24bit/192kHz Digital Crossover?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
The miniDSP chips are capable of doing 24/192 it's just that, as abraxalito says, it eats at the resources.

Personally I don't understand why the miniDSP guys don't just allow you to use higher sampling rates. Sure you might run out of computational power, but if you do you can easily swap to something lower.

The 2x8, as far as I am aware, uses the top line sigmaDSP line of chips from analogue devices and these are more then capable of doing 24/192 for a 4 way active system.

Pardon my beginner-ness, but wouldn't it simply be a matter of upgrading the microprocessor and memory to something more powerful that can handle 192kHz operations? Anyway, looking at blu-ray stats, there are only a couple flicks that actually run at 192kHz, with the majority at 96 or below.

The next question then, and I realize this is now off-topic and in the wrong section, is which would be better: home theater processor -> passive xo -> single amp -> speaker? home theater processor -> active xo -> bi amp -> speaker?

The main issue would be that by using an active XO, there would be another analog to digital conversion, but I'd have the ability to fine tune any crossover. The home theater processor has built in room EQ so would this negate any benefit of using an active XO?
 
Pardon my beginner-ness........

The main issue would be that by using an active XO, there would be another analog to digital conversion, but I'd have the ability to fine tune any crossover. The home theater processor has built in room EQ so would this negate any benefit of using an active XO?

Hi Frosteh,
A good active crossover or well designed passive crossover is not the issue.

Room EQ has absolutely nothing to do with the crossover which controls the individual drivers output. That said of course the EQ can boost or cut frequencies otherwise finely honed by your crossover, passive or active.

Room eq is a minefield or a lottery, it will mess with the (hopefully) carefully crafted output of the speakers in an attempt to flatten the frequency response in your listening room, that is according to your best attempts with microphone positioning for given listening positions. Sometimes it can be helpful in the bass.

It looks like you are obsessed with audiophile notions or concepts of purity without the most basic understanding what you want to build or control. I think you should do a lot more reading before tackling any building work.
Even 24bit audio is not proven to be superior to 16bit for playback, although the extended data available on blu-ray discs often brings better sound:
LAT News The (non)sense of high definition audio
also:
24/192 Music Downloads are Very Silly Indeed

Incidently active can be digital or analogue, my Pluto's use analogue signal processing for instance.
 
@frosteh

something like this ?

- 2 channel in, and 6 channel out.
- Input analog, digital AES/SPDIF, USB audio input.
- Control by USB and PC software.
- Resolution, 24 bits / max 192 kHz sampling.
- A/D converters AKM, AK5385A. D/A converters AKM, AK4395.
- DSP TI , optional adding a second one to increase processing power.
- Power supply only 2x12V, all regulators low voltage on board, optional use of our HxR audiophile regulators.
- Special clock design with extremely low jitter.
- Use of sample rate converter.
- Special gain section to adjust at several different sensitive. Guaranteed highest SNR level.
- Optimized for low EMI.
- Size only 105x105mm.
- Optional IR remote control.
- Optional graphical display.
- Optional input relays section to use it as a preamp.
- Digital volume control, we will write our own firmware, target is to have performance the same as with an analog volume control. Will definitely sound much better as the PGA volume control chips!
 
Last edited:
Pardon my beginner-ness, but wouldn't it simply be a matter of upgrading the microprocessor and memory to something more powerful that can handle 192kHz operations? Anyway, looking at blu-ray stats, there are only a couple flicks that actually run at 192kHz, with the majority at 96 or below.

Both yes and no. All the sigma DSP chips (as far as I am aware) are capable of working at 192kHz. The memory usage, per filter section, is the same regardless of the sampling frequency.

The chips themselves, run at a specific frequency and this is exactly the same whether or not you're using 48kHz or 192kHz. Just for the sake of argument and this calculation wont be accurate by any means but is fine for this example, but lets say that the sigma chips core runs at 192kHz, we also want to do processing on a 192kHz stream. As the core is running at the same rate as the data is coming in, you can only perform one calculation per sample. Now you swap over to 48kHz, as the core rate = 4x the sample rate you are now able to do 4 calculations per sample. If each calculation represented a standard biquad filter then you'd be able to produce a 12dB/octave electrical filter at 192kHz, but at 48kHz you'd be able to cascade 4 filter blocks to produce a 48db/octave electrical filter.

Now the miniDSP products use different chips from the sigmaDSP family. The standard version is set to work only at 48kHz, it will do 192kHz, but as it's core runs 'slow' it wont allow you to do much processing when running at 192kHz. Interestingly though, you'd be able to run most standard two way loudspeakers through the basic miniDSP product and still have enough processing power, it wont do much beyond this though. As it's easy enough to switch between different sample rates in the software, I don't know why the miniDSP guys feel it's necessary to impose a 48kHz limit. Most of the people using the miniDSP products are more then capable of realising that they are running out of processing power and should switch to something lower, or reduce the filter count.

One area where 192kHz is appreciably better then 48kHz is with digital delay. You can only delay by a minimum of 1 sample period and at 48kHz this = ~20us, this is far too long a time to make accurate adjustments for time aligning. 5us however, as offered by 192kHz is good enough.

The 2x8 and 8x8, using the fastest of the sigmaDSP chips, is basically what I've built for myself. I run mine at 192kHz and also use it to actively crossover a 4 way. This is a reasonably complicated design and I've only used up around 40% of what the chip is capable of. If the 2x8 really does use the ADAU144x family of chips then it is more then capable of working at 192kHz.

