Which sound card to get...

**netchris** · 01-24-2007, 08:33 PM

Ok well I found a FIR filter, so maybe this thing could be used for active crossover, but I'm not the expert here, sooo....

http://www.rekkerd.org/sweetboy-vst/ (it's listed as T-SLEDGE)

How much difference would FIR really make?

**XC-C30** · 01-24-2007, 08:57 PM

Originally Posted by durwood

Not sure what you mean by vertical vs horizontal but time delay makes your car sound really good from the seat you tune it from. It's what we call a one seat tuned car. Bad for your passengers but good for you

If you want it to sound the same on both sides of the car, you have to go kick panels or horns. There are tradeoffs for both.

It won't sound horrible to your passenger, but it will sound like all the sound comes from the right side of the car to the passenger if you go the time delay route.

you might point and tune your tweeters to the center of your front window, right under your rearview mirror.... that way I think it will sound best for both driver and passenger.... It maybe won't sound as good as when you would locate the stage right infront of the driver, but this way the passenger also will have a fairly nice soundstage..... So unless you're competing, this -to me- seems the best option. Wether you go with on-axis or off-axis tweets, well, that's an other story, but you already know that, Durwood

**netchris** · 01-25-2007, 12:09 AM

It wouldn't be best to point both tweeters right between the heads of the driver and passenger? Why would I point the tweets right below the rear view mirror? I never have my head there..

**Abmolech** · 01-25-2007, 12:41 AM

Have a read up of ambiphonics and crosstalk.

**netchris** · 01-25-2007, 01:10 AM

I'll definitely look into it! I was under the impression that on-axis listening was the best though, and I (maybe wrongly) assumed that on axis = pointing the speaker at your head?

**Abmolech** · 01-25-2007, 01:44 AM

On axis is parallel to your eyes.

Crosstalk is the amount of similar signal between (normally) stereo chanels.

Example most "engineered" stereo (monophonic recording into stereo) has 60% (electrical) similarity of the channels. This is ideal of maintaining a centre image.

Depending on who you talk to 70 % and above is considered monophonic.

Acoustic crosstalk is the same imformation acousically summing. Stereophonic use this to give the impression of left right at lower (time sensitive) frequencies. A "proper" stereo recording .

When listening to a pair of speakers in a room both ears will hear the signals produced by both the left and right loudspeakers. The timing differences associated with this acoustic 'crosstalk' between the two channels and each ear lie at the core of the 'stereo illusion'. This is what allows us to perceive phantom images between the speakers, and coincident-microphone and stereo-panning techniques (which employ only level differences between the two channels to convey the spatial information) rely entirely on this acoustic crosstalk to work properly. A large portion of the left speaker signal will go to the right ear of the listener, and similarly a large portion of the right speaker signal will go to the left ear of the listener

When listening to ordinary stereo material via headphones, this interaural timing information is missing — we have only the differences in level between the two channels to go on — and hence the stereo images become non-linear and ill-defined. In fact, most people perceive the individual sound sources to lie on a line running directly through the centre of the head, instead of being portrayed in front of us as they would be with loudspeakers.

While it is true that if coincident microphones of 20 cm apart, you should have 580 millisecond delay between left and right microphone, to simulate a 90 degree azimuth, crosstalk will reduce this to 250 milli seconds. This results in a 30 degree loss in separation down to the recommended 60 degree speaker position, allowing the correct placement but with the loss in depth. Some have tried to reduce this crosstalk with various DSP logarithms, producing an increase depth, but at a serious loss with a VERY tiny sweet spot and any head movement completely collapses the illusion.

Thats stereophonic in a "nut shell".

**netchris** · 01-25-2007, 03:37 AM

Ah very nice information! It definitely gives me new perspective.

I've learned a lot about speaker selection and how to set up an active crossover, but I now realize that I took speaker placement for granted. I thought for sure the best way to aim the tweeters was just to point them at your head, but clearly I was wrong. I do understand a lot of what you're saying, but some of it escapes me.

While it is true that if coincident microphones of 20 cm apart, you should have 580 millisecond delay between left and right microphone, to simulate a 90 degree azimuth

But isn't 580 milliseconds at the speed of sound more like 200 meters and not 20 centimeters? It takes .588 milliseconds for sound to travel 20cm. I'm confused..

crosstalk will reduce this to 250 milli seconds.

Why is that? Is it due to the acoustic summing? Is it just the listener's perception? Are acoustic summing and the listener's perception the same thing?

This results in a 30 degree loss in separation down to the recommended 60 degree speaker position, allowing the correct placement but with the loss in depth.

So why is 60 degrees the correct speaker placement if there is a loss in depth?

Thank you for opening my eyes to the importance of proper speaker placement. It's another facet of car audio and even audio in general that I will have to learn, but I had better know it before I plunk down a load of cash on a system!

**XC-C30** · 01-25-2007, 04:17 AM

Originally Posted by netchris

It wouldn't be best to point both tweeters right between the heads of the driver and passenger? Why would I point the tweets right below the rear view mirror? I never have my head there..

your soundstage has to be in front of you, and preferably as high as possible... (about equal to your ears).... If you go to a concert, you won't go standing on stage where the musicians are, just because that's where the sound is coming from, do you?

**durwood** · 01-25-2007, 07:53 AM

Speaker placement in a car is a total guessing game. If you ask any top SQ competitors they will tell you it takes time for deciding on placement. You have many different surfaces the screw with the sound. Some are hard surfaces which reflect the sound (Glass) some are soft surfaces that absorb (padded dash, seats, etc). Most people when installing tweeters or kick panels will temporarily place the speakers until they find the right position or placement. I ended up on axis because I didn't like the way they sounded off-axis. Every car is different.

Ambolech as always with great super technical info.

Sometimes you really put my brain to work.

There's probably something wrong with my reasoning here, but why couldn't we just make a hard cut band pass filter right at our crossover frequencies with no attenuation or dB slope at all? For example, 99Hz would come out of the sub at full volume, and 100Hz wouldn't come out of the sub at all.

What you speak of is an "ideal" filter but it's near impossible to do such a thing. The closest you can get is with a really steep slope i.e. 72 db/octave or more. FA will do up to 42 db/octave which is still incredibly steep. 24 db/octave is popular due to its a complete 360deg out of phase and you are less likely to notice phase issues. If you design crossovers right though, sometiems the phase changes can work to your advantage if you need some cancelation at a particular frequency IMO.

It is yet to be decided if the Full version of FA makes that big of a difference with the phase correction vs the Lite version without (Phase Arbitrator). The developer of FA seemed to think I wouldn't need it, but I chose to buy the full version anyway and I will test for myself. I bought before xmas though and got a discount and the lite version was also included.

**Abmolech** · 01-25-2007, 11:23 AM

Depth is "lost" because the acoustic summing (crosstalk) changes the perception of distance for the bonus(?) of width. That is instead of 90 degrees it is 60.

Turning a driver off axis changes the amplitude/phase. This can help with the perception of localisation of sound.

Full "brick-wall" linear phase crossovers are available, see
DEQX, TACT, Lake etc.

Anytime you apply a filter (crossover EQ etc) it has to delay the other frequencies (phase). Logically you can't speed them up (hence time delay is is sub set of phase adjustment), the amount of delay depends on how much you attenuate them. For example a 6 degree slope = 90 degree phase shift, 12 degree slope = 180 phase shift.Etc.

If you have all your drivers on the same filter slope then they are all delayed by the same amount. The "interesting" stuff happens when you use different slopes.

Advantage of less degree slope:

1\ less phase shift
2\ "forgiving" matching of drivers. If you could imagine a pyramid the "base" section is made by the sub (no pun intended) then the woofer makes the next section up etc.

Your crossovers and time alignment determine how well each section of the pyramid aligns itself. 6 degree slopes allow much more misalignment, because the edges are "feathered".

As you increase the slope, the less forgiving the alignment.

Disadvantage of flatter slopes, is the driver has a much greater expected bandwidth to reproduce. This can put them into nasty cone breakup etc. The other problem is different slope matching in the crossover area, is considerably prolonged. (that is two drivers are playing the same frequency, and often don't blend well)