Active Designs & Their Favorability

[Frank Dernie]

One thing I forgot to consider, is that with the DSP models, there is an additional latency added to the signal chain. Genelec claims that it's a constant 3-5ms depending on the model (3ms for 8330A, 5ms for the 8351 IIRC) with or without GLM calibration applied.

I don't know how significant this kind of delay would be when it comes to audio cues in say, gaming.

EDIT: Well a quick test showed that it was fairly easy for me to detect a 2ms timing difference on audiocheck, with 10/10. Wonder if that applies to this situation, though.

This is the reason I haven't bought a modern dsp speaker to improve the dialogue channel in my home theatre.

 

[ElNino]

One thing I forgot to consider, is that with the DSP models, there is an additional latency added to the signal chain. Genelec claims that it's a constant 3-5ms depending on the model (3ms for 8330A, 5ms for the 8351 IIRC) with or without GLM calibration applied.

I don't know how significant this kind of delay would be when it comes to audio cues in say, gaming.

EDIT: Well a quick test showed that it was fairly easy for me to detect a 2ms timing difference on audiocheck, with 10/10. Wonder if that applies to this situation, though.

Keep in mind that modern TVs average around 9-10ms of latency for video processing. So there's no problem integrating minimum phase DSP (typically 3-5ms) into modern video rigs -- the DSP is faster.

 

[lc155]

Keep in mind that modern TVs average around 9-10ms of latency for video processing. So there's no problem integrating minimum phase DSP (typically 3-5ms) into modern video rigs -- the DSP is faster.

And how would that compare with modern PC gaming rigs, with 120Hz+ monitors?

 

[ElNino]

And how would that compare with modern PC gaming rigs, with 120Hz+ monitors?

Refresh rate is actually a separate concept from the display's internal video processing latency. Good IPS gaming displays are still in the 9-10ms range for video processing latency. A lot of older (non-CRT, obviously) displays have considerably longer processing latencies (15-20ms is not uncommon).

There are displays where you can disable virtually everything in terms of processing. Then you can get closer to 2-5ms on IPS panels on some displays, which is actually a good match for minimum phase DSP.

 

[Purité Audio]

This is the reason I haven't bought a modern dsp speaker to improve the dialogue channel in my home theatre.

Modern ones have low latency modes, the Kiis is infinity variable, and you can set up as a preset, the 8Cs are fine in their full on latency but they have a low option too.
Keith

 

[Kal Rubinson]

For some reason I thought that studio monitors tweeters are designed to be more direct like a beam , and hifi speakers tweeters are designed to do the exact opposite so you won't be locked into the "sweet spot"

This is a variable for all speakers and, of course, it is measurable.

 

[andreasmaaan]

To give some perspective to what latencies of 3 or 5ms are like perceptually: If you play the trumpet, the "latency" between the time you blow (buzz) into the mouthpiece and the sound emerges from the horn at the other end is about 4ms.

 

[lc155]

To give some perspective to what latencies of 3 or 5ms are like perceptually: If you play the trumpet, the "latency" between the time you blow (buzz) into the mouthpiece and the sound emerges from the horn at the other end is about 4ms.

Yeah, I since read that it takes 3ms for sound to travel a metre, so the added delay by DSP would just be like sitting slightly farther back. I don't think any couch gaming PC folks have ever complained of delay resulting from that.

 

[ElNino]

I do think it's worth ensuring that you have audio delay compensation so that the audio matches the visual, but it's easy to do this on modern TVs even if your DSP doesn't offer delay functionality by itself. (It's a different story obviously if you're doing complex FIR-based EQ with longer latency than the display processing.)

 

[q3cpma]

GLM is just a simple parametric EQ, there is nothing "advanced" about it.. sorry for perhaps pointless correction, just caught my eye..

Doesn't it correct phase too? That'd put it at least one level above basic for me. I do also remember something about working with multiple subs, which only Dirac with an extension costing a pretty penny has too; though it doesn't correct the final summed response, I think?

