Post 1996 Mariah started to sing with what's been known as a 'trade mark' whispery style of singing. Most people do not sing like this and most people do not really enjoy listening to it either.
The artificial 'airiness' of the Alicia Keys track was actually quite good for upper end balance imo. Studios tend to add an upper end tilt to most vocals to give them a more 'polished' and 'impressive/attractive' type of sound. It's sort of become the norm.
The artificial 'airiness' of the Alicia Keys track was actually quite good for upper end balance imo. Studios tend to add an upper end tilt to most vocals to give them a more 'polished' and 'impressive/attractive' type of sound. It's sort of become the norm.
I actually agree with both these statements. I think it all comes down to listener fatigue. As with any work we can get tired and fatigued. Producers, mix engineers, mastering engineers all know this. We must put in long hours with the music listening critically for small variations and make adjustments, and trying to capture the right "vibe". But at some point we get tired and we can no longer work properly and then we need a break. And before releasing a track we always let it sit for awhile and come back to it with fresh ears to get a holistic impression before starting work again or realizing that it is "done" (final decision). Sometimes we have a "wow that is better than I remember" or "damn what was I thinking there".
I personally would not trust myself on first listen to make an informed final judgement, even if I could easily identify problems or excellent qualities. But that first impression is very important to remember and come back to.
Here was my process for this test (also a good critical listening exercise for me)
- first impression notes on all the drivers for all the tracks
- eliminate as many problem drivers as possible
- several hours critical listening (over several days and sessions) taking notes, allowing some opinions to slowly form. Allow the preference to shift around
- some blind random shuffle tests to see if I can pick the drivers accurately based on the qualities I identified
- rest
- one final session, once through with fresh ears on my final top picks and make my personal preference choice.
It is also worth saying again that this test is more like wine tasting preference and not really about an "objective best". There are too many variables here to pass an objective judgement on these drivers : driver interaction with the foam-core nautiloss enclosure, FAST XO effects, room interactions affecting the recording, mic placement, frequency response of the recording microphones, mp3 compression, playback system effects. We can just get an impression of these drivers from these listening tests and the measurements, and this can help us decide if a particular driver is worth buying and trying in our own builds.
Last edited:
Leona Lewis would also have been a good choice for a vocal that would exercise a speaker's capabilities. She has great clarity and incredible range.
I really don't like her voice. I can't stand her really slow and laboured vibrato when singing softly.
Rachelle Ferrel's has an excellent voice, both in terms of technique and the range of things she does with her voice. Her Individuality album is also very well recorded and sounds sublime.
Fair point that the bass is mostly handled by the RS225 but it is unchanged for every mid under test so if there is a difference in the bass it can only be due to the mid. The harmonics of the bass notes (in the recording) stretch into the midrange bandwidth, especially due to the 2nd order xover slope, so you get harmonics of harmonics.
What I was hearing as 'HD' may have also been partly a Qts bump on the midrange driver or IMD.
This is quite true. The mid/tweeter under test has a -12dB/octave Bessel high pass filter set at 350Hz (electronic xo frequency, but acoustically it is closer to 500Hz due to where the woofer slope crosses it.) So at 250Hz, the mid/tweeter is only down -12dB, and thus one would hear some contribution. The woofer on the other hand, has a -24dB/oct Butterworth low pass filter and this quickly stomps out almost any leakage of the woofer into the mid range to reduce coloration of the mids by the woofer's cone breakup or resonance modes. This is particularly more important when using an aluminum cone driver like the RS225-8. It's a great driver with almost unique specs at this price point, however, benefits from a steep crossover slope to remove mid range muddiness that can result from shallow filter slopes.
In a way, the Harsch XO with an asymmetric slope arrangement gives the best benefits for us to experience a truer representation of the intrinsic "character" driver under test, yet still have a consistent source of bass provided by the FAST woofer to enable an apples-to-apples comparison.
True. I have played the game as if I were in the market for such a driver, and had to decide which one to buy.
Hmmmm.....
Some time with clip 6 gave a certain quick judgment which followed with further listening with clip 2.
Reviewing reference for both clips they are overwhelmingly better; solid bass and good stereo mix. Reducing this to mono of right channel for clip 2 reference makes it quite dry, but bass is still overwhelmingly too solid. So I high passed with 1024 point FIR clip 2 tracks and clip 2 reference. Dropped into multi-track, tweaked timing for better solo to solo switching. With frequent solos of ref v other tracks I sense better differentiation of X's room sound and direct sound of speaker setups. Via this a completely different set for top three emerges for me.
Also did some normalization tweaking; ref clip 2 is hot compared to test tracks. I'll repeat process with a few additional checks and see what I get.
Some time with clip 6 gave a certain quick judgment which followed with further listening with clip 2.
Reviewing reference for both clips they are overwhelmingly better; solid bass and good stereo mix. Reducing this to mono of right channel for clip 2 reference makes it quite dry, but bass is still overwhelmingly too solid. So I high passed with 1024 point FIR clip 2 tracks and clip 2 reference. Dropped into multi-track, tweaked timing for better solo to solo switching. With frequent solos of ref v other tracks I sense better differentiation of X's room sound and direct sound of speaker setups. Via this a completely different set for top three emerges for me.
Also did some normalization tweaking; ref clip 2 is hot compared to test tracks. I'll repeat process with a few additional checks and see what I get.
The next round will be a finals round with top 2 from each round plus one of my choosing. Maybe in the future I can have a whole new set of tests and can include Fountek if someone provides it. Which model by the way? Fountek aluminum cones tend to ring quite a bit.
Are you saying the recorded sound clips don't have the bass of the reference? An 8in woofer with Linkwitz can only do so much. I know that when I play this with my open baffle setup with dual 12 in woofer's it sounds so much better because the weight of the bass is much better. However, this test is really about the mid and highs and some leakage of the mid bass from the driver under test down into the 150 to 200Hz range as result of -12dB/octave high pass.
That is neat that your FIR can remove or reduce the effects of the lively room I have. However, I would caution that such advanced post processing of the recorded sound that I provided may not be as true to the source as just listening. The recording mic is 0.5m away and the first reflections come after the first direct undisturbed wave arrival at the mic. I am not sure an anechoic room recording would sound very good - dry as you say.
Thanks for your comments on how you are coming at a conclusion but odd that with the convolution the answer is flipped around.
You might be surprised what you can hear with earbuds. Listen to clip 2 with your buds and see if you can hear a difference.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.