Ascilab C8C Active Speaker Review

[Mort]

so 8C/KH420 guys pony up

That's a $25k ask. Of course they will be returned, but it's a tall order.

 

[Curvature]

For subs, performance is distortion-dominated. You buy the biggest, lowest-distortion sub with the highest output at the frequencies of interest, up to whatever your budget and aesthetic considerations dictate. For speakers, the radiation pattern always come first and everything else is contextual information for the purchase decision. There are no means for conducting high-quality, controlled blind tests meant to detect and grade small differences in nonlinearity by consumers. We only have measurements to fall back on, and we can't fully interpret the measurements.

The speakers mentioned are all excellent: KH420, 8c, C8C. They present varying trades in performance based on their design. There is no clear, definitive, optimal choice.

Let's look at preference scores. Remember that preference scores are useful because they follow a consistent method of calculation. They don't mean nothing and they help interpret the data when the underlying measurements are done well.

The max difference in preference scores under any circumstance is 0.6. The attempts to correlate the score with listener preference suggest that a perceptually relevant threshold is reached with differences of 0.7 or greater. As such, even if differences are heard, the results here suggest that no clear preference would be found under blind circumstances with these speakers.

If we wanted to extend the reasoning beyond that:

• Cardioid speakers perform a little better in-room for SBIR. The Neumann speaker is therefore eliminated.
• The C8C has lower distortion than the 8c. So the 8c is eliminated.
• Confounding factors are that the Neumann speaker has potentially better reliability and may play louder overall, the 8c has arguably better control of the radiation pattern in the midrange, but the C8C extends lower. And small differences in each for the listening window and vertical radiation.
• Then there are aesthetics, availability, price.

If I had the money I would buy the C8C because it looks cool and performs excellently. That's not a definitive judgment, but it's a decent one.

 

[illusonic]

If anyone says "X speaker sounds awesome!" that is a valid data point that it sounds awesome to one person. (But at least one person likes it!)

What if everyone says speaker X sounds awsome but measurements suggest too high distortion? It seems that's the situation that may come closer to reality for D&D 8C than your single subjective result.

 

[amirm]

What if everyone says speaker X sounds awsome but measurements suggest too high distortion?

Everyone can detect linear (frequency response) difference but not everyone can hear distortions. Nor would they play the content that would make such distortion audible. So if all else is equal, I would take lower distortion irrespective of what everyone else says.

 

[illusonic]

Everyone can detect linear (frequency response) difference but not everyone can hear distortions. Nor would they play the content that would make such distortion audible. So if all else is equal, I would take lower distortion irrespective of what everyone else says.

From my experience the ability to detect linear as well as nonlinear distortion varies over individuals. Both can be trained but a majority of untrained people will only be able to clearly detect substantial deviations from linear frequency response (e.g. 3dB bass/treble shelve in comparison to a reference) and high levels of distortion (e.g. 10% IMD in fullrange or 2-way speakers, 100% THD can be utilized for lower octave bass enhancement algorithms with positive listener preference).
In an "all else is equal" case, I would also tend to lower distortion but I would listen first. Unfortunately, there are usually drawbacks like bigger size or higher cost that come with lower distortion. And in case of limited bass output, harmonic distortion sounds like more bass. Since I got a demo of a distortion compensation system on an 8" bass, where direct A/B comparison of about 12dB difference in distortion was possible, I don't care much about bass distortion. And I know that many would prefer to keep distortion for bass enhancement in that specific case.

 

[illusonic]

For speakers, the radiation pattern always come first and everything else is contextual information for the purchase decision. [...]

Let's look at preference scores. Remember that preference scores are useful because they follow a consistent method of calculation. They don't mean nothing and they help interpret the data when the underlying measurements are done well.

The max difference in preference scores under any circumstance is 0.6. The attempts to correlate the score with listener preference suggest that a perceptually relevant threshold is reached with differences of 0.7 or greater. As such, even if differences are heard, the results here suggest that no clear preference would be found under blind circumstances with these speakers.

