 
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
|
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
In Reply to: Is there anybody equipped to perform the following test? posted by Penguin on November 17, 2002 at 16:18:28:
It's not going to work.By the way,w hy do this in an anechoic chamber?
JJ - Philalethist and Annoyer of Bullies
Follow Ups:
Then use a player piano :-) The point is to try to see what are the main differnces between a recorded instrument sound, and a real instrumnet sound. The primary recording environment should add nothing to the recorded instrumnent sound, The space cues should only come from the secondary recording space. It may be more realistic to record in mono.I can not prove it but i think the primary sound of an instrument has a lot to do with the interaction of the instrument with the sorounding space. Recording it in the anechoic chamber would remove the interaction. PLaying it back will not add back the interaction, while playing the instrument in the same space would interact with the space. using the secondary recording would enable the use of subtraction if one could guarantee that both passages were identical.
The question i seek an answer for is a simple one. Why can one tell if one listens to a recorded sound, voice or instrument, without looking twice? If audio memory is as bad as people want to make it out to be, how come 90 out of 100 will identify the recorded sound in one try?
BTW the second recording is just to make the whole thing easier to repeat, and to be able to give both samples to a large number of people. let them listen to it, the only thing they have to do is to answer the question about which recording is the recording of the real instrumnet in the listening space :)
dee
;-D
You want to find out what differences people hear between live and recorded sound.So what you are trying to do is to compare the information a mic placed in the normal listening position would pick up from:
(a) a live musician in the room, and
(b) a recording in which the image location of the musician is in the same place as the live musician in (a).
And the reason for the anechoic chamber recording is so that you have a "pure" instrument/singer sound to play back which contains no room information, just as the sound made by the instrument/singer in a real room contains no room information at the precise instant it is produced.
Yes, it would be a very interesting comparison but I'm not sure it's quite as valid as it appears. After all, we never hear that "pure" sound since we always musicians in a space with acoustic properties and that's the case in a recording studio as well so I'm not sure it's a good comparison to make.
Also, is it possible to judge the "virtual location" of a recorded musician in the room accurately enough to place the real musician or instrument there? I can judge their direction from the listening seat quite well, but I find the virtual depth tricky. It's almost as if the recording overlays another space over the room and I don't have the reference points required to judge depth in that space. Different speakers portray space and depth differently as well, though the angular location of where things are in the left to right dimension is usually more consistent across speakers.
I wonder whether the difference might not, in part, derive from the fact that there is always room acoustic information on our recordings and then our own room adds another set of such information and the two different sets of acoustic information "clash" in a way which is never heard when listening to live music. I would think there has to be something distinguishable when the recording, for example, contains information indicating that the performers are placed in a large concert hall and the listening room acoustic is telling you that you're in a 13' x 17' x 8' space with an alcove opening off it behind you and a large open doorway on the right side (my room). That kind of "location acoustic" confusion never occurs in real life. It doesn't have to be an audible artifact that is produced - simply having the 2 sets of information presented simultaneously should be enough to cause a level of confusion as to space that isn't present at a live performance and that may well be enought to tip us off. It's also interesting to consider that the scale of this problem will vary depending on the difference between the size and acoustic of the performance space and your own room, so the effect should be more pronounced on some recordings than others. I find symphonic recordings among the worst simply because the space difference is so large and I know there's no way an orchestra can fit into my room, so the scale is always way off for both the acoustic space and the people space occupied by the performers. It never seems as bad with a performance of one or two musicians, even if they're recorded in a live venue much larger than my room, because the scale of the performers seems more appropriate even though the acoustic space seems "off", especially when there's audience noise like applause.
Then there's always the issue of whether your speakers can reproduce the full frequency range and dynamic range of a live instrument. While there are some which can probably do this for a piano, given the bottom note on a piano is around 28 Hz most speakers will fail that and the tendency of speakers to compress dynamics at high volume means many fail there as well. Things will definitely get worse there too as you add more and more musicians.
No doubt someone like JJ can come up with a lot more differences between the live event and the reproduced, but I think we have a fairly good handle on what the differences are and I don't think that we can eliminate them all with better equipment and room correction. Frankly, I'm staggered at times at how good a job we can do at reproducing music. When you stop to think about it, it does seem rather implausible that you can produce the illusion of a musical performance in your room out of the bits and pieces that go into our systems, doesn't it?
