Relationship between C50 Clarity, Directivity, and Soundstage

[Fredygump]

I would like to discuss the relationship between clarity, speaker directivity, and the subjective soundstage or immersiveness of a stereo speaker system.

I have been operating under the assumption that a speaker with narrow/ controlled directivity will have more direct sound than a more omnidirectional speaker. But this theory seems largely untested since speakers with narrow or constant directivity seem rare. I think it is clear that modern speaker designs trend to use dome tweeters and small woofers.

I have prototype speakers with narrow directivity (60x40 degree pattern). The subjective detail and clarity is quite good, and the clarity C50 measurements are ~18dB across the full range.

Anecdotally, when I search for people discussing clarity, the number that keeps coming up is ~10dB. I assume this value corresponds to widely available wide directivity speakers.

Also anecdotally, I have seen discussions here and elsewhere about how reflections can be a good thing since they create a broader sound stage and sense of immersion. Listeners seem to like these reflections, and audio professionals try their hardest to eliminate them. And then forums will make people think the holy grail of audio perfection is a mastering studio level acoustic treatment that dramatically reduces reflections, even though the person in question may really be searching for the illusion of an immersive sound stage. Making a room more dead when we want live sound is obviously counter productive.

I feel that there is potential to correlate measurements like C50 or C80 to subjective listening experience? Then a person could measure their system and understand how their system fits on the scale, and therefore they can make a more informed decision about crafting the perfect system for themselves. This could inform speaker choice, acoustic treatment decisions, etc.

I think it would be great if we could come up with generalized ranges of measurements that correlate to different subjective experiences. We could use the scale of wet -> dry like a mix engineer would use, where "wet" is massive reverb from the room that reduces clarity, and "dry" is approaching an anaechoic chamber? It shouldn't be too hard to correlate measurement values?

 

[kemmler3D]

Listeners seem to like these reflections, and audio professionals try their hardest to eliminate them.

As with all things audio, it depends.

The concept of either a completely dead studio (in the sense of being actually anechoic) was never popular, I think RFZ (reflection-free zone) is still pursued by some but is also considered by others to be out of date.

Making a room more dead when we want live sound is obviously counter productive.

Unfortunately I think this also depends on the recording.

A stereo-mike recording of a live concert with all the hall reverb captured as-is on the recording might sound very "live" on a good system, in a super-dead room.

An artificially mixed studio recording with realistic panning but relatively little reverb, played on good omni speakers in a moderately reflective room might sound "live" as well.

If you switch them around, the outcomes will be quite different.

And then there are many recordings that have no relationship to a real-world physical layout of musicians, and getting a "live" sound is a nonsense proposition to begin with. For example, all the piano recordings where the keyboard spans the width of the stereo image (low notes on left, high notes on right).

I feel that there is potential to correlate measurements like C50 or C80 to subjective listening experience?

Sure, I think so.

I think it would be great if we could come up with generalized ranges of measurements that correlate to different subjective experiences.

Per the above in my comment, I think it would end up being a loose correlation / rule of thumb but could be useful for rough diagnosis. On the other hand, if someone can provide C50 / C80 in a forum thread, in all likelihood they're running REW and can just provide the waterfall, ETC, and more, so I am not sure it would advance the state of the art of recommendations on forums.

I think it would be great if we could come up with generalized ranges of measurements that correlate to different subjective experiences. We could use the scale of wet -> dry like a mix engineer would use, where "wet" is massive reverb from the room that reduces clarity, and "dry" is approaching an anaechoic chamber? It shouldn't be too hard to correlate measurement values?

I'd bet you money someone has already done this and published something with AES or similar. C50 is a pretty old metric and it's been in use in professional acoustic work for a good while. Might be worth a look...

 

[DVDdoug]

There's a long thread Properties of speakers that creates a large and precise soundstage .

I didn't read it all and I don't know what the consensus is...

My intuition tells me that a large soundstage is not compatible with a precise soundstage, but sometimes intuition is wrong.

Floyd Toole says that EARLY reflections help with intelligibility.

I've also read that listening an anechoic chamber can give the impression of the sound coming from inside the head (like a lot of people get with headphones) so apparently you need some reflections for a realistic soundstage illusion.

even though the person in question may really be searching for the illusion of an immersive sound stage. Making a room more dead when we want live sound is obviously counter productive.