 

[Frank Dernie]

Modern ones have low latency modes, the Kiis is infinity variable, and you can set up as a preset, the 8Cs are fine in their full on latency but they have a low option too.
Keith

Indeed Keith but still too much for a centre channel with all the other channels old fashioned I fear.
I was thinking of maybe a Neumann KH420 as an excellent conventional active speaker. I want a 3 way or I may as well stick with the M33

 

[maverickronin]

Keep in mind that modern TVs average around 9-10ms of latency for video processing. So there's no problem integrating minimum phase DSP (typically 3-5ms) into modern video rigs -- the DSP is faster.

And how would that compare with modern PC gaming rigs, with 120Hz+ monitors?

10ms is gaming monitor territory. If you grab a random monitor or TV it's likely to be much worse. A random monitor is usually better than a random TV but 50-100ms isn't uncommon.

 

[hege]

Doesn't it correct phase too? That'd put it at least one level above basic for me. I do also remember something about working with multiple subs, which only Dirac with an extension costing a pretty penny has too; though it doesn't correct the final summed response, I think?

No, it doesn't do any phase (time-domain) correction. It in no way compares to Dirac in that regard.

What it can do is adjust subwoofers delay (time-aligning), but don't know if it does that with 100% accuracy. I do know that it doesn't even set the subwoofer levels correctly, if you have multiple subs you have to manually adjust the levels like -3dB or -6dB. It's really as basic as an auto EQ can get.

 

[lc155]

No, it doesn't do any phase (time-domain) correction. It in no way compares to Dirac in that regard.

What it can do is adjust subwoofers delay (time-aligning), but don't know if it does that with 100% accuracy. I do know that it doesn't even set the subwoofer levels correctly, if you have multiple subs you have to manually adjust the levels like -3dB or -6dB. It's really as basic as an auto EQ can get.

From what I understand, GLM 4 will do phase correction, no?

Still, for the asking price of the system, it really should do more than this.

 

[hege]

From what I understand, GLM 4 will do phase correction, no?

Where do you read this?

If you see any reference to "phase correction", it would refer to aligning subwoofer to mains, not any fancy time domain speaker phase linearization.

Still, for the asking price of the system, it really should do more than this.

Dunno. I guess any fancy stuff would depend on if the hardware is even capable of complex FIR-stuff. And they simply might have decided long ago that any phase related stuff is not audible. Though I think Ilkka advertised S360 as being the most linear phase speaker they have, so I guess it means something.

 

[Duke]

Member @jtwrace recently interviewed Earl Geddes. By way of introduction to Earl's work, here is the data on his top-of-the-line Summa: GedLee LLC

Earl designed both passive and active versions of the Summa. The link below is cued up to the part where Jason asks him to compare the two approaches. This part lasts from 32:39 to about 34:55:

Waveguides, Acoustics and Speaker Design with Dr. Earl Geddes - YouTube

Short version: Active is a little better, and the gap narrows as the passive design is optimized.

 

[dfuller]

I don't particularly care for DSP designs as I'd rather not have the latency, but I'm firmly a fan of biamping with active crossovers. Plus, less gear to worry about.

 

[Purité Audio]

Contemporary active designs D&D8C/ Kii allow you to adjust latency, they can be used latency free, with the kiis for example the only disadvantage of choosing no latency is the loss of phase coherence.
Keith

 

[stevenswall]

Member @jtwrace
Short version: Active is a little better, and the gap narrows as the passive design is optimized.

Seems like the problem then is manufacturers purposefully choosing to not compete with active designs. The ones that brag about their crossofters seem to in several cases, talk about how simple it is or how few components it has, because they don't know about or don't care about interference patterns and lobeing, or or distortion.

 

[Duke]

Seems like the problem then is manufacturers purposefully choosing to not compete with active designs.

Or seen from another angle, maybe the problem is passive crossover designers failing to adequately optimize their designs. Earl mentions Tom Danley as a designer whose thoroughly-optimized (and correspondingly complex) passive crossover designs achieve excellent results.

The ones that brag about their crossofters seem to in several cases, talk about how simple it is or how few components it has...

Funny you should mention that. I've had people at audio shows literally say, "Oh, I can tell by the sound you're using a very simple first-order crossover." Nope! But I did pattern-match in the crossover region, and that matters.

Imo parts count is pretty much of academic interest only (as long as parts quality is not sacrificed). What matters are the acoustic end results and the impedance curve the amplifier sees.