David Aiken
The real purpose is to isolate the room effects. as in reverbration, delay etc, from the interaction of the same space with an instrument or voice. Take a piano for example; in the anechoic chamber you will only record the sound of the piano and when you play it back you will get the room effects and no interaction. When you play the instrument you will get both. I am wondering how much of the "realism" comes from the interaction between the source and the space. If there is no such thing then you would not be able to differentiate the recording of the recording in the acoustic space vs the recording of the real instrument in the same space. The secondary recording is the re just to make things a lot more comfortable, no need to get acousticly transparent curtains and a bunch of other things to make the test reliable, you can just give the recordings to a large number of people and ask them the simple task to identify which one of the two recordings sounds more real. Also if you use a high quality digital player piano, you could take several takes and subtract the anechoic recording from the acoustic recordings and see what the actual differences are. On avarage it should be the interaction of the space with the instrument, since both recordings should have the same room effects.
dee
;-D
When you play the recording from the anechoic chamber back, the room effects WILL interact with the sound of the recording in the same way that they will interact with the sound of a live instrument, at least on one level. The interactions won't be quite the same - the piano has a different radiation pattern to a speaker, and a stereo speaker pair will have a different radiation pattern again.If my guess at "confusion" due to the presence of 2 different sets of acoustic environment cues is right, you will reduce that situation to one set of cues and it will be the same environment providing them in both cases, but the radiation patterns of the speaker will still be different to those of the live instrument and there will still be the other issues I raised about frequency response and dynamic handling ability of the speaker vs the live instrument so the experiment may end up telling you more about the specific speaker and how it reproduces that particular instrument than anything more generalisable.
I'm not certain it will get you all that much closer to understanding why recordings sound different since recordings aren't normally made in an anechoic chamber so the sound of the recording you make is likely to be very different from the sound of a normal recording of the same piece. You may just end up knowing why a recording made in an anechoic chamber sounds different to the sound of the same instrument live rather than why a live recording of the instrument in a normal performance venue sounds different to the sound of the live instrument.
A simpler test to get a handle on the effect of space on recordings and what then happens on playback isn't too hard. Do you have a large cathedral-type church nearby that has a choir who have released a recording made in the church. Go and listen to the choir live in the church and buy one of their recordings. Listen to the recording on your system at home and see what you think of the difference in the overall ambience as well as the way the voices sound.
My room has some acoustic treatment and I listen near field which also works to reduce room effects. When I put on something recorded in a reverberant cathedral or similar large space, I can "sense" the scale of the space when listening to the record but I am also aware of the "sound" or "feel" of my room as well. The space on the recording tends to dominate but I have an L-shaped room and one side is very different to the other, and I can hear that difference to a greater or lesser degree depending on the record, but it is there so there's always something of my room present along with the sense of the space in which the recording was made. Listening to live music there is always only the sense of one space, the space you are in.
Actually that's a lie but only a little one. I do have a recording of David Hykes and the Harmonic Choir recorded live in New York in a venue that didn't have the reverberant acoustic they liked. They digitised the signal from the mixer, sent it by telephone cable to an abbey in France where they had previously recorded. It was played back on speakers there in real time, miked and redigitised then retransmitted to New York where the feed from the French abbey was mixed into the live feed from the group on stage and played through the speakers so that the audience could listen to the group live with the acoustic of the abbey mixed in. The recording incorporates the acoustic from the abbey as well as from the live site also so playback means you have 3 sets of acoustic info when you count your own room. I have never heard of any other live performance where the performers deliberately chose to mix the acoustic information from the live venue and another space in that way.
David Aiken
David,##the room effects WILL interact with the sound of the recording in the same way that they will interact with the sound of a live instrument##
Yes the interaction with the sound would be the same, but is it that really what makes us recognize the differences, or is it the interaction of the soundfield reflected and direct with the instrument that creates a very specific sonic signature that we recognize as the real instrument vs. the recording of one done in a anechoic chamber. If you play back the sound recorded in a non reverbrant space than the opporunity the alter the sound of the instrument is missing, you only have the opportunity to alter the sound of the speakers. The rest of the room effects would be the same, but having the room feedback into the instrument is not present. The question is really a rather simply minded one: do we recognize something as real because it get acoustic signatures from the resonant space around it or is it just the sound being reverbrated, altered in frequency.
dee
;-D
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
I think i buy it wholesale. I was looking for something that one could call the essence of recorded sound vs live sound, but i guess you are right it sounds a lot more complicated.
dee
;-D
NT
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.Perhaps that explains why Mariah Carey sounds the way she does through my Vandies.
> how come 90 out of 100 will identify the recorded sound in one try?1. Where did you get the 90% from?
2. Keep on going, I am very interested. I do run tests like this (note: like) using a live sound, mike mic-amp, in one room, and amp + speaker in another room ( no recording ).
Oh, if one could seperate all the variables, if.
Jens
You're also leaving out the fact (well, you are and you aren't, actually, but you're leaving out the interaction with the recording system at least) that instruments do not radiate a uniform soundfield, and any recording of one will be missing gobs of soundfield information.Let's start at the beginning. What are you trying to achieve? I think you have an idea here, but it needs some refinement.
JJ - Philalethist and Annoyer of Bullies
This post is made possible by the generous support of people like you and our sponsors:
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2024, Audio Asylum, all rights reserved ]