When listening to stereo recordings I like to use a "hall" or "theater" setting on my AVR for the "feel" of a larger room. More "immersive" as you say. But I DONT like the sound of an over-reverberant small room. I have a fairly typical living room and the short "small room" reflections aren't bad enough that I notice them.

And then forums will make people think the holy grail of audio perfection is a mastering studio level acoustic treatment that dramatically reduces reflections

The BIG problems are usually in the bass range where you get standing waves. That creates big bumps and dips/cancelations at different frequencies and in different places in the room. That can be improved with bass traps to trap the reflected waves, smoothing-out the bumps and the dips. "Casual" room treatment doesn't help with bass.

And if you are going to treat your room (which I haven't) I always recommend, "diagnosis before treatment".

Oh... I ran across something written previously by Amir:

A dipole speaker surely generates spatial effects that are not real and were never heard in the studio. Combine two such speakers nave you are just dealing with fantasies as opposed to mimicking what someone may have setup.

That seems to align with my intuition of "large" and "precise" being opposing goals.

 

[Fredygump]

That seems to align with my intuition of "large" and "precise" being opposing goals.

On accident I set one of my speakers with a delay, like 20ms. I played some music, and it sounded amazing, as in big and immersive...despite being played on directional speakers. But then I switched to something with dialogue, and I instantly knew something was terribly wrong!

So yeah, your Amir quote seems on point!

I agree about bass issues, atleast if it is a mono-subwoofer issue! My speaker project was primarily focused on improving low frequency response, and I'm tempted to say that full range stereo pair, without multiple subwoofer tricks, might be good enough. This would be especially true if we believe what Toole said about how we intuitively compensate for how sound behaves in a room.


Back to Clarity, I am able to find that a C80 of a large music hall is typically 1 to -4dB. And I found another reference saying that C50 of >3 is good. These numbers are for "big" rooms. I know our small rooms are different, so we can't automatically assume those numbers mean anything to us.

There is not much information to find when simpy searching. There are some papers on researchgate, but usually I can't access the information on that site...

 

[Arindal]

I feel that there is potential to correlate measurements like C50 or C80 to subjective listening experience?

Clarity indeces like C50 and C80 were introduced to evaluate aspects of speech and complex music quality in reverberant rooms, originating from real sound sources not containing meaningful reverb pattern themselves (speakers, instruments, vocalists or sound reinforcement speakers with mostly dry signals alike).

The moment you evaluate the sound of stereo systems where localizable real sound sources and phantom localizations vary, the calculations seemingly do not apply anymore. Additionally, recordings of acoustic events such as live concerts in a concert hall, contain an unknown degree of meaningful reverb pattern having its own influence on clarity and transparency. The way reverb on the recording and additional reverb created in the listening room blend, is seemingly depending on completely different factors compared to the situation on which C50 and C80 calculations are based. In most of cases, additional reverb in the listening is coming earlier, and does not offer a precise tonal match.

how reflections can be a good thing since they create a broader sound stage and sense of immersion.

Would say it very much depends on the particular case and whether (discrete) reflections deteriorate the localization stability and to which extend they blend seamlessly with the ambience on the recording or not.

A certain degree of diffuse reverb in the listening room is desired in order to compensate for the reflections missing from a significant percentage of angles towards the listener. The question is how do we perceive the perceive the angles, and at which point such reverb can dominate the reverb on the recording. In this regard, tonality and initial time delay of the reverb become important factors.

My personal recommendation would be to make sure that indirect sound in a listening room is sufficiently linear in terms of tonal balance compared to the direct sound. Constant directivity speakers (at least over the localizable frequency bands) and even room treatment (not overdamped) are contributing to that. The second aspect is sufficient diffusion in order not to mess up the localization and allow reverb on the recording and reverb in the listening room to blend seamlessly.

Two simple experiments are usually helping to test that: One is to check the localization stability of monaural content like dry spoken word and how it changes depending on increasing the listening distance in steps of typically 20cm. If there is a sudden decline in stability, you can be sure of some discrete reflections getting dominant, usually very bandwidth-limited. Particularly with untreated rooms and speakers offering uneven directivity (for example tweeters with broad dispersion pattern), this can happen pretty quickly.

