Propeller Head Plaza

Posted by David Aiken (A) on November 20, 2002 at 13:35:08

In Reply to: A few more attempts at this :-) posted by Penguin on November 19, 2002 at 18:53:52:

I didn't say that the actual interactions would be the same, just that the room would interact in the same way, a much more general statement. The actual room interactions between the actual piano in the room, and playback of recording in the room WILL be different, even though the room is the same.

The reason for difference in the actual interactions is that the piano and the speaker will radiate their respective sounds differently and this will influence the reflection of sound around the room. Even if we used only one speaker and placed it at exactly the same place in the room as the piano, the reflected sound patterns would be different simply due to the difference in radiation patterns. Since you were talking in terms of a stereo recording, playback involves 2 speakers located in positions that will be different to that of the piano and that will set up some differences of its own as well. Just try moving a speaker around and see what happens to its sound in different locations, or play with the toe-in angle to see what simply changing the mix of direct/reflected sound plus some changes in the reflected sound due to moving of the early reflection points along the wall can do for a simple demo of some of these effects.

I don't know what the scale of the differences would be, but it will also be very speaker dependent since different speaker types (eg planar vs box) have wildly different radiation patterns and there are significant differences among speakers of the same type.

As I also said, the speaker will introduce some frequency and dynamic changes of it's own to the piano sound. You could hear them by playing the recording back in the anechoic chamber and comparing that to the sound of the piano in the chamber. The 2 sounds won't be quite the same to start with, even if the recording was perfect which it can never quite be.

I sympathise with what you're trying to achieve. We definitely don't have the recording and loudspeaker technology to do what you're suggesting and come up with a playback sound that won't be different to the sound of the actual piano in the room, so you will definitely hear differences. The end result, however, is going to be that those differences will only be true for one room and one set of speakers in a specific placement setup (and also for one overall system as well but I think the speakers will make the most significant contribution to the overall system sound). It could definitely be a useful test for the speaker designer as an indication for areas of further improvement in his design and you could possibly generalise on the results in relation to speakers of the same type with reasonably similar radiation patterns.

You'd ideally want to use the piano used for the anechoic recording when doing the room comparison and that will definitely limit the usefulness of the test since the piano isn't something you can slip in the box with the recording and market at a reasonable price. It could be a useful thing for an acoustic lab which rented out an anechoic chamber to have as part of the services they offer, if they could afford to have a piano around gathering dust the rest of the time when it wasn't wanted and major speaker companies like Revel and B&W could probably afford the piano and already have the anechoic chambers so they might find it useful.

As a side comment, I remember reading somewhere that one speaker company, Vandersteen I think, used recordings of familiar noises like a shovel scraping on cement for test material rather than music. Their reasoning was supposedly that the sounds didn't carry the emotional and other "baggage" that music often holds for us and that listeners made much better assessments of how "accurate" the speakers sounded than when music was used. I like using singers when I'm auditioning speakers since we all tend to be reasonably familiar with how a voice sounds and we can identify a lot of voice colourations very easily (nasality, chestiness, throatiness, etc) even on unfamiliar voices - example: you can tell whether someone has a cold just listening to the first sentence or two they say, even if you've never met them before. That makes them a very good standard when you can't do a direct comparison with the original. I also find that if you get voice sounding right, everything else seems to fall into place around it a lot better than it does when the speaker doesn't do voices well, so voice is my "possible/no way" test for weeding the list of speakers down to the short list for serious consideration.

David Aiken