Palmer Orbit 11 preference scores are 6.26, 7.45 (with EQ), 7.16 (with Sub) and 8.35 (with both).
If your reasoning above holds true and considering speaker cost, any sane person would apply EQ and potentially a sub to the Orbits and be done.
Yet, this is probably not what you meant to say. So why would anyone want to invest multiple times the Orbit price to buy one of the speakers above?

 

[amirm]

From my experience the ability to detect linear as well as nonlinear distortion varies over individuals. Both can be trained but a majority of untrained people will only be able to clearly detect substantial deviations from linear frequency response (e.g. 3dB bass/treble shelve in comparison to a reference) and high levels of distortion (e.g. 10% IMD in fullrange or 2-way speakers, 100% THD can be utilized for lower octave bass enhancement algorithms with positive listener preference).

I don't know what formal experience you are drawing from. I can play two different speakers and get anyone off the street in a blind test to correctly identify them. Yet, hardly any will be able to hear artifacts in MP3 encoding that are clearly audible to me. At my last job, we tested large group of audiophiles on detecting lossy codec impairments. The failed miserably compared to our trained listeners.

Importantly, not only will you need to be trained to hear small non-linear impairments, but also need to have the right content which exposes the weakness in the system. This is far harder than having them hear frequency response variations.

Unfortunately, there are usually drawbacks like bigger size or higher cost that come with lower distortion.

Not really. In the same price class, there is massive variation in distortion between speakers.

 

[Curvature]

Palmer Orbit 11 preference scores are 6.26, 7.45 (with EQ), 7.16 (with Sub) and 8.35 (with both).
If your reasoning above holds true and considering speaker cost, any sane person would apply EQ and potentially a sub to the Orbits and be done.
Yet, this is probably not what you meant to say. So why would anyone want to invest multiple times the Orbit price to buy one of the speakers above?

All things being equal, that's exactly how to use the score. More context is needed to make a purchase decision in any case.

The kind of EQ informing the score change has to be applied on an anechoic basis per speaker, the specific speakers you bought, not the speakers measured in reviews. The techniques to achieve this at home on a limited basis require some knowledge about not only measurement, but speaker design and filter creation and the kinds of corrections that are feasible and practical (don't try to simply invert the measured response, for example).

@Nuyes has showed that self-noise in the Orbit 11 speakers is mostly luck of the draw, which points to inconsistent manufacturing at least of the amp electronics. They also have limited low bass output, requiring a sub below 100Hz, and the cardioid pattern extends only to about 300Hz or so, which doesn't address the most problematic SBIR region and makes them similar to some large baffle bookshelves and some floorstanders in terms of pattern control.

 

[illusonic]

I don't know what formal experience you are drawing from. I can play two different speakers and get anyone off the street in a blind test to correctly identify them. Yet, hardly any will be able to hear artifacts in MP3 encoding that are clearly audible to me. At my last job, we tested large group of audiophiles on detecting lossy codec impairments. The failed miserably compared to our trained listeners.

Importantly, not only will you need to be trained to hear small non-linear impairments, but also need to have the right content which exposes the weakness in the system. This is far harder than having them hear frequency response variations.

Related experience on my side is occasionally demoing acoustic systems to usually untrained listeners in A/B comparison. No formal listening tests. Participants are interested in clear differences that can influence a consumer buying decision, not just noticeable differences like in formal studies. If I demo something, I do hear differences in tonality and distortion that relate to measured performance - many people don't. A convincing demo for untrained listeners has to include striking differences like 6dB bass level or extreme distortion. The latter usually includes strong IMD as THD alone mostly goes unnoticed.

Based on this experience, I can't confirm your statement that everyone hears linear distortion but only some hear nonliner distortion. In both cases audibility depends on the level of distortion as both can reach levels that are impossible to ignore. However, deviations between speakers at low playback levels are usually more noticeable in frequency response than in distortion. If put this way, I would agree. Turn up the volume and distortion can also become obvious.

Compression artefacts are a typical example for the blessings of ignorance. It's good if one can enjoy for example youtube content or that special song from a streaming service that is only available in MP3. Besides that, most people would not want to pay for differences that are inaudible to themself even if a specially trained person may be able to detect them.