The second test is to listen to some stereo recording with complex timbral mixture, like orchestral pieces, and a balanced reverb, ideally originating from a concert call that is familiar to the listener. The less the tonal balance of the reverb and the subjective proximity/depth-of-field are changing over a vast range of different listening distances, the better the subtlety of the additional reverb and the more likely the two patterns will blend.

If the quality of the reverb in the listening room is excellent, you will most probably reach a tipping point at overly large listening distances at which the proximity disappears completely making way for a ´fully distant projection plane´ with the complete recording drowning in reverb which is not tonally colored. Speakers offering a colored off-axis response in the sense of increasing directivity index (as well as in treble-overdamped rooms) the reverb will appear to be increasingly dull and midrange-laden, subjectively separating phantom sources and assigned reverb, so you will not get to this tipping point of loosing proximity.

Does it make sense?

the person in question may really be searching for the illusion of an immersive sound stage

My advice would be to listen to surround or immersive recordings with a channel-discrete setup, if one wants immersive soundstage. It cannot be achieved by taking a stereo recording and pimping it with dominant additional reflections in the listening room.

A stereo-mike recording of a live concert with all the hall reverb captured as-is on the recording might sound very "live" on a good system, in a super-dead room.

I tested that in an anechoic chamber, and it works indeed, if the level of the reverb on the recording is sufficient and the pattern gives meaningful information (not just decorrelated artificial reverb).

 

[Curvature]

I would like to discuss the relationship between clarity, speaker directivity, and the subjective soundstage or immersiveness of a stereo speaker system.

I have been operating under the assumption that a speaker with narrow/ controlled directivity will have more direct sound than a more omnidirectional speaker. But this theory seems largely untested since speakers with narrow or constant directivity seem rare. I think it is clear that modern speaker designs trend to use dome tweeters and small woofers.

I have prototype speakers with narrow directivity (60x40 degree pattern). The subjective detail and clarity is quite good, and the clarity C50 measurements are ~18dB across the full range.

Anecdotally, when I search for people discussing clarity, the number that keeps coming up is ~10dB. I assume this value corresponds to widely available wide directivity speakers.

Also anecdotally, I have seen discussions here and elsewhere about how reflections can be a good thing since they create a broader sound stage and sense of immersion. Listeners seem to like these reflections, and audio professionals try their hardest to eliminate them. And then forums will make people think the holy grail of audio perfection is a mastering studio level acoustic treatment that dramatically reduces reflections, even though the person in question may really be searching for the illusion of an immersive sound stage. Making a room more dead when we want live sound is obviously counter productive.

I feel that there is potential to correlate measurements like C50 or C80 to subjective listening experience? Then a person could measure their system and understand how their system fits on the scale, and therefore they can make a more informed decision about crafting the perfect system for themselves. This could inform speaker choice, acoustic treatment decisions, etc.

I think it would be great if we could come up with generalized ranges of measurements that correlate to different subjective experiences. We could use the scale of wet -> dry like a mix engineer would use, where "wet" is massive reverb from the room that reduces clarity, and "dry" is approaching an anaechoic chamber? It shouldn't be too hard to correlate measurement values?

C50 was developed in architectural acoustics to measure speech clarity in classrooms and the like. An equivalent measurement for concert halls is C80.

These are crude metrics and heavily depend on the size, construction and furnishing of the room far more than they do on speakers.

More useful metrics are found in the room texture analyzer by Alejandro Bidondo: https://alejandrobidondo.wordpress.com/ These are new and potentially promising ways to analyze how rooms respond and shape reflected energy. Again they are not developed to analyze speaker/room interactions, but their precision gives them an edge worth trying.

 

[Heinrich]

I have prototype speakers with narrow directivity (60x40 degree pattern).

Is that achieved throughout the whole frequency band, or is it just a sharp pro/ horn used with a single bigger(?) woofer/mid from the typical 1kHz on?

 

[youngho]

I feel that there is potential to correlate measurements like C50 or C80 to subjective listening experience? Then a person could measure their system and understand how their system fits on the scale, and therefore they can make a more informed decision about crafting the perfect system for themselves. This could inform speaker choice, acoustic treatment decisions, etc.

