What frequency response, in car, for subs, sounds the best?

[duanebro 10+ year member]

Ok, got about 30 min. I'll look all that over and see what I can do.

Take your time, I check back in the morning.

Thank you,

Duane

 

[keep_hope_alive Premium Member]

I used to use an ECM8000/SoundDevices USBpre/Laptop RTA setup. Now borrow the the Casella CEL-490 i use for work measurements ($7k meter). yea, it's flat.

the referenced equal-loudness contour is an important concept to grasp but it doesn't tell the whole story. yes, it explains why we prefer a loudness boost at lower frequencies and not at higher frequencies. it also explains why a flat RTA response (using Pink noise) results in a lack of bass.

the sensitivity of our hearing also plays a huge role in how we listen. i actually have 1/3 octave graphs of my own hearing (measured as SPL and FPL). I had this done at a hospital university's audiology lab (test was free since it's not a typical test). I also have custom molded ear plugs with flat 15dB filters (not free).

My right ear has a helluva dip - that I blame on the phone. i have nothing to substantiate that... call it a hunch.

personally, i like a gentle increase below 400Hz (with a notch at 250Hz for typical recording "mud"). I pull down 5k-6kHz a bit as well.

 

[keep_hope_alive Premium Member]

the right ear notch i have affects my tuning for a sound stage. it's right at the transition between PLD and IID. //content.invisioncic.com/y282845/emoticons/frown.gif.a3531fa0534503350665a1e957861287.gif

 

[Moble Enclosurs 10+ year member]

I agree with reducing road noise.

awesome.

Sorry I don't understand this. If the noise is louder than the radio, then you can't hear the radio.//content.invisioncic.com/y282845/emoticons/confused.gif.e820e0216602db4765798ac39d28caa9.gif

This is based on if you were to rta the response. The rta does not c,are whether it is noise or sound. So, if the bb noise is louder in some places or even all, since it essentially acts on an era like pink or brown noise, the rta will consider this a smoothed response to a point of the weighted function used. So, what I meant was, the bb noise will smooth the response in a measured aspect, not seerating sound from noise, like our ears do. //content.invisioncic.com/y282845/emoticons/biggrin.gif.d71a5d36fcbab170f2364c9f2e3946cb.gif hope that helps a bit

True, which is why many sq guys have very nice home stereos, It is way easier to have a low noise floor.

i could not agree more.

Can you find a reference to this? I found this: Equal-loudness contour - Wikipedia, the free encyclopedia If I am reading this correctly we are least sensitive at lower and higher frequency,and most sensitive around 2-5K.

Ironically, this is a great reference and something very similar to what I used in the military. All my references are from there so I cannot show any if I had it because my clearance was a tssci(highest possible) and when I got out, I had to shred everything I noted. But basically, if you look at this curve in the pic(Wikipedia is amazing by the way for having this), then you notice that a dip occurs between your referenced frequencies of 2-5kHz, and below 2khz...actually shown in that graph, closer to 1khz, the sensitivity increases(and I say sensitivity because though it is referenced in db, all other factors are constant in these measurements) nearly consistently down to the lowest frequencies. So, the higher the db(referenced), the higher our sensitivity to those frequencies are. If you were to mirror that response curve, you will get something that is realistic to an sq setup response for natural output. Now, keep in mind that yes, at lower frequencies we do tend to lose audible sensitivity, but this is counteracted by physical sensitivity such as vibrations...these will enhance the perspective of loudness even though they are inaudible. The same goes for higher frequencies when pain occurs from natural distortion of tweeters. Does that make sense? So, have you ever turned up a system, and the tweeters are painful? A great example would be most Mb quart drivers. This is what I am talking about for sensitivity...not measured sensitivity like it may be explaining such as simply listening to the source. I'm incorporating sensitivity with increased distortion or physical pressure changes. Because that makes a difference in an audible experience as well. Take bass shakers as a perfect example.

Um, you kind of lost me here. As far as Equal-loudness contours, Doesn't this not matter? (did I just hear someone hit the groan button ) The basic/simplified reason: If the mics recording the music have a flat FR and the speakers playing it back have a flat FR - Then sound comes out of the speakers the same as it went in!

See last

So if you know your car's transfer function, then you can design around that for the best SQ?

