Thank you!
For myself, the B&K curve is way too dark as well. It's like all the shimmer of the music has been sucked out.
My preferred curve is relatively flat from low end to about 1kHz, then a gentle slope down by 3 or 4 dB at the end of 20kHz. Not that my ears can hear anything above 15kHz, but I guess you could say I like my music to cut through when it comes to cymbals and higher details.
Like wesayso mentioned, to each his/her own. Music is about preferences, in every aspect of it.
My preferred curve is relatively flat from low end to about 1kHz, then a gentle slope down by 3 or 4 dB at the end of 20kHz. Not that my ears can hear anything above 15kHz, but I guess you could say I like my music to cut through when it comes to cymbals and higher details.
Like wesayso mentioned, to each his/her own. Music is about preferences, in every aspect of it.
So what about a flat curve? That's probably a bit to bright on the top end?
I can only suggest doing as I did. Start with the curve closest to your liking and listen to it for a couple of days on end. After that period, take a day of tweaking to get it closer while listening to a variety of music.
If you find something you like, measure what you've got and promote that and use it as the new target.
Repeat the above until there's no need to tweak anymore.
Sounds simple but for me, it wasn't. It took me quite some time, during which I noticed I liked different curves for different parts in my stage. If I locked in the overall shape, sometimes the phantom part was lacking in definition. Other times the sides became to bright. That's when it got complicated... I introduced the mid/side EQ to overcome that hurdle. Around that time Pano posted the http://www.diyaudio.com/forums/multi-way/277519-fixing-stereo-phantom-center.html thread and I recognized my listening experience in that paper and Pano's description of it.
The JBL graph shows how much variation between people exists. I could act real proud that I like the "trained listeners curve" but what does that mean. Nothing, right? It's still a preference, though the extended listening sessions and learning to listen (I've done that for years on end) and comparing it to real world sound perception might have something to do with it.
Personally I think our previous gear might have an influence too. I like a warm lower midrange, not too much, but just a hint that feels like a warm blanket in songs makes me feel good. I blame my previous speakers for that. I've had them for 20+ years.
There's no right or wrong in preference. Here, the tiniest differences may make a huge perception difference. As long as these tiny differences span a large area in octaves. And in a room with low early reflections the inter aural comb pattern is going to play at least some role. How big that is can be noticed by playing something that was mixed for multichannel on your stereo set. It isn't huge in terms of level differences, but it is enough to make quite a difference in perception.
You're spot on with the PPT from Sean Olive. That's where the reference comes from. It is mentioned on a lot of HT and other Audio forums and which is how I found it long ago looking for room targets (somewhere between 2008 and 2011).
A lot of factors will weigh in when judging room curves. It's certainly not a set it and forget it kind of deal. Power response, room reflections, the room liveliness itself all play a big part. And when to measure it and where to set it.
I prefer to only adjust the first arriving wave front trough the use of the FDW.
I hate the amount of shimmer I get on a lot of speakers. It's just not what I hear when listening to people making music right in front or around me. I'm not talking about a Rock concert here.
I'm also quite sensitive for Esses and Tee's in reproduction. Turns out that was mostly a phantom problem (for me at least). No problems like that with my specific mid/side EQ. Measuring my ears officially left and right showed they both measure exactly the same and are 'above average' in sensitivity at about 6 KHz.
What I hate is speakers that have a spike in the shimmer range. Often speakers that people refer to as "hurting their ears" after a while. A nice gentle and clean slope going downwards is good for me.
When I say shimmer, it's meant as, for example, a regular cymbal sound (drummer here) and not completely killing the nice wash available in a dynamic recording, unfortunately not present in most over-compressed pop these days.
So, for me, a 10dB drop is too much on nice and balanced recordings, On the other hand, cheap mp3 recordings/mastered tracks would benefit from such a drop to make them listenable!
Just as you previously mentioned that there is a lot of information in the lower range of the FR that we often do not have, listening on systems that do not have much output under 50Hz, I feel there is a lot of information as well towards the higher frequencies as well, giving it that little extra 3D feel to the sound.
Wish you happy tinkering in the next few days! (or weeks?)
