Hyundai Veloster N Car Stereo Misadventures

Discussion in 'Cars, Motorcycles, Boats, Airplanes Talk' started by purr1n, Oct 5, 2022.

  1. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,771
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    @ChaChaRealSmooth asked me if I would lend him my Veloster (and presumably let him drive it hard) if he ever down here in Texas, and I was like sure. I'm not possessive of my stuff, and besides, I can't take it with me when I die. Stuff comes and goes. People come and go. Impermanence is a fact of life. Anyway, this led the conversation about the car stereo. I mentioned that I should take a frequency response measurement of the Infinity 8 speaker sound system (Harmon, JBL Samsung, etc.) and he said that he would be interesting in seeing it. I do have plans to upgrade the speakers. I'd love the put in one of those DSP/amps, but this could be difficult given that the headunit sends proprietary CAN-bus (with SPDIF) signals to the factory DSP/amp unit. Nothing would be stopping me from inserting a DSP/amp after the factory amp though, to de-fuckify the sounds as the factory unit actually does some EQ, lifting up the highs.

    2021-hyundai-veloster-n-dct-455-1608564205.jpg
    Photo from Car and Driver: https://www.caranddriver.com/photos/g35031104/2021-hyundai-veloster-n-gallery/?slide=32

    Generally, the "technicalities" of the sound system isn't bad. It's better than most premium sound system I've heard from most cars. However, the tonal balance is screwy and the built-in EQ is too limited to be able to correct it. There a lot of bass, but the the emphasis is well below the mid-bass. That is, it's not excessive boomy. In fact it's a nice touch because it allows us to hear bass over road noise (those 19" wheel low profile 235/35 tires do us now favors) and the exhaust, which in N mode retards the timing when the pedal is lifted. This means the exhaust valve is partially open before full combustion so some of the gas ignites after the exhaust valves causing loud pops. (Except it's not as lame as the BMW guys who do this intentionally via aftermarket ECU tuning). There's actually a reason for this because it keeps the turbo spinning reducing lag when we get back on the gas. The main problem is small midrange suckout and very bright highs. Maybe a small peak around 8kHz that causes some sizzle.

    So here is the frequency response that I took with the microphone pointed straight forward where my head would be in the driver's seat.

    Veloster N
    Frequency Response
    Microphone straight forward where my head would be in the driver's seat
    upload_2022-10-5_20-52-47.png

    So wait, that looks pretty flat right? Nope. Remember, flat is fricking bright sounding for a listening position measurement. Flat is good for an on-axis speaker measurement three meters away, but oftentimes when we put speakers in a real room with furniture, carpet, and listen a few feet away, the frequency response is different at the listening position. B&K did some research back in 1970 and wrote an AES article on this. Their research was better than anything Sean Olive did in his entire life regarding frequency response targets. To summarize, the B&K curve stipulated a 1db drop every octave from 200Hz up. This by the way is fairly consistent with real-room listening position measurements with many speakers that measure flat on-axis three meters (or closer) away.

    B&K target (listening position)
    B_and_K.PNG

    So imagine the above frequency response near-field inside a car, with a glass windshield that is reflecting the highs off the dash tweeters. It's bright! Even then, if we observe the frequency response, we will note a rising response from the 500Hz and up.

    Anyway, did Sean Olive's Harmon research have something to do with this?
     
    • Like Like x 9
    • Epic Epic x 5
    • Agreed, ditto, +1 Agreed, ditto, +1 x 1
    • List
  2. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,771
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    So anyway, there's the B&K curve. Another widely accepted target at sound studios is similar. It's simply a 1db per octave downward slope. The main difference is that this curve is bassier than the B&K since the B&K is flat from 20Hz to 200Hz. I'd take either one depending upon my mood. I just don't like bright.

    So let's take that frequency response measurement and overlay the targets.

    Veloster N
    Frequency Response
    Microphone straight forward where my head would be in the driver's seat
    No EQ adjustment
    upload_2022-10-5_21-11-33.png
     
  3. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,771
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    The first thing I tried was bumping up the mids on the EQ +4. (There are only bass, mid, and high EQ settings). Here is the result. The mid EQ setting bumps up the region from 200Hz to 2kHz. This was too wide of a range, especially toward the lower mids and upper bass which didn't need any correction.