The next question then, and I realize this is now off-topic and in the wrong section, is which would be better: home theater processor -> passive xo -> single amp -> speaker? home theater processor -> active xo -> bi amp -> speaker?

The main issue would be that by using an active XO, there would be another analog to digital conversion, but I'd have the ability to fine tune any crossover. The home theater processor has built in room EQ so would this negate any benefit of using an active XO?

The answer to this is fairly simple. Do you need the extra capabilities of the DSP unit in the system to allow you to use your loudspeaker design properly? A passive version of what I'm doing just wouldn't work, so for me it would be a no brainer. The trouble with your situation is that adding in a DSP unit under those circumstances invites the same kind of problems that the Behringer DCX2496 is known for - ie altering the volume in some way before the final D/A process. What you really need is an 8 way volume control, such as the one designed by Jan Didden, that is retro fitted into the DCX2496. This then allows the gain of the entire audio chain to work at a high level, both in terms of the ADC before the DSP and the DACs that run after the DSP, as you only alter volume right at the end of the process.
 
Hi Frosteh,
A good active crossover or well designed passive crossover is not the issue.

Room EQ has absolutely nothing to do with the crossover which controls the individual drivers output. That said of course the EQ can boost or cut frequencies otherwise finely honed by your crossover, passive or active.
Sorry I should have clarified. I know that EQ does nothing regarding which frequencies go to which speaker (hence the point of a XO). The advantage to most active XO's is that you can fine tune the various blips in response given a certain room. To do that in a passive XO is nearly impossible without redesigning part of the XO for new components. However, like you said, room EQ does allow you to vary the levels of each frequency, thus negating the need for an active XO, provided a basic passive XO is already in place. With this in mind, I'm wondering if there's any advantage to using an active XO over a passive XO when room EQ and processing is available.
Room eq is a minefield or a lottery, it will mess with the (hopefully) carefully crafted output of the speakers in an attempt to flatten the frequency response in your listening room, that is according to your best attempts with microphone positioning for given listening positions. Sometimes it can be helpful in the bass.
This is exactly why I'm wondering if it makes a difference between a good passive XO or an active XO if we're going to be messing with room EQ anyway. I know it's often a crap-shoot, but that's how it will be with an active XO anyway.
It looks like you are obsessed with audiophile notions or concepts of purity without the most basic understanding what you want to build or control.
This has been debated many, many times before. I, along with the others who go the 24/192+ route, could care less whether or not it's been proven to work better; we've likely read about the debate but still choose to use it. I just want to know that I have it if I want to use it if it someday ever is proven to be better (Or not! I don't care!). Please understand this before knocking around my understanding of audio.

@lduarte: No link! :p
 
Hi Frosteh,
Sorry if part of my response sounded a little snooty.
The benefits of "Tuning the blips for a given room" (as you say) isn't the only benefit of active design. Active crossovers can be far more flexible than passive.
A stable active high order crossover will remain useful for the highest possible sound quality for given drive units even when the processor's DSP has tailored the frequency response. Obviously we want the crossover to remain a constant.

An active crossover has theoretical benefits at the high volume levels experienced in Home Theatre playback, probably more so in systems with smaller less sensitive drive units. ATC have produced interesting material on this.
 
A stable active high order crossover will remain useful for the highest possible sound quality for given drive units even when the processor's DSP has tailored the frequency response. Obviously we want the crossover to remain a constant.
I don't really know what you're getting at here. How would the crossover change with DSP enhancements?

I don't plan on changing my speakers, and for simplicity's sake let's assume I won't, so is there any flexibility advantage to the actives considering this?
 
However, like you said, room EQ does allow you to vary the levels of each frequency, thus negating the need for an active XO, provided a basic passive XO is already in place. With this in mind, I'm wondering if there's any advantage to using an active XO over a passive XO when room EQ and processing is available.

This is exactly why I'm wondering if it makes a difference between a good passive XO or an active XO if we're going to be messing with room EQ anyway. I know it's often a crap-shoot, but that's how it will be with an active XO anyway.

"However, like you said, room EQ does allow you to vary the levels of each frequency, thus negating the need for an active XO, provided a basic passive XO is already in place"
I didn't say this.

The benefits of going active should be exactly the same as any other time. The only difference being that the mic on a HT processor is meant to correct for perceived imbalance so theoretically you would not need this if they worked, producing a balance you actually preferred.
Not all active give you extensive EQ control options.
 
I don't really know what you're getting at here. How would the crossover change with DSP enhancements?

I don't plan on changing my speakers, and for simplicity's sake let's assume I won't, so is there any flexibility advantage to the actives considering this?

I was referring to the DSP in the processor (room correction on the processor) not anything in the speakers. In other words just because you have changed the input signal this does not detract from the benefits of the active crossover.




Processor - Active crossover - amp's -direct to drive units, no passive crossover components in the way.
 
Last edited:
*Modified for clarification (hopefully!)*
The benefits of going active should be exactly the same as any other time. The only difference here (compared to a straight stereo system) being that the mic on a HT processor is meant to correct for perceived room acoustics imbalance so theoretically you would not need to modify the tonal balance with the speakers own EQ settings if the room correction toys on the AV system worked perfectly, producing a balance you actually preferred, without the need for extensive tone control options on an active speaker (just one small potential benefit/feature of actives).
Not all active give you extensive EQ control options.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.