Thanks for the link. I burst out laughing at this:
"In Fig. 14 the author has modified the original data to
separately show the result of evaluations by trained and untrained
listeners. This is compared to the small room prediction
from Fig. 13(a). The “all listeners” average curve
is close to the predicted target, except at low frequencies
where it is apparent that the strongly expressed preferences
of inexperienced listeners significantly elevated the average
curve. In fact, the target variations at both ends of the
spectrum are substantial, with untrained listeners simply
choosing “more of everything.” An unanswered question
is whether this was related to overall loudness—more research
is needed. However, most of us have seen evidence
of such more-bass, more-treble listener preferences in the
“as found” tone control settings in numerous rental and
loaner cars."
So true, and yet it shows the kind of elitist mentality that audiophiles tend to develop and does them disservice in the presence of others.
Regarding Wesayo's post, couldn't agree more that it is hard to get the balance right and small changes can be highly audible.
Good inputs on preferences of house curves, often speculate why at all do we prefer as soon as speakers is fair smooth corrected these downward tilts as frq goes up, anybody has any clues ?
Is it because stop band for microphone under recording process adds speedier VHFs than rest of band and reproducing speaker double this effect being HF limited so we want it masked a bit, also there could be break up resonances up very high for microphones and reproducing speakers which would probably benefit same masking curve.
There seems so much small stuff that influence into these final tunings that wesayso had been very good teacher about and shared here, at least for me it had been very helpful to learn from and open up do own experiments and get real improvements and experience in long run.
Really happy all the more modern development into speaker sector that we now a days can enjoy, for example not long ago into DSP sat using Linkwitz transform real time for my systems high pass slope into cooperation with room, having a known roll off can be transferred to any wishes and sitting rolling it one Hz at a time and bang when it hit 63Hz hair raised on my arms and sensed rhytm for tracks increased to excellent. I'm really not a dancer but from that time on its impossible to not start a bit of dancing when standing up and music plays, so little was changed in correction but it had so big a influence on final sensed musicality.
Is it because stop band for microphone under recording process adds speedier VHFs than rest of band and reproducing speaker double this effect being HF limited so we want it masked a bit, also there could be break up resonances up very high for microphones and reproducing speakers which would probably benefit same masking curve.
There seems so much small stuff that influence into these final tunings that wesayso had been very good teacher about and shared here, at least for me it had been very helpful to learn from and open up do own experiments and get real improvements and experience in long run.
Really happy all the more modern development into speaker sector that we now a days can enjoy, for example not long ago into DSP sat using Linkwitz transform real time for my systems high pass slope into cooperation with room, having a known roll off can be transferred to any wishes and sitting rolling it one Hz at a time and bang when it hit 63Hz hair raised on my arms and sensed rhytm for tracks increased to excellent. I'm really not a dancer but from that time on its impossible to not start a bit of dancing when standing up and music plays, so little was changed in correction but it had so big a influence on final sensed musicality.
Just as you previously mentioned that there is a lot of information in the lower range of the FR that we often do not have, listening on systems that do not have much output under 50Hz, I feel there is a lot of information as well towards the higher frequencies as well, giving it that little extra 3D feel to the sound.
Agreed! It's almost impossible to find a balance that suits all purposes. It didn't stop me from trying. And it worked out way better than anticipated.
True high dynamic range recordings are more gentle on the ears.
I remember standing in a venue full of people, with a load of sounds blasting my ears at high SPL and enjoying myself fully! I could not help but think: Why do I not mind the overall balance here, the finer nuances etc. This was at a Rock concert
. I'm much more scrutinizing at home.
When it clicks, it can be magical. But for it all to click has been quite the journey!
It's easy to be impressed for a relative short time, its much harder to say impressed for weeks, or even months on end! Every thing I try has to pass a long term test for that reason.
Did you catch the Sean Olive PPT ra7 linked to? There's some explanation there, though I cannot verify nor vouch for it. It does make sense to me.
What I do know is: flat FR does not work for me in my room. Nor in my Car. Some of this 'might' even depend on the true SPL level people listen at.
That's one variable I try and take away by knowing the level as measured by my RadioShack meter. All plots are calibrated with it too and while it may not be the end all of SPL meters, it will all correspond pretty good to "a" reference level. It's so easy to go really loud with a clean sound. It isn't needed for a thrilling experience I.M.H.O.
