Understanding Audio Frequency Response & Psychoacoustics (Video)

[Bruce T]

Sorry, I didn't realize you were talking about car audio instead of home audio. Yes, loudspeaker measurements are far more useful.

 

[Art of sound]

Sorry, I didn't realize you were talking about car audio instead of home audio. Yes, loudspeaker measurements are far more useful.

CEA 2034 apply for car audio as well. if ERDI is smooth then on axis and reflected sounds maintain the same tonality and tonality matters per harmon research upto the point to preferring loudspeakers or headphones that were tuned to the curve over the same ones that weren't. I have to say i always add a sub bass EQ to the EQ profiles from Oratory and Crinacle.
So we can't measure technicalities like resolution, soundstage, positional accuracy, timbre etc objectively and thats what bothers me as subjective stuff is not how we move forward as humans

 

[Bruce T]

Yes, meeting good standards is a great starting point for judging loudspeakers and headphones. I like products that perform well both objectively and subjectively. I'm not a fan of products with bad measurements, even if they sound good. Too many speaker designers fail to take the room interaction into account. Unlike a lot of people on this forum, I have a limited headphone background. I'm still using a pair of Sennheiser HD 600, probably 10-20 years old. No headphone amp or EQ.

 

[Art of sound]

Yes, meeting good standards is a great starting point for judging loudspeakers and headphones. I like products that perform well both objectively and subjectively. I'm not a fan of products with bad measurements, even if they sound good. Too many speaker designers fail to take the room interaction into account. Unlike a lot of people on this forum, I have a limited headphone background. I'm still using a pair of Sennheiser HD 600, probably 10-20 years old. No headphone amp or EQ.

 

[Art of sound]

1)Mean percentage of 2nd or 3rd harmonic distortion were not highly correlated with listeners’ distortion rating ( p = <0.05). Higher order harmonics is more likely correlated with listeners preferences so all measurements with 3rd harmonics are moot.
2) Midrange and bass are the most important sensitivities in olive's research and also align with what i observed in youtube and asr observing specific people's peferences in speakers, headphones etc, Most preferred speakers are the ones with less midrange colorations and a fuller bass.
3) spatial qualities pg 8 will likely take a larger preference factor if inputted into a study as loudness and distortion are getting more irrelevant with better chips, drivers will have to be removed. ( look at arendal 1961 $800/pair and they barely distort at 96db @1m )

Predicting Loudspeaker Sound Quality Erin's Audio Corner.pdf

Shared with Dropbox

www.dropbox.com

 

[Bruce T]

Wow, what an incredibly detailed and exhaustive study. Most people don't grasp the work, cost, and facilities involved. I laughed at the negative correlation between the speaker accuracy ratings by Consumer Reports and Harman. Similar to what Amir has said, people prefer speakers that are flat, neutral, and have great bass vs. uneven, colored, and thin. Yes, the midrange and bass are critical. I'm not surprised the low-order harmonics aren't that audible. They found direct sound and early reflections matter more than sound power. Spatial qualities are important, but to some degree that depends on individual preference. Personally, I'd rather have good soundstaging/imaging over dynamics.

 

[Audiofool1Q84]

I stumbled onto the thread after watching the video (which at first I thought was a real gem). 126 posts in I'm still confused from the perspective of a relative newbie who is trying to learn more about this field. There seems to a slightly heavier leaning towards the view (based on presented evidence) that lower frequencies are less discernible (counter to the original video). For someone trying to correct for room gain/modes does this mean:

1) Measure (using preferred method)
2) Apply heavier smoothing up to transition frequencies (similar to REW description)
3) Apply broader or narrower filters to deal with peaks?? (this part I'm truly confused about - if we are less sensitive to changes, i.e. broader ERB in this area, why bother with narrow filters?)
4) Don't mess too much with frequencies above 500-1000 Hz (unless looking to deliberately make tonal changes?)
5) Use less smoothing at higher frequencies (but then shouldn't we use sharper/narrower filters if assumption is greater sensitivity to changes here?)

Interestingly while the counter views have presented some evidence, Amir (whom I truly respect for his huge efforts in ASR) hasn't come right out to address this specific debate. If I'd go only by the video there certainly seems to be a fair margin for actually achieving the opposite with peq as intended.