    Mid +4
    upload_2022-10-5_21-15-0.png
     
  4. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,771
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    So I figured the next step was the back-off the mids (some things just cannot be fixed, but let's compromise with what we have) and drop the highs. This time around, let's do +2 mid and -3 high.

    Mid +2, High -3
    upload_2022-10-5_21-20-32.png

    I had intended to do -4 or -5 on the highs, but you guys get the idea. FWIW, before I even measured anything, I had already dialed in by ear: Mid +2 and High -4. It's nice to see measurements confirming. As for the bass, I am leaving it as is because of the aforementioned issues with road and exhaust noise. Besides I'm a closet basshead (as long was it doesn't bleed into the midbass and up).

    Some takeaways:
    1. I am absolutely sure there's some EQ going on in the factory HU or amp because that kind of extension to 20kHz is not possible with dash mounted tweeters (even if they are pointing up at and bouncing off the windshield). I will measure to confirm.
    2. I am curious if dropping the xover point on the tweeter would help fill in the mids. Pretty sure the factory tweeter is going on have a single inline cap on it. A bigger cap should do the trick. The front speakers (dash and door) are sent via a single wire pair per side, so I know the tweeter and woofer do have not have separate amps with active crossovers. I am going to pull the factory tweeter on the car to examine any xover components, measure it's sensitivity, impedance, etc. so that I can purchase a suitable tweeter. (I plan on reusing the woofers because they actually sound pretty good and mate well with the characteristics of the door). I am hoping the 93db sensitivity Morel will be a good fit.
    3. I may test fabric materials in front of the tweeter as a means for physical EQ, seeing if I could get different results.
    To be continued. Car audio is a pain in the ass these days.
     
    • Like Like x 4
    • Epic Epic x 1
    • List
    Last edited: Oct 5, 2022
  5. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,771
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    Quick note on the Veloster N drive modes and exhaust. The Elantra N has the same motor and modes (the turbo is bigger, but Hyundai nerfed it to have same HP as Veloster). The looks of the Elantra have grown on me since I've been seeing them in person. They look better on the street than in photos.

    Here's to California:
    https://www.motor1.com/news/614016/cop-elantra-n-owner-stock-exhaust-loud/
     
  6. Kernel Kurtz

    Kernel Kurtz Friend

    Pyrate Contributor
    Joined:
    May 19, 2018
    Likes Received:
    1,677
    Trophy Points:
    93
    Location:
    Winnipeg, Canada
    There probably is. On my car manual EQ is very limited, but you can choose between a number of different built in sound profiles with appropriate coding software. The same amp is used in multiple models so there is some DSP going on for all of them, it's just not granular enough beyond that.
     
  7. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,771
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    So I took the tweeters out. I was fully prepared with the service manuals, trim removal tools, mechanics tool set. I had forgotten how such endeavors always lead to F-bombs that all the neighbors can hear. At least the weather is good outside. The last time I did this was when I was 18 or 19. (I ripped apart the interior of the car to install power from the battery, amps, speakers subwoofer box, etc.)

    Dealing with the corner dash was a bitch. The tweeter panels were so flush with the top of the dash panel that I could not jam a trim remover into it. I had to first

    remove the trim panel under the A-pillar. What made things hard was the screw in a hard to reach place. I was eventually able to angle a screwdriver in there without stripping the screw.

    PXL_20221006_151714421.PORTRAIT (1) (Medium).jpg

    Easy after the screw is removed. The funny thing is that the service manual says (in an unclear way) to just rip this part out from the clip and get a new one. I got what the manual meant after seeing this.
    PXL_20221006_165027023 (Medium).jpg

    Even after I got this part off, the tweeter grill would not pop off easily. I did find a trick: stabbing the ice pick tool into the corner of the grill, pressing it in, and lifting it up. Super easy.
    PXL_20221006_165114310 (Medium).jpg

    The rest came out with a socket wrench (a low profile ratchet would have helped immensely) near the windshield. I thought this would be another pain in the ass because of the aggressive rake angle on the Veloster N's windshield, but it was straightforward.