I think it would be great if we could come up with generalized ranges of measurements that correlate to different subjective experiences. We could use the scale of wet -> dry like a mix engineer would use, where "wet" is massive reverb from the room that reduces clarity, and "dry" is approaching an anaechoic chamber? It shouldn't be too hard to correlate measurement values?

I believe that C50/C80 will be inversely related to T30/60. You might be interested in reading the following:

Laukkanen 2014
https://users.aalto.fi/~ktlokki/Publs/mst_laukkanen.pdf
“The results of the preference tests clearly showed that mixing engineers prefer quite dry rooms (T60 of 0.15 - 0.20 s) and interviews confirm that the stereo image and the amount of room reverberation are the most important factors for them. In contrast, mastering engineers seemed to prefer more lively rooms (T60 of 0.30 - 0.40 s) and the frequency balance was the most important factor for them. It was also noticed that the preference rating varied between different music samples, especially among mixing engineers.”
“The preference rating seemed also to vary between different music samples…on the basis of the listening tests, it seemed that different music genres need different kind of treatment in control room.”

Bech and Lokki 2019
https://users.aalto.fi/~ktlokki/Publs/JASMAN_vol_146_iss_5_3562_1.pdf
Sound field reproduction using spherical loudspeaker array in anechoic chamber, listeners told ““Imagine that you are in a typical residential room, listening to a 2-ch stereophonic reproduction over loudspeakers.”
Four perceptual constructs comprising attribute clusters

1. Reverberance: relates to the later energy [of the sound field], “excellent relation” to RT30 and early decay time
2. Width and envelopment: relate to the earlier energy of the sound field
3. Bass
4. Proximity, negatively correlates to width and envelopment, “strong correlation” with clarity index 50 (C50) and direct to reverb ratio (DRR)

“Assessors systematically preferred the sound fields with lower RT. In our study, the most preferred acoustical conditions presented fields that evoked the sense of being less reverberant and less wide and enveloping. The sources were perceived as closer to the listener, exhibiting high levels of proximity. It is also important to note that the current results suggested that a negative preference is apparent for acoustical conditions with RT higher than 0.4 s”
“One could attempt to alter the DRR within a field by means of directivity control in the loudspeakers, aiming to evoke certain perceptual aspects that would otherwise be dominated by the room’s natural acoustical field.”

Kantamaa 2020
https://aaltodoc.aalto.fi/bitstream...Kantamaa_Olli_2020.pdf?sequence=2&isAllowed=y
“Increase of the low–middle frequency [80-600 HZ] directivity improved clarity, reduced sound colouration, improved virtual sound image definition and transient reproduction compared to the conventional loudspeaker directivity.”

Lokki 2022 (just focus on the mixing rooms)
https://acris.aalto.fi/ws/portalfiles/portal/56976875/Riionheimo_Lokki_Movie_Sound_Part2.pdf
“The perceived sense of space matches better with the measured C50 (scale inverted) in the middle frequencies than the measured reverberation time T30.”
“Perceived and measured clarity match well.”
“The width was difficult attribute to evaluate, indicating the word itself is ambiguous, especially with surround sound. The width of the soundscape is affected by the angle of the left and right screen speakers, the volume of the surround speakers, and the envelopment caused by the room reverberation.”
“The ratings for brightness match well with the level of the high frequencies above 4 kHz of the electroacoustical responses.”
“The perceptual distance matches better with measured clarity C50 in the middle frequencies than the actual listening distance.”
“The assessors preferred somewhat clear and dry sound over reverberant and distant; however the room should not be totally dead nor too bright.”

Lokki 2024 (just focus on stereo testing, also observe variance of preference with program material and work vs leisure listening)
https://research.aalto.fi/en/public...-exploring-the-impact-of-listening-room-on-so

Young-Ho

 

[DanielT]

To get a "big" sound, with reverb added, it can be like this:

Speaking of music that is mixed, recorded in a way that makes basically any Hifi system sound good. Not for nothing that they are used at Hifi fairs:

The DIY knew what he was doing when he recorded this particular section of the song. It's right there that sparkling sweet spot with the guitar section that is perceived as "high resolution", "clarity", and so on. The song that is in the Spotify link above:

With that said,these large DIY AMT based speakers together with bass boxes sounded damn good no matter what was played,

 

[DanielT]

You have to experiment and see what you like. However, too much or too little reverb is probably not a very pleasant experience for most people, for example this: