Well I don't know that I'd have to counter everyone's points, but it may be worth discussing a few, which the OP and I have both addressed. From reading the test I think it is well constructed and has a good chance of success.
Some points worth considering:
- Recording technique. Much of this is taken care of by recording in an anechoic chamber. As explained, this eliminates any recorded acoustics which would be mixed and confused with the playback room acoustics. Recording in a well treated room with directional microphones is another way to this, but perhaps not as good. The distance of the mic(s) needs to be well chosen, and that is usually done thru trial and error. The idea is to make the recording sound like a cello that's playing in the playback room. Double acoustics can muddy the sound and be an easy clue.
- Keep it simple. Recording a single instrument raises the chances of success, mostly because of mic technique and speaker placement on playback.
- Playback level. This is critical, as has been discussed. A level that is not matched to the reference will still sound good, but is less likely to fool you than a tightly matched level. If the listener is not directly comparing real to recorded, then level matching is less critical. But level does remain important for a sense of realism.
- Visual cues. One would hope for a blind test, but they aren't always needed. E.G. I've done a test were a string quartet was playing in one room, and it was piped live to speakers in another room down the hall. Listeners could go back and forth between the two to compare and criticize the reproduced sound - and that was useful.
- Fooling people. As any magician knows, it's not too hard to trick people - if you set out to fool people, you usually can. Edison did it with his Diamond Disk (which is remarkably good) by training his singers to sound like the payback.
That's not the goal here, this looks like a very honest test, not a trick.
Well, I agree with you on those.
Mic placement is a little tricky and would hopefully match listener placement. Listener placement is too compromised in a show demo situation. But then if you have a larger crowd, a lot of the sound will depend on the acoustics created by the crowd themselves so maybe not so critical.
Also the listener, if recording in mono, should also only listen with one ear. That might sound like nonsense but I say that on the basis of a single mic v stereo hearing and the differences in what each ear will hear from a quite large source (cellist) v the small source (speaker). Simply moving one's head will reveal the width and height of the source - they would need to be sufficiently far away so the cellist was tending more toward a point source.
The mono mic will record an average of tones reaching it. The real cellist will be sending different tones in different directions so you will further being experiencing different tones reflecting off different surfaces/different absorption and therefore have a much more varied, spacial frequency response again hitting your ears differently due to their spacing. Your brain will interpret that as a larger, real source I think.
And the cellist will move a lot if they are anything like the usual expressive cellist with hopefull the speaker stationary (I do kind of hate that - because musicians playing the same instrument appear to move in the same ways, it feels very contrived to me! Damned musos.. ). So even more variation in tones hitting stereo ears.
It's not atypical experimental practice to reduce as many variables as possible when investigating a particular. Focus first on just direct sound before expanding to source and reproducer dispersion effects. From my perspective it makes perfect sense in this case. I don't know how you could sensibly approach the latter two without fully understanding the first.
Thats true - the speaker would have to be an omni to behave the same as the live sources and excite the room in the same way, especially the cellist.
So now I see the point someone made of needing to listen in an anechoic chamber too .
I think people make this more difficult than it is.
I have seen the Duke Ellington Orchestra, playing with Cleo Lane at the Hammersmith Odeon, as it goes.
You need a great vocalist for a great recording. I never knew if I preferred Frank Sinatra, Bobby Darin or Jack Jones. And I mustn't discount Barbara Streisand and Marvin Gaye to be more inclusive.
Curiously, Kevin Spacey does an extremely good impression of Bobby Darin (who, if you didn't know, had a heart-condition that was going to kill him by the time he was 35, so he wasn't messing about...): YouTube
All the technical stuff is just nonsense, IMO.
I have seen the Duke Ellington Orchestra, playing with Cleo Lane at the Hammersmith Odeon, as it goes.
You need a great vocalist for a great recording. I never knew if I preferred Frank Sinatra, Bobby Darin or Jack Jones. And I mustn't discount Barbara Streisand and Marvin Gaye to be more inclusive.
Curiously, Kevin Spacey does an extremely good impression of Bobby Darin (who, if you didn't know, had a heart-condition that was going to kill him by the time he was 35, so he wasn't messing about...): YouTube
All the technical stuff is just nonsense, IMO.
You ain't kiddin.
So much conjecture in this thread. If you've ever done this, or at least tried, you'll know it's not nearly as complex and difficult as made out here. Really, it isn't.Here's an example. I can put on my Fostex headphones and set up my AT2020 microphone in the kitchen or living room. Once the levels are matched I can listen to my wife talking or whatever else is going on in the room and hear no difference in tonal balance or realism, none. Yes, I hear a little more room tone thru the mic, and also there is no stereo effect if I pay attention, but otherwise it's identical. No difference at all in tonality or naturalness.
And how different is that? Single mono microphone randomly placed, a stereo headphone, no blind test. Far from ideal test parameters. And yet they sound identical, except for a little difference in room tone. Being somewhat more careful with the setup will yield better results.
Anyhow, I say let's wait for the results of this test. This is someone actually making an effort to do something worthwhile, rather than just engaging in conjecture.
If the test is not successful, then you can let fly the "I told you so" and "That could never work because XYZ" comments.
FWIW, I don't think people here are widely off course with the comments and conjecture, it's just that the problems of acoustics, reflections and directivity are, IME, far overstated. With some care and experience, they just aren't as much of an issue as many people think.
And that's a good thing! We are lucky the illusion works and that we hear realistic music coming out of paper cones.
And that's a good thing! We are lucky the illusion works and that we hear realistic music coming out of paper cones.
sincerely,
I do not even understand why it wants to record a sound inside an anechoic chamber.
there is no way to get good sound.
you would put square wheels on your car,
to see what effect does?
maybe I do not speak English well.
Maybe you do not know what you're doing ..
What's worse?
honesty and humility, if we want to learn something.
From what I gather, the loudspeaker system is designed for production work: specifically mixing sound sources that are substantially devoid of room effects (ie. reflections), and then adding-in room effects.
Pro's generally don't want to add-in sound from the control room when they do this (..the worst usually resulting from the mixing console itself depending on speaker placement and absorb-er application to it.)
-it's not an "end-user" goal.
You're right! If the goal of the recording is to let a recreational listener enjoy the music, you need reflections and reverb. But that is not our goal here. Our goal is to replicate the sound of an instrument in a room. We record in the anechoic chamber instead of a normal room, because we don't want to record the acoustics of the room. If we did that, during playback listeners would hear the sound of the instrument with the recorded acoustics, and added to that the acoustics of the listening room. That's two times acoustics, instead only one.
Something along those lines would probably sound more beautiful, but will it resemble a live instrument in the room better?
It's designed for flat direct response, placement close to the front-wall and constant directivity. Great for production work, but also for in a modern home environment, with usually less than perfect acoustics and limited placement options.
For this particular demo, a wider dispersion speaker would probably be better, to better mimic the radiation pattern of the instruments. Also, there's the issue that at least with the cello there is no one microphone position that accurately captures the character of the instrument. During try-outs in de anechoic chamber I clearly heard the sound from the cello change as I walked in a circle around the player.
The reproduction of the spoken voice sounds very similar to live, although the live voice sounds a bit more spacious. Like I said before keys on a chain sound very lifelike. I listened to the snare drum too and I believe it sounded very lifelike, but we didn't yet directly compare it to live.
The recorded cello misses a natural richness. The cello seems to suffer most from the limitations of the setup. First its sound in the anechoic chamber is captured with only one microphone, which does not take into account the very complex radiation pattern of the cello. Second, the speakers used are unidrectional, yet still relatively wide dispersion up to 90 degrees off-axis. Generally those properties are favorable for playing recorded music in a room, but for this particular use it's not ideal.
Good point. You use two ears to locate the orientation, distance and size of a source. With only one ear, it will probably be more difficult to hear the difference between live and reproduced. But then again, for most people listening with only one ear is unnatural and it kind of feels like cheating to me.
FWIW, I don't think people here are widely off course with the comments and conjecture, it's just that the problems of acoustics, reflections and directivity are, IME, far overstated. With some care and experience, they just aren't as much of an issue as many people think.
And that's a good thing! We are lucky the illusion works and that we hear realistic music coming out of paper cones.
Yes, I guess that is the difference between talking about something and doing it.
But hey, people on forums like to think and talk about what we've thought - that's mainly what I do.
as far as the illusion of live performance: since I started listening to open baffles designed with controlled directivity in mind, I can tell you that I prefer listening to well recorded live events which sound the best on those and I can tell the difference in venue acoustics as well.
I am not an expert on the studio subject but I do not believe in mixing individually recorded instruments. Mixing a few microphone locations may be necessary, I am not sure, but anything that involves altering the sound while listening through another pair of studio monitors or headphones is a risk. I believe Linkwitz is far ahead on those discussions with the industry now.
We already know that binauraly recorded sound with mics built into the headphones can offer a near 100% reproduction quality when played back on expensive equalized headphones (I tested it myself with HeadAcoustics hardware I have acces to through my job). The problem is trying to get the same tonal and phase relationship ear-to-ear with sound coming from two points only (assuming stereo), rather than say 4 instruments and a vocal. the only remedy apparently is to disperse the sound again and sum it up at the ears coming from different directions in order to cheat the brain into believing.
Thus I agree with the guy saying here that experimenting with an individual instrument and in an anechoic chamber has little value.
I am not an expert on the studio subject but I do not believe in mixing individually recorded instruments. Mixing a few microphone locations may be necessary, I am not sure, but anything that involves altering the sound while listening through another pair of studio monitors or headphones is a risk. I believe Linkwitz is far ahead on those discussions with the industry now.
We already know that binauraly recorded sound with mics built into the headphones can offer a near 100% reproduction quality when played back on expensive equalized headphones (I tested it myself with HeadAcoustics hardware I have acces to through my job). The problem is trying to get the same tonal and phase relationship ear-to-ear with sound coming from two points only (assuming stereo), rather than say 4 instruments and a vocal. the only remedy apparently is to disperse the sound again and sum it up at the ears coming from different directions in order to cheat the brain into believing.
Thus I agree with the guy saying here that experimenting with an individual instrument and in an anechoic chamber has little value.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.