    Here's the back of the tweeter grill. If I got larger tweeters, I'd need to pull out the dremel to cut out that cylindrical lip. I do like what Hyundai did here. They don't allow the tweeter to just play into a grill, but focus the sound outward.
    PXL_20221006_153654957 (Medium).jpg
     
  8. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,771
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    The tweeter actually has a foam ring around it that sits on the lip on the inside of the grill. Yes, it's rigged to my impedance measurement instrumentation (ignore the AD converter). I wanted to get an impedance measurement to know the nominal impedance and to also get an actual curve in case I wanted to do some tweaks or do something super fancy.
    PXL_20221006_225728085.jpg

    Hyundai Veloster N
    Front Dash Tweeter
    Impedance (ohms) vs Frequency
    upload_2022-10-6_21-5-2.png

    The tweeter actually ain't bad. It's real tweeter. 4-ohms nominal instead of 2-ohms. It's cheap, but it's decent and not a POS 2-3" paper cone found in most other "premium" car audio systems. The resonant frequency is higher than I'd like at 1.6kHz for a tweeter; but heck, it's cheap.

    So I guessed it. Inline to the tweeter is a single cap. A Bennic electrolytic 4.7uF special. It could have been much worse part, like a China special. Bennic is decent. I appreciate Hyundai for not participating in a race to the bottom because I've seen worse on fancier auto brands. With the nominal impedance around 4.7uF and 4-ohms more or less, I'd gather the xover point is 8kHz. With a single cap, it's first order with a slow roll-off. Still 8kHz is super high for a tweeter. I'd surmise that the woofers door behave well (or are EQ's into behaving well) high up. Or most likely, the center speaker (I have not popped this open yet) is filling in to lower treble.

    Here's the impedance measurement with the cap in place. Note the scale is different.

    Hyundai Veloster N
    Front Dash Tweeter (with inline cap)
    Impedance (ohms) vs Frequency
    upload_2022-10-6_21-14-48.png
     
    • Like Like x 1
    • Epic Epic x 1
    • List
    Last edited: Oct 6, 2022
  9. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,771
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    Regardless the tweeter did sound grainy and scratchy so I decided to solder on a film cap in parallel to the electrolytic. This is noname 1uF part from China (that sounded good enough to be used as the bypass cap in the EC ZDS power caps and feedback loop electrolytic). Normally I dislike doing this (paralleling mediocre caps with good quality small caps aka "bypass" with speaker xovers, but this isn't exactly a high-fidelity application. Anything will most likely improve the sound.

    PXL_20221006_234523444.PORTRAIT.jpg

    Now this does change the overall capacitor value now to 5.7uF. I figured would move the xover down in frequency just a bit, maybe filling in that midrange hole. The downside is because the stock xover on the tweeters is so high in frequency, I may end up bringing up the upper mids. Heck, I was willing to take the risk because the HD600 and HD650 (headphones that I love) do have slightly elevated upper mids in specific spots. This of course is all conjecture because it's hard to say what would happen because of xover phase interactions, especially from a midrange center speaker and two woofers at the doors.

    Finally, I cut a small piece of mask material and inserted it into the grill. No pictures. I didn't want to cut the highs too much because the EQ seemed to be doing a good job. I could afford to lose a bit of air (the stock audio sounded sounded too airy to me and I dislike this). The idea was that I wanted to smooth out the highs and cut down that sizzle and zippiness. It's a light touch. I didn't want to do a Mr. Speakers Doggy Treat that cured the ills but killed the patient.
     
    Last edited: Oct 6, 2022
  10. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,771
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    So I fired up the car stereo and played back some tunes from Spotify. Yeah, this is really nice. The gritty graininess of the highs is much improved. The overly zippy transients are no longer, to the point where I wonder if I will miss them. (In the long term, when I blast the radio, I know I won't). The problem is now that things are smoother sounding, it's more obvious of how mediocre these tweeters (and the door speakers, and the mid speakers sound). Again, from a technicalities view, the Veloster N's car stereo is still better than most other systems. And now from a tonal response and timbre, far superior. I actually dialed in slightly different EQ settings this time around. As I mentioned earlier, changing the xover frequency may do unexpected things. Let's at a look at some measurements.

    upload_2022-10-6_21-45-52.png

    I don't think there is a big difference. I replicated the conditions the same as best was I could. These measurements were done with averaging where I waited until they were very stabilize and reproducible. I would say that with the tweaks (red line), the frequency response is smoother. The last octave is smoother without the small dip at 12kHz and bump at 16kHz. There dip at 2.8kHz is filled in albeit there is now a small bump 2kHz. However looks smoother. The lower xover point does seem to lift 800Hz to 4kHz just teeny bit. Unexpectedly 400Hz are 800Hz are lowered (and this isn't necessarily a bad thing when we recall that the Mid EQ setting has its affects well into the lower mids.
     