Not really. In the same price class, there is massive variation in distortion between speakers.

My list of potential drawbacks included also size. An important factor in many cases.

 

[traind]

I don't know what formal experience you are drawing from. I can play two different speakers and get anyone off the street in a blind test to correctly identify them. Yet, hardly any will be able to hear artifacts in MP3 encoding that are clearly audible to me. At my last job, we tested large group of audiophiles on detecting lossy codec impairments. The failed miserably compared to our trained listeners.

Importantly, not only will you need to be trained to hear small non-linear impairments, but also need to have the right content which exposes the weakness in the system. This is far harder than having them hear frequency response variations.


Not really. In the same price class, there is massive variation in distortion between speakers.

For the average person who loves listening to music, is it an advantage or a disadvantage ( disadvantage meaning, it is a bit of a curse) to be a trained listener? Or stated another way, is ignorance bliss in this regard?

 

[Curvature]

For the average person who loves listening to music, is it an advantage or a disadvantage ( disadvantage meaning, it is a bit of a curse) to be a trained listener? Or stated another way, is ignorance bliss in this regard?

Once you learn to hear dynamic compression, a ubiquitous studio procedure (different from the audio coding discussed above), you can't unhear it. Then you are forced to ask yourself about what you like and not in application, and there's a whole dimension of music to explore as a result. An unavoidable fact of gaining knowledge. Not so different from learning to listen to the sound of speakers rather than only listening through them to the music alone.

 

[illusonic]

The kind of EQ informing the score change has to be applied on an anechoic basis per speaker, the specific speakers you bought, not the speakers measured in reviews. The techniques to achieve this at home on a limited basis require some knowledge about not only measurement, but speaker design and filter creation and the kinds of corrections that are feasible and practical (don't try to simply invert the measured response, for example).

I was under the impression that one can apply exactly the EQ proposed at spinorama.org as it was determined based on anechoic data. Now you introduce additional barriers asking for anechoic measurements for the exact speaker that receives the EQ. In this case, the preference score is worthless as it only applies to the measured sample.

@Nuyes has showed that self-noise in the Orbit 11 speakers is mostly luck of the draw, which points to inconsistent manufacturing at least of the amp electronics.

Dozens of users gave feedback on self-noise. Nobody is bothered past 1m. I don't spend up to 8 time the money to be able to stick my ear to the tweeter without hearing noise.

They also have limited low bass output, requiring a sub below 100Hz

If the Orbits need a sub below 100Hz, D&D 8C and Kii 3 also need one.

the cardioid pattern extends only to about 300Hz or so, which doesn't address the most problematic SBIR region and makes them similar to some large baffle bookshelves and some floorstanders in terms of pattern control.

KH420 has no cardoid extension at all. And the most problematic issue of front-wall reflection is adressed by the Orbits as you can see here.
The "similar to large baffle bookshelves and some floorstanders" claim begs for examples from your side.

 

[Curvature]

I was under the impression that one can apply exactly the EQ proposed at spinorama.org as it was determined based on anechoic data. Now you introduce additional barriers asking for anechoic measurements for the exact speaker that receives the EQ. In this case, the preference score is worthless as it only applies to the measured sample.

Manufacturing tolerance of a few dB is normal, although very good manufacturers of course seek to minimize the variance. Neumann and Genelec are particularly consistent. Score optimization via EQ on spinorama.org is an example of what can be done, not an instruction manual on what works best in a given setup. At home, I would use anechoic data in reviews as a rough guide about what to expect and use my own measurements to make final decisions.

Dozens of users gave feedback on self-noise. Nobody is bothered past 1m. I don't spend up to 8 time the money to be able to stick my ear to the tweeter without hearing noise.

I've spent a long time with active monitors and their hiss and I'm sensitive to it. Your call on what's valuable to you and how you choose to interpret user experiences. 15dB SPL hiss at 10cm (equivalent to 0dB at 0.6m) is my limit, which is around what @Nuyes measured for most of the range in one sample, but I'm uncomfortable with the variability.