Yes and no. The transfer function changes in an environment for every single inch of listening positions available, and just by turning your head, you change the functions response. But an averaged function can be obtained. In a vehicle, there are actually symmetrically only 3 usable functions for every position in the vehicle. Once these are averaged independently with whatever resolution you choose...meaning number of listening positions, then they are averaged as a single response curve. So, accurately compensating for this perfectly is impossible for sq, but that also depends on the preferences sq response as well. And it is not just for sq, and actually is much easier to design around spl based setups due to the averages needed and changes that occur. But yes, a more accurate sq response can also be obtained than just putting speakers in and facing them to the listening point. A lot(and I mean a lot) can be involved than just the transfer function variables in setting the system up. But to answer the question more focused to the function, it is a yes and no answer and depends on what the goals are before they can be obtained or not. In some cases, it will not help, and will hurt the sq response, but that is why equalizers exist.

Please know that I am not trying to argue, I am trying to learn. I will try to rephrase my original question. What sound do most car audio guys, (Not the sq guys, as I think we are the minority here.) like to hear? What makes them or the kid on the street say "That is the best system I have ever heard! I want my car to sound like that!"

Oh, no problem at all. I hope this helped a little. As far as the general consensus of a full range sound, that is obviously relative, but majority...i have to agree that an anechoically flat response is generally not ideal. That more bass is considered to obtain a rich sound from a system. And if you look at the curve from the wiki source, you will see that the lower the frequency, the more output is needed to accommodate for lost (audible as they show it) sensitivity in that region.

 

[duanebro 10+ year member]

Moble Enclosurs - Thank you for your response. You, of course, design a lot of enclosures/boxes for people. It there a fr you shoot normally shoot for? I don't mean in half space, but what the user will hear in the car? I see many designs that have should have a bump around tuning, normally at the low 30's. Is this what most people like? I was trying to find out by using measurements, but I can't get a large enough sample. I have a feeling that many people don't really know what they like. Or what they are actually hearing.

I use a omnimic system to take measurements, it is not a $7K measurement suite, but is fine for most DIY testing. I am new to this in general. I am trying to learn how people perceive what they hear and use that info to help design my own audio system that sounds right to most. I know this is not flat, at least in the sub frequencies.

 

[keep_hope_alive Premium Member]

equipment used - this is a better mic, an Earthworks M50 with the SoundDevices USBpre

Two mics are shown in the same graphs below - one in passenger seat and one in drivers seat. Mics are at teh headrest.

I noticed some distortion and some harshness that I couldn't tune out. Then i realized that stupid MX (Media eXpander) was turned on. So I flattened the EQ and reset the turning. I noticed that I didn't need much EQ - it is mostly flat - maybe 6 bands are adjusted for the mids and tweeters.

All crossovers are set, I tuned without the sub. I do have some sub adjustments left for a better blend.

Soundstage height is great, nice and wide... and it does two seats quite well.

This was after I flattened the EQ - in 1/3rd octave without sub

This is same as above in 1/24th octave without sub

This is everything running - 1/3rd octave with sub

This is the end result after some tuning effort - without sub

This is the same as above - current tuning in 1/24th octave without sub.

This effort sounded really nice. I was very happy with the car at that point.

 

[duanebro 10+ year member]

Very nice, I should post what I have, I can't eq out much as my HU's peq is limited. I notice that you have the sub louder than the rest. So do I. //content.invisioncic.com/y282845/emoticons/smile.gif.1ebc41e1811405b213edfc4622c41e27.gif

 

[keep_hope_alive Premium Member]

yes, i have a gentle rise for midbass and subbass due to our reduced sensitivity to low frequencies. flat on an RTA and "flat" relative to our hearing are two different curves.

I had minimal EQ there. everything has changed since that measurement, and i'm changing my entire front stage (drivers and enclosures) so i'll have to make new measurements when that is done.

 

[GammaRadiation 10+ year member]

I usually like it with in a 3-4 dB range centered around 35-40Hz. I usually run my subs bandpass from 20-60Hz. Above that I usually go for a "jazz" like EQ setting. Increased response near lower end guitar notes and in the vocal range. My ears are very sensitive to 10kHz + so I like to go with a light roll off up there. This also makes me hate most metal dome tweeters.

 

[keep_hope_alive Premium Member]

you'd probably love a full range driver in lieu of a tweeter. very natural, smooth roll off above 10kHz

somthing like this:

 

[GammaRadiation 10+ year member]

Thanks for the suggestion. I've got some 1" silk domes that I like but I will definitely keep it in mind for the future. //content.invisioncic.com/y282845/emoticons/smile.gif.1ebc41e1811405b213edfc4622c41e27.gif

 

[Moble Enclosurs 10+ year member]

Moble Enclosurs - Thank you for your response. You, of course, design a lot of enclosures/boxes for people. It there a fr you shoot normally shoot for? I don't mean in half space, but what the user will hear in the car? I see many designs that have should have a bump around tuning, normally at the low 30's. Is this what most people like? I was trying to find out by using measurements, but I can't get a large enough sample. I have a feeling that many people don't really know what they like. Or what they are actually hearing.
I use a omnimic system to take measurements, it is not a $7K measurement suite, but is fine for most DIY testing. I am new to this in general. I am trying to learn how people perceive what they hear and use that info to help design my own audio system that sounds right to most. I know this is not flat, at least in the sub frequencies.

you're welcome. As far as a general response, I do not have one, unfortunately. I combine the listeners goals of a response range with driver response operation in a specific enclosure choice and balance that with room resonances and placement and positioning to obtain a response curve. And the driver, I usually let choose the response range and tuning factors over any other preferences, because if you go out of the range of its capabilities for efficient output or exceed any limitations by trying to make it do more than it's capable of, then you might find yourself not enjoying the outcome as much as you like.

 

[keep_hope_alive Premium Member]

this feels like a thread-jack. but you asked about RTA's.

//content.invisioncic.com/y282845/emoticons/smile.gif.1ebc41e1811405b213edfc4622c41e27.gif

it took these a few hours ago. first time i've RTA'd in a few months. i see a few issues - such as my midbass level (80-180Hz)

with Pink noise from 1995 IASCA disc - volume at a normal listening level

playing music - NIN (The Fragile - Left, Track 2) - this is the top end of my listening volume - i rarely go higher than this.

Taken with teh CEL-490 held 2 inches in front of my face - i am sitting in the driver's seat. I did not move it around to see if the dip at 500Hz resolved itself (indicitave of phase interference)

 

[dumple 5,000+ posts]

how much does phase have an influence on this? say speakers in a rear deck in or out of phase of the subwoofer

 

[keep_hope_alive Premium Member]

phase has everything to do with response in a car (or any other audio system)

in simple terms - we use phase (or phase interference) to describe the arrival of an audio signal to our ears. when you have one full range driver (no crossovers) and no reflecting surfaces you have no phase issues. introduce reflecting surfaces and you get reflections that arrive later - and you get phase interference. introduce multiple drivers and you get various arrival times (and the additional reflections) and you get more phase interference. simply put - when you have more than one arrival time you have phase issues.

rear deck speakers will be manhandled by the trunk subwoofer. regardless of what the speaker wants to do, the pressures generated by the sub will move that cone. this is why rear deck speakers with trunk subs are usually a bad combination (unless the rear deck speakers are fully enclosed).

but keep in mind that phase relates to Path Length Difference (PLD). so depending on the frequency and PLD - you may have constructive interference (they add) or destructive interference (they cancel). as frequency changes so does the amount and type of interference. of course, signal level plays a role in this as well.

when viewed in a narrow band (even 1/24th oct) RTA, you can see phase interference as nulls and spikes. when the drivers are properly aligned in time/distance/phase then the spikes go away and response is smoother. nulls don't have much influence in 1/3 oct or 1/1 oct levels, but spikes dominate. i have the ability to process 96kHz 24-bit .wav recordings with 1 Hz FFT resolution. but it takes longer and requires a custom program we've developed. it can be a useful tool when trying to resolve phase issues. using TrueRTA and 1/24th octave resolution allows for real-time viewing and adjustments - much more useful when tuning in a car, IMO.

i'll be looking into my 500Hz dip, i could tell i was missing some midrange. I have a random collection of speakers from the 90's in the car now. Waiting to build new baffles and enclosures for the new drivers.