With Jriver's Volume leveling it catches most surprises so I don't get caught off guard by a low dynamic range recording blasting my ears off for being much louder. Once you really know your system it's pretty amazing what our senses can pick up.
I'll continue to try to connect the dots I see to the dots I hear
There is a straightforward reason for preferring a dropping response. What we are talking about and correcting is the steady state response, i.e., it includes the direct sound and all the reflections. This is what is meant generally when you say "room curve." You are correcting the sound power, not the direct sound. Keeping that in mind, if you make sound power flat, it will result in a rising on-axis response (assuming traditional cone/dome drivers and increasing directivity with frequency). This sounds bright because the on-axis response is rising. Remember again that on-axis response (direct only, no reflections) is king. The ideal situation is when you get the on-axis response flat and the speaker directivity plus room absorption make the HF response droop downwards. For whatever reason, this sounds natural to us.
Maybe because DRC uses frequency dependent windowing, the traditional drooping room curve, such as that in the B&K curve, needs a second thought. Maybe the windows are so short at HF that it is not accounting for the power response and instead just correcting for the direct response (which may or may not be on-axis) and so a drooping curve doesn't sound right.
It is extremely important to distinguish targets for the room response and the on-axis response. In my experience, sacrificing the on-axis response to target a certain room curve never sounds right. The direct on-axis response must be flat in order for it to sound right and then the speaker topology (crossover points, cone diameters, etc) and room absorption must work together to get the right balance. That is why throwing EQ at the room response is rarely the perfect solution. It is a good starting point though.
Last edited:
By the way, did anyone catch the latest review done by Mitch? https://www.computeraudiophile.com/ca/reviews/devialet-phantom-gold-loudspeaker-review-r654/
Interesting speakers! I like Mitch's writing style too, always backed up with some intriguing references. Not to detract from the current subject, I just happened to stumble upon it and thought I'd share. Mitch and Bob Katz were of great help to me when I started out using FIR filters. They asked all the right questions on several forums that were helpful for me in my start up. Next to Greg (gmad), their journey gave me some good clues to get me started and almost running straight away. I love that sharing that just might help others. Which is why I post such long boring pieces of text on a regular basis (lol).
Interesting speakers! I like Mitch's writing style too, always backed up with some intriguing references. Not to detract from the current subject, I just happened to stumble upon it and thought I'd share. Mitch and Bob Katz were of great help to me when I started out using FIR filters. They asked all the right questions on several forums that were helpful for me in my start up. Next to Greg (gmad), their journey gave me some good clues to get me started and almost running straight away. I love that sharing that just might help others. Which is why I post such long boring pieces of text on a regular basis (lol).
This morning I had my first real listening session for this new re-done test setup. Much longer than previous sessions and listening to a (very) wide variety of songs.
I'm quite pleased! Smooth sounds and it worked on all material I threw at it.
It worked so well that I even put on some Led Zeppelin! (That's saying a lot for me)
It made me drift off into the music, forget all about peaks and dips, reflections and diffraction, woofers and tweeters... just music and nothing but the music.
Had a fun morning, wishing I had more time to spare. After all, it's all about the enjoyment of the music...
I still have some ideas to improve upon what I did. That will take even more time though... for now I've got to run! Maybe one more song before I go? Oh no, not that again....
(back in the danger zone, always running late because I want to sneak in another song or two)
I'm quite pleased! Smooth sounds and it worked on all material I threw at it.
It worked so well that I even put on some Led Zeppelin! (That's saying a lot for me)
It made me drift off into the music, forget all about peaks and dips, reflections and diffraction, woofers and tweeters... just music and nothing but the music.
Had a fun morning, wishing I had more time to spare. After all, it's all about the enjoyment of the music...
I still have some ideas to improve upon what I did. That will take even more time though... for now I've got to run! Maybe one more song before I go? Oh no, not that again....
(back in the danger zone, always running late because I want to sneak in another song or two)
Last edited:
I must say, I rather like the most recent remasters. They were the first rendition of me that brought back the memories I had of the vinyl versions.
I've spend quite a bit of time on a vinyl Led Zeppelin collection including some early pressings, bootlegs and foreign singles that included otherwise unreleased (at least for me) songs on the b-side. (now all available in abundance)
Somehow the CD versions never gave me the same satisfaction as those vinyl disks. Chances are I was over-romanticizing the vinyl.
