25 replies to this topic

[mrock]

I must say that I have noticed that the dsp chips used in today's head units are of very poor quality and they really do degrade the sound. They don't help much at all.

[shizzle]

looks like you answered your own question

[mrock]

just want to see what other people's opinions...lol

[~thematt~]

It is my opinion that yes, you answered your own question

[mrock]

well, it's good to know...Then if the majority thinks like that why don't the manufactorers stop installing them and spend more money on better internal components?

[Fury♫]

Would you (the average consumer) pay $1500+ for a headunit?

[Pulse-R]

I agree that many newer head units have poor quality sound processing capabilities.

I do not agree that it is the DSP chip causing the 'problem', but rather the budget code used to effect the 'adjustments', and the budget design implemented.

Every DSP chip is pretty similar internally, it has inputs and outputs, and firmware inside to make it do what it's told.

The problem arises when a programmer has to cram in more features. This requires more code and more processing speed.
Head unts and processors used to have a dedicated DSP chip for processing, but now they have the DSP doing EQ+TA+XO and then throw in a few codecs as well, and bluetooth, and iPod control.. all of a sudden, the ancillary functions need so much space there's very little left for any audio adjustments. A good example of how it used to be done is the Alpine PXA-H701. It has 2 DSP chips for time alignment and equalisation. A seprate DSP/CODEC chip is used for Digital audio decoding and crossovers.

Given that a processor only has so much program memory, it's not hard to see that with all the extra 'features' in modern head units, there has to be a trade-off. This tradeoff is in the precision of the complex calcuations used to perform EQ and crossover functions. To do it right requires 10x more processing than the approximations used in many budget implementations.

The truth is, cheap sells and most people can't hear the difference anyway - or don't care.

- my 2c.

[Big_Valven]

Good points Pulse.

I think the stickiest point here is that it's hard to actually pinpoint the DSP chip as being at fault. Thing is, there is any number of points in a headunit where manufacturers can cost-cut, and more often than not a DSP can degrade sound quality without actually setting TA and EQ.

A unit could have one of the most cutting-edge DSPs available, and still be let down by rubbish conversion or line drivers, for example.

To directly answer your question though; my DSP doesn't affect sound quality one bit... but that's cos I don't have one... yet

Edited by Big_Valven, 16 May 2010 - 09:14 PM.

[TMM]

Big_Valven, on 16 May 2010 - 09:13 PM, said:

To directly answer your question though; my DSP doesn't affect sound quality one bit... but that's cos I don't have one... yet

+1. Best to just avoid any processing unless you can afford a decent one.

[mrock]

Why use didgital when you can alter it passivly...

[Big_Valven]

You're not serious are you?

Some digital processors may affect sound quality but passive networks affect it MUCH more, even if in different ways.

[mrock]

I am serious... For me it's easier to get the results, plus there is no fluctuation( you can alter the frequency you want). And you can get exactly what you want. Why are passive components so bad? Your speakers will never be 100% distortion free. Plus some of the dsp's actually introduce their own distortion. And the options are endless without costing you an arm and a leg. You have choices of 6,12,18,24 dB, butterworth, linkwitz, bessell, cheybesky, series notch and parallel notch, zobel, and the best one of them all "Contour". With the electronic devices, you are limited.

[br85]

I'd say passives are FAR more limited since you're stuck with the settings you paid for.

Active processing for similar $$$ is much more flexible, and doesn't suck up power or create phase anomalies to the same extent that passives do.

And what do you mean there's no fluctuation? There's HEAPS of fluctuation in passive components over time, when the weather changes, etc. DSP units have no such problem.

[mrock]

By fluctuation( I worded it incorrectly), You can alter the exact region you want to. Also, I don't mean by bying off the shelf crossovers. Forget that rubbish... I mean putting them together yourself. And it's not expensive going to Jaycar and doing alterations. Give me an example on how would you smooth out the peaks on a piezo tweeter with an active eq an crossover? It's not an easy task. This way I just bay a cap and resistor and whola...Bobs your uncle.There is alot of experimenting, but that is what makes it fun.

Edited by mrock, 21 May 2010 - 09:06 AM.

[br85]

I see you aren't aware of active notch filters. Hard to get 240v stuff working in a car but there is plenty of people who've done it. I stand by my statement that your jaycar caps, resistors and inductors will change values and shift frequency and q over time, but dsp will not. I didn't think you were talking about off-the-shelf crossovers, by the way.

MOTHER OF GOD I WANT ONE, actually make that I WANT ONE HUNDRED:

DEQX active crossover and everything under the goddam sun you could ever ask from dsp

Ok, I get that it's super expensive, but there are people who've spent WAY more than 10k on a system before, it's not that far out of reach.

Attached Files

•  hdp3.gif   5.48K   0 downloads

Edited by br85, 21 May 2010 - 09:55 AM.

ss-rotel, on 14 September 2010 - 11:05 PM, said:

you dont some much hear, but fell the sound

• Back to top