.diyaudio.com/community/...rkaudio-alpair-5-2-in-evansound-3-75l.393909/
wchang: "Yeah, but... If seated far, the very high frequency "air and sheen" cannot be heard, and if 5khz is attenuated the music loses dynamics -- both macro (impact) and micro (articulation).
That said, the way an up-firing coherent driver(s) project a "bodily" image midair is fascinating. Ghostly suspended between driver and ceiling...."
Right, but I mean angled up 15-20 degrees or so to get the off axis frequency response independent of toe in etc..
Sorry about these manual quotes, I'm having trouble with the quote function for some reason.
That said, the way an up-firing coherent driver(s) project a "bodily" image midair is fascinating. Ghostly suspended between driver and ceiling...."
Right, but I mean angled up 15-20 degrees or so to get the off axis frequency response independent of toe in etc..
Sorry about these manual quotes, I'm having trouble with the quote function for some reason.