  11. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,771
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    Here are the results with the tweaks and some EQ. The EQ settings are now -3 highs, +3 mids, and -1 lows. Note that without the tweaks (material over tweeter and xover cap bypass and value increase), I typically dialed in -4 or -5 on the highs. Also before, I used +2 on the mids. I found that could away with +3 on the mids because with the tweaks, the upper bass / lower mid emphasis wasn't as strong as before. Also, since the highs aren't so nasty, I dialed down the bass just a bit.

    upload_2022-10-6_22-11-35.png

    Final result, for now...
    upload_2022-10-6_22-15-12.png
     
    • Epic Epic x 4
    • Like Like x 3
    • List
  12. penguins

    penguins Friend, formerly known as fp627

    Pyrate
    Joined:
    Sep 26, 2018
    Likes Received:
    3,451
    Trophy Points:
    113
    Location:
    SouthernCA
    Some BMWs have this stock now though...

    Anyways, this is nervosa level, but do you think the B&K target may have changed given that cars were also pretty different inside in the 70's vs today + music is different, etc. Agreed still better than Harmon curve or whatever although I also wonder if people take the Harmon curve more seriously than it's actually intended to be.
     
  13. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,771
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    Not really because the B&K target is taken at the listening position. So if cars were different back then (more glass, less sloped glass, harder materials), it doesn't matter. The measurement is taken where the listeners head is at. The B&K deal is not just a target, it's also a method that takes into account the location of the listener and the surroundings.

    It just so happens that speakers that measure flat on-axis nearfield tend to follow B&K or the 1db per octave downward slope when measured at the listening position in a typical room (like with furniture).
     
  14. LetMeBeFrank

    LetMeBeFrank Won't tell anyone my name is actually Francis

    Pyrate
    Joined:
    Aug 4, 2017
    Likes Received:
    3,758
    Trophy Points:
    93
    Location:
    Jackson, Mi
    I was having a similar issue with my 300 recently after replacing a pioneer head unit with a Sony XAV AX5500 for Android auto. Right away I noticed the highs were boosted. I looked at the eq settings and didn't see anything in there, so I dropped everything from 1k to 20k by 6db (the lowest you can go) and it STILL sounded bright, but also muddy as hell. I tweaked it into a slope and got it sounding mostly ok but it was still piercing in the higher frequencies.

    I searched high and low online, eventually deciding to read the whole manual, and I found the problem. Sony ships these radios with DSO on by default, but it's not in the EQ settings. From Sony "The DSO (Dynamic Stage Organizer) creates vivid sound image as if there were speakers on the dashboard." I already have dash mounted Infinity 3.5" coaxials with 1" tweeters, so this setting made it sound horribly bright. This should be in the EQ settings as a switch, on or off, and it should NOT be on by default. So dumb.

    Now that it's sounding good I'm super happy with it though.

    [​IMG]
     
  15. Kernel Kurtz

    Kernel Kurtz Friend

    Pyrate Contributor
    Joined:
    May 19, 2018
    Likes Received:
    1,677
    Trophy Points:
    93
    Location:
    Winnipeg, Canada
    Yes, my 240 pops and burbles on lift right from the factory. It also came with "Active Sound Design", basically fake engine noises that play through the speakers. Sadly a lot of cars have this sort of thing nowadays. Coding that out was the first thing I did when I got the car. I actually like the real engine noises betterer.
     
  16. penguins

    penguins Friend, formerly known as fp627

    Pyrate
    Joined:
    Sep 26, 2018
    Likes Received:
    3,451
    Trophy Points:
    113
    Location:
    SouthernCA
    I still need to block those noises out myself.
     
  17. Kernel Kurtz

    Kernel Kurtz Friend

    Pyrate Contributor
    Joined:
    May 19, 2018
    Likes Received:
    1,677
    Trophy Points:
    93
    Location:
    Winnipeg, Canada
    I think I remember you have an M240 as well. Bimmercode on your smart phone makes it easy. I've recently moved on to Protools and it is scary how many parameters you can modify at will. With great power comes - great fear of borking something LOL.
     
  18. cyypher

    cyypher New

    Joined:
    Nov 27, 2023
    Likes Received:
    0
    Trophy Points:
    1
    Location:
    UT

Share This Page