The remasters just work for me. I've been playing the high res versions and didn't even try the regular CD's. Not that I'm a snob on that sort of thing, but if it works, it works.
I'd say to play it all well balance is key. Easier said than done though....
I've spend quite a bit of time on a vinyl Led Zeppelin collection including some early pressings, bootlegs and foreign singles that included otherwise unreleased (at least for me) songs on the b-side. (now all available in abundance)
Somehow the CD versions never gave me the same satisfaction as those vinyl disks. Chances are I was over-romanticizing the vinyl
.The remasters just work for me. I've been playing the high res versions and didn't even try the regular CD's. Not that I'm a snob on that sort of thing, but if it works, it works
.I'd say to play it all well balance is key. Easier said than done though....
Last edited:
I decided it was time to join in on the fun fluid was having with measurement averages. I've always wanted to try it, never got around to it though. This was my first measurement session in over a year.
REW has grown, that's a fact. I'm not even running the latest version yet, but I really should.
Here's my raw average result (with pré EQ included) of my right side, compared to the center point measurement:
This is shown with a 6 cycle FDW in REW. The measurements where averaged without using any window, unsmoothed.
It kind of tells the story pretty good though, why I was happy with correction based on a single point measurement, even when listening over a larger area. Absorbing the early reflections, combined with the averaging the line arrays do themselves makes it clear this isn't a normal type of speaker. It certainly has its advantages, with just a little help from straight forward damping techniques and the modern DSP we have at our fingertips today.
I had seen this before (when I made my decision to trust in the single point technique) but didn't save the results that time. They were more rough than this example, as I didn't have all panels up yet at that time. (at that point in time the Frequency Dependent Windows were not included either)
I will try to get the best correction I can based on these new averages. Sadly my time is more limited these days but I'll try and show the results (as I've always done).
I ran out of time, so I don't have my correction ready yet. Just thought it would be interesting enough to show. Plus serve as a reference point for fluid.
A note for fluid: old fashioned averaging of 8 measurements (a little closer than the last numbers we used for you, to cover my main seating area) and summed with (A+B) in pairs. The summed total was brought down exactly 18 dB to match the center measurement. Averaging was done from the outside pairs towards the center. 4 pairs, each summed in pairs again and averaging those last two. (+6 dB + 6dB +6 dB = 18 dB) The top end has a high shelf (of 3.5 dB starting at 3 KHz, Q= 0.707) in pré EQ as that gave me better correction results. No other corrections than that shelf on the high end above 3 KHz in pré EQ.
REW has grown, that's a fact
. I'm not even running the latest version yet, but I really should.Here's my raw average result (with pré EQ included) of my right side, compared to the center point measurement:
This is shown with a 6 cycle FDW in REW. The measurements where averaged without using any window, unsmoothed.
It kind of tells the story pretty good though, why I was happy with correction based on a single point measurement, even when listening over a larger area. Absorbing the early reflections, combined with the averaging the line arrays do themselves makes it clear this isn't a normal type of speaker. It certainly has its advantages, with just a little help from straight forward damping techniques and the modern DSP we have at our fingertips today.
I had seen this before (when I made my decision to trust in the single point technique) but didn't save the results that time. They were more rough than this example, as I didn't have all panels up yet at that time. (at that point in time the Frequency Dependent Windows were not included either)
I will try to get the best correction I can based on these new averages. Sadly my time is more limited these days but I'll try and show the results (as I've always done).
I ran out of time, so I don't have my correction ready yet. Just thought it would be interesting enough to show. Plus serve as a reference point for fluid.
A note for fluid: old fashioned averaging of 8 measurements (a little closer than the last numbers we used for you, to cover my main seating area) and summed with (A+B) in pairs. The summed total was brought down exactly 18 dB to match the center measurement. Averaging was done from the outside pairs towards the center. 4 pairs, each summed in pairs again and averaging those last two. (+6 dB + 6dB +6 dB = 18 dB) The top end has a high shelf (of 3.5 dB starting at 3 KHz, Q= 0.707) in pré EQ as that gave me better correction results. No other corrections than that shelf on the high end above 3 KHz in pré EQ.
Last edited: