闲聊桌面音箱的测试 - 鬼斧神工119的文章 - 知乎
https://zhuanlan.zhihu.com/p/611898115
From the designer. Translation by a friend.
I’ve written about the CEA2034A standard for loudspeaker testing a while back. There are some limits as to where this standard can be used. Due to the variety of speakers and their intended usage, sometimes different test methods are more applicable.
Smart speakers, Bluetooth speakers and desktop speakers have more appropriate test methods, some of which comply with or originate from the IEEE 1329 standard.
Put simply, the speaker is placed on a desk and tested at a predetermined height/angle. Of course, this is just one configuration. For some products, other factors, such as listener posture, will influence measurement configurations as well as design choices. For example, the difference in tweeter section between the first-gen HomePod and second-gen HomePod2, which I’ve analysed before.
In a similar manner, some speakers made for desktop use (In the general sense, including some Bluetooth speakers and smart speakers) have their own target responses. There is some in-depth research in the industry, but as it’s not been published yet, I cannot discuss specifics here. What can be said, however, is that there are differing evaluation standards, test weights, and target responses for speakers versus CEA2034A and other more “traditional” standards, meant for the evaluation of domestic speakers placed in a living room setting and at a distance from the listener.
Of course, many passive loudspeakers and active loudspeakers without a DSP preset for desktop scenarios cannot fully solve the problem of response variations from desktop reflections, but at least some of them are developed with reference to the aforementioned standard for desktop speakers, rather than CEA2034A, in their R&D process, and I did mention a few of these speakers in previous articles. This also applies to some soundbars with a desktop mode.
Besides, certain desktop speakers (again, in the general sense) have a different emphasis on aspects of their performance. Some loudspeakers are made for usage scenarios that place more emphasis on On-Axis performance (and Early Reflection and related Directivity, or course) of the speaker. One such scenario would be when the speaker is placed in a living room with their acoustic axis properly aligned towards the main listening position. Some will cite that horn loudspeakers could have On-Axis response variations due to diffraction around the edges of the horn, these variations can be considered insignificant If they only occur across a limited area. (Translator Note: probably meaning over limited bandwidth and angular intervals)
However, some usage scenarios dictate a shorter listening distance than usual, such is the case when listening is conducted in a desktop setting with the user sitting near the speaker. The distance between the listener’s head and the speaker itself is short enough that the listener’s head and ears can no longer be considered as a single point in the sound field. Suppose listening is conducted at 50~60cm away from the speakers, with the speaker and the head in an equilateral triangle. Even if the acoustic axis of the speaker points toward the middle of the listener’s head, the angular deviation between the two ears and the acoustic axis of the speaker would be 10~12 degrees. If the speaker points toward one of the two ears, the angular deviation for the other ear will be even greater. Therefore, an average of the speaker response across this angular interval is more suitable for some products designed with this scenario in mind.
Vertically, the difference in each user’s height and various tables’ and chairs’ heights also need to be taken into account. It should also be averaged over an interval, barring the use of an adjustable stand. (Translator Note: He’s basically talking about a Listening Window, but probably with a bit more vertical interval)
On top of all these, some speakers built for short-distance listening need to have their reflections assessed differently.
There are different test standards and methods of assessment for loudspeakers for different usage scenarios, such as professional audio and car audio. In the past, I followed the corresponding standard when evaluating these speakers designed for specific, alternative usage scenarios. I will talk about more of these “alternative” standards in the future