If the Orbits need a sub below 100Hz, D&D 8C and Kii 3 also need one.

I would agree on the Kii 3 (compression graph: link) but not the 8c (link). We don't have clear data for the Orbit 11s (the chart extends only to 100Hz: link), but the distortion levels would suggest compression is likely.

KH420 has no cardoid extension at all. And the most problematic issue of front-wall reflection is adressed by the Orbits as you can see here.
The "similar to large baffle bookshelves and some floorstanders" claim begs for examples from your side.

The most problematic range in rooms is 100-300Hz, where EQ tends to be only somewhat effective at best (even with FIR filters) and velocity-based porous absorbers have little impact.

Here are a few examples (-6dB contour):

• Revel F328Be: https://www.spinorama.org/speakers/Revel F328be/ASR/asr-v2-20210415/SPL Horizontal Contour.html
• JBL M2: https://www.spinorama.org/speakers/JBL M2/ErinsAudioCorner/eac/SPL Horizontal Contour.html
• MoFi SourcePoint V10: https://www.spinorama.org/speakers/MoFi SourcePoint V10 Master Edition/ErinsAudioCorner/eac/SPL Horizontal Contour.html
• Ascilab C8T: https://www.audiosciencereview.com/...e-about-to-launch.54794/page-139#post-2493493
• Genelec 8361A: https://www.spinorama.org/speakers/Genelec 8361A/ASR/asr-vertical/SPL Horizontal Contour.html
• Genelec S360A: https://www.spinorama.org/speakers/Genelec S360/ASR/asr/SPL Horizontal Contour.html
• Audio First Designs Cadentia 3: https://www.spinorama.org/speakers/Audio First Design Cadentia 3/Audio First Design/index_vendor.html

Biggest difference to cardioid designs is the widening of the radiation pattern as frequency decreases before control is lost. Again, your call on what you would consider relevant and which contextual factors matter to you. If you are price sensitive, need speakers now and not later, willing to use and integrate a sub, are ok with aesthetics and size, and understand the relative tradeoffs, then the Orbit 11s are a good choice.

 

[illusonic]

Manufacturing tolerance of a few dB is normal, although very good manufacturers of course seek to minimize the variance. Neumann and Genelec are particularly consistent. Score optimization via EQ on spinorama.org is an example of what can be done, not an instruction manual on what works best in a given setup. At home, I would use anechoic data in reviews as a rough guide about what to expect and use my own measurements to make final decisions.

Have you seen the Genelec 8351B measurements from stoneeh? Doesn't look like particularly well controlled tolerances when compared to ASR measurement. Doesn't look like a particularly flat frequency response either. Compare that to the Orbits on the right plot.

I've spent a long time with active monitors and their hiss and I'm sensitive to it. Your call on what's valuable to you and how you choose to interpret user experiences. 15dB SPL hiss at 10cm (equivalent to 0dB at 0.6m) is my limit, which is around what @Nuyes measured for most of the range in one sample, but I'm uncomfortable with the variability.

Don't know where you're located. In Europe this concern can be solved by the return period for online purchases.

I would agree on the Kii 3 (compression graph: link) but not the 8c (link). We don't have clear data for the Orbit 11s (the chart extends only to 100Hz: link), but the distortion levels would suggest compression is likely.

If the Orbits require a sub depends on SPL requirements of the user. Max SPL graph from stoneeh at 10% THD shown below. Add some 8-12dB room gain and 6dB for a second speaker and the plot explains why these speakers actually work quite well in typical living rooms even without a sub. The "Orbits need a sub" statement is no universal truth.

The most problematic range in rooms is 100-300Hz, where EQ tends to be only somewhat effective at best (even with FIR filters) and velocity-based porous absorbers have little impact.

Here are a few examples (-6dB contour):

• Revel F328Be: https://www.spinorama.org/speakers/Revel F328be/ASR/asr-v2-20210415/SPL Horizontal Contour.html
• JBL M2: https://www.spinorama.org/speakers/JBL M2/ErinsAudioCorner/eac/SPL Horizontal Contour.html
• MoFi SourcePoint V10: https://www.spinorama.org/speakers/MoFi SourcePoint V10 Master Edition/ErinsAudioCorner/eac/SPL Horizontal Contour.html
• Ascilab C8T: https://www.audiosciencereview.com/...e-about-to-launch.54794/page-139#post-2493493
• Genelec 8361A: https://www.spinorama.org/speakers/Genelec 8361A/ASR/asr-vertical/SPL Horizontal Contour.html
• Genelec S360A: https://www.spinorama.org/speakers/Genelec S360/ASR/asr/SPL Horizontal Contour.html
• Audio First Designs Cadentia 3: https://www.spinorama.org/speakers/Audio First Design Cadentia 3/Audio First Design/index_vendor.html

Biggest difference to cardioid designs is the widening of the radiation pattern as frequency decreases before control is lost. Again, your call on what you would consider relevant and which contextual factors matter to you. If you are price sensitive, need speakers now and not later, willing to use and integrate a sub, are ok with aesthetics and size, and understand the relative tradeoffs, then the Orbit 11s are a good choice.

For me the biggest advantage of cardioid speakers is the reduction of front wall reflections. This gives better focus in sound image and avoids notches in the frequency response due to cancellation by reflected sound.

Genelec 8361A comes close due to large size but still has considerably more sound leaking to the rear side below 2kHz than the Orbits. Due to the rear port, it can't be placed on the front wall to avoid notches caused by wall reflection. Placement remote from the wall moves the lowest lambda/4 distance into the omnidirectional frequency range. The resulting notch can't be compenated by EQ. All other speakers you listed leak even more sound to the rear.

 

[dfuller]

velocity-based porous absorbers have little impact.

Well, more at typical thicknesses. You need to push it up to 8-10" but that starts to take up considerable space.

 

[BELGIT]

Anybody received news on when to expect the ascilab C8C and BX8C together be shipped from Audiophonics?

Mine said mid June but no news afraid it will be delayed again...anyone else has received info from Audiophonics?

Cheers

 

[Curvature]

Well, more at typical thicknesses. You need to push it up to 8-10" but that starts to take up considerable space.

They still don't do much below 200Hz. Around there you have to switch to pressure-based.

 

[dfuller]

They still don't do much below 200Hz. Around there you have to switch to pressure-based.

Agreed though this depends on room size. And those take up even more space, so...

 

[Bruce Morgen]

Not really. In the same price class, there is massive variation in distortion between speakers.

One of many aspects of speaker performance to which "You get what you pay for!" frequently does not apply -- all the more reason to go by credible measurements rather than price or, for that matter, brand reputation among the "audiophile" contingent.

 

[Curvature]

Don't know where you're located. In Europe this concern can be solved by the return period for online purchases.

@stoneeh made measurements of the hiss. His sample had 40dB SPL at 10cm: link. Around 3min20sec. That's as high as PA speakers. This makes the speaker impossible to accept--for me. Based on some comments here over the years, some people are amazingly tolerant or amazingly oblivious to hiss. My first experience with an active speaker with silent hiss was the Neumann KH80, with hiss around 15dB SPL at 10 cm. There I actually have to put my ear to the tweeter.

I found compression measurements by @Nuyes: link. Starts around 4min.

Have you seen the Genelec 8351B measurements from stoneeh? Doesn't look like particularly well controlled tolerances when compared to ASR measurement. Doesn't look like a particularly flat frequency response either. Compare that to the Orbits on the right plot.

Out of interest, I took the time to plot them all. The level of variation is a little better for Genelec than it is for the Palmer speakers, though I wouldn't call either manufacturer's spec frequency response all that representative.

That said, @stoneeh's low frequency measurements look off. Groundplane is accurate, and gated measurements are certainly accurate over 1kHz, I wouldn't contest that, but I think it must have been cold outside, cold enough to cause the woofer surround compliance to change. Nothing to do with break-in, unlike what he says in the video.