A Subjective Blind Comparison of 3in to 5in Full Range Drivers

[ra7]

Hi BYRTT & X,

I remember reading somewhere, psychological studies reveal that people tend to choose the one appear earlier (or placed higher in the list etc). This applies to many things including menu in restaurants, goods displayed on the racks, name list for everything, and maybe more...

I think this principle more or less reflects the stats we have here.

Let's face it, it's impossible that all voters are in the same degree of seriousness to this. Most would do it casually, if they went through all clips. (I know, several members have contributed their thorough comparing methods. Very good, but I'd bet they are the minority.)

So, for many of the voters, it's very likely the case of the following situation:

A: ooh, very bright, somewhat sizzling, ouch that electric guitar...

B: ah, much mellower, so 'comfortable' (compared to the previous one), this is pretty much it

C: er..., sounds plain, I'm not quite sure....

D: I lost patience...

No offence to anyone, it's just human nature.

So this test is so interesting that not only for audio but also reveals our (human) behavior.

What do you think?

I completely agree with your assessment, CLS.

I tend to use female vocals as a benchmark, but I am not familiar with the singer in track 1 so I had no preconceived Idea of how she should sound. I went with what I thought probably was the most natural, for her voice (without knowing what it actually should sound like) See a problem here?

Had I been familiar with the track/vocalist, I quite possibly would have voted differently. This is also another interesting aspect. How many of us have essentially flat systems? One thing Earl Geddies challenged me on at some point was that if we listen to inaccurate for a long time that conditions us to thinking that is normal. When we hear accurate we think it sounds wrong. Over time however the accurate will win out.

Our preferences can be swayed by years of previous listening experiences which form our own perception of what is right.

C'mon, Tony. Norah Jones? She's good. Buy "Come Away With Me" already!

You make some good points. We get accustomed to the systems we listen to on a daily basis. Any change from that can be perceived as less good. But over the long run, flat and neutral always comes through. Once you get used to flat, it is easy to spot resonances and bad driver behavior.

CLS, based on my listening I'm not surprised by the rankings at all. Only surprise is A. I didn't think it would do as well as it is. Otherwise they're scoring as they sound IMO.

Hmm.... I am surprised by the votes. It makes me think that people didn't take it very seriously. I listened to the files on three separate occasions and using different gear before voting. It is what it is, I guess.

 

[ra7]

ra7,
Think will help a lot mask them EQ flat to microphone point in space, but trained ears probably still can find inherent differences and signatures.
Show here my A10.2 low Q sealed box measured at half a meter with and without EQ. At nearfield listening the EQing sounds great and running squarewave sweep reveal some nice ones here and there in sweep when EQed. Impulse response is cleaned up and so is step response, but have a look group delay the driver wrap steep timing distortion at ~1,6kHz. The timing wrap is about 1ms and this is big compared 1,6kHz period is 0,625ms see tuning fork. Is it right to assume the best drivers will react to minimum phase EQ corrections and clean timing too when frequency response is corrected or they have flat response and flat timing as a minimum phase pass band device. I guess a instrument playing a note in the timing distortion area will have its original sound changed, example it was relaxed at the studio monitor where produced to now a forward note or whatever.
As sidenote the flat EQ doesn't sound good in room only on axis, but i actual use the flat EQ and add EQ smile at HF end to spread good sound into room, but now don't sit on axis its a real 4" cone that can beam hot 🙂.

With a driver that is beaming, you might get away with tilting the top slightly to make it sound good. But I've found that I could never get an acceptable balance on all recordings. The correct balance always starts with a flat on-axis response and then a uniform off-axis response that is usually narrowing as the frequency increases. Remember that what you perceive is a combination of the loudspeaker directivity and your room absorption. If you have moderate absorption and moderate directivity (the most likely occurrence in an average home with furniture, some drapes, etc, combined with regular two-way or three-way speakers), then the balance is just about right. This is just pure coincidence, given the typical driver sizes and the way acoustics works. When you have full-range drivers, or constant directivity horns, you could need special absorption to get the balance right.

 

[pnix]

I tend to use female vocals as a benchmark, but I am not familiar with the singer in track 1 so I had no preconceived Idea of how she should sound. I went with what I thought probably was the most natural, for her voice (without knowing what it actually should sound like) See a problem here?

Yes, that's one of the issues with subjective listening tests. The listener believes he knows what the voice should sound like without actually knowing how the original sounded i.e. how it sounded in the recording studio.

 

[sippy]

Yes, that's one of the issues with subjective listening tests. The listener believes he knows what the voice should sound like without actually knowing how the original sounded i.e. how it sounded in the recording studio.

ANYONE can use ANY female voice as reference...... and theres plenty of females out there.

BBC Radio 4 / BBC World Service is a good place to start if you Can't get one (a female) to talk to you.
Instant Live Reference!

 

[CLS]

But how do you know what exactly her voice in real life sounds like? Do you know any of them and they talk (or sing) to you in person?

We can recognize the sounds of acquaintances through the phone (or other lo-fi devices), but not the other way round -- if we hear through the machine first, we mostly don't have enough clue to restore it to the original.

 

[wintermute]

A great test (for your own speakers) is to connect a mic up and get someone you know well to talk into it. That will very quickly tell you if there are any tonal imbalances (assuming that the mic doesn't impart it's own colour of course).

Any other voice whether it be TV, radio, CD or whatever can only be judged based on previously heard reproductions. These will influence our perceptions.

I think due to some of the deficiencies in my old 3 ways that I listened to for probably more than 20 years, I have developed a tolerance to certain shortcomings which others would probably find intolerable. My current speakers are quite flat but are a bit light on the bass below 200 Hz.

ra7 I shall do a bit of a youtube search for Nora Jones, I know the name just have never listened to her! I'll more than likely get the CD you recommended 🙂

Tony.

 

[irribeo]

Debut cd probably sounds like producer saw best for sales. If you want to hear her voice you need less "produced" live recording, 1 take vs 200+ takes, official (dvd) releases probably already over 10 takes LOL. Or later albums. I believe she mentioned going to another producer to get more natural/real sounding recordings, a less "produced" sound.

 

[Philosophil]

One thing Earl Geddies challenged me on at some point was that if we listen to inaccurate for a long time that conditions us to thinking that is normal. When we hear accurate we think it sounds wrong. Over time however the accurate will win out.
Tony.

I find this very interesting for a number of reasons. Has this 'tendency for accuracy to win out over time' been documented scientifically? If so, can you (or anyone else) point me to the studies that support this?

Thanks.

 

[xrk971]

Wouldn't the actual 40 second long sound clip source (a wav file or whatever was used upstream of the signal chain) prior to it being fed into the speakers tested here be the *reference standard* ? After all, the speaker cannot be asked to do anymore than to actualize the sound as manifested in the content of the recorded bitstream.

Maybe I should post the 40 sec long 320kbit MP3 (now I can fit that because of the 2MB .asc file extension trick) of the actual source for comparison ? You could go and listen to YouTube but that won't guarantee that you are listening to the same source material.

But let me remind folks that you are now talking about which driver is flattest and sounds the most like the source. The definition of a studio monitor - that requirement can easily be found by looking at the freqiency response curve for flatness and balance. From the previous thread, we all know which driver was flattest and would give the closest sound to the original source material. But this test here is for what you enjoyed listening to the most, so the flattest measuring driver may not be the winner - or so we shall see after all is revealed.

 

[wintermute]

Yes that would be the best reference xrk, because that allows us to compare how that sounds on the system we are evaluating on and then it gives a baseline as to what the colourations that the various speakers in the test are adding.

However I think it shouldn't be posted until the key is posted 🙂 It may very well change the way people will/ would have voted!


Philosophil, I think that Toole's research is probably the best known, I was just posting (paraphrased) what Dr Geddes had said to me, and my own personal oppinion.

Dr Geddes challenged me because I was stating that I thought it was more important to taylor the sound to what I liked than to make it accurate. In the end I decided to try what he was saying and go for accuracy over what I thought sounded good. At first I didn't like it, but I persevered, and I haven't gone back 🙂

Now I still maintain that there are factors other than just a flat on axis response which are important, and I'm not saying otherwise. I think once the key is revealed and my subjective impressions released I will have some explaining to do 😉 I know that I am very surprised myself, and not 100% sure what to make of it!

Tony.

 

[sippy]

Listening to anything via youtube is like listening to it down a phone line via a toilet bowl (imho).

My speakers have a few 'issues' so If I had 40 secs of the original track, I could use that as reference, dial my brain into not hearing their issues and go from there.

Having said that, I'm still waiting for my drivers to stop altering as I dosed the whizzer cones lip with Zig 2 way glue Monday night - I 'think' I've managed to get rid of the 'cardboard tube honk'..... YAY!

 

[xrk971]

Headphones, and good ones are really the best way to do this unless you happen to have some studio monitor speakers and a sound treated mixing room.

But, even on your own speakers, if the source and the test driver sound close - then the driver is doing its job well. Just harder to tell.

 

[wintermute]

for Philosophil http://www.theaudiocritic.com/back_issues/The_Audio_Critic_28_r.pdf the first article by toole.

sippy, the comment about youtube was just so I could get an idea whether I like the tracks 😉 try before you buy so to speak.

Tony.

 

[irribeo]

If you do not want influenced voting the thread should have been locked after posting the soundclips.

 

[wintermute]

Perhaps we could run a second poll, with the same tracks (perhaps re-named) and the reference tracks and ask which speaker is the most accurate, as opposed to which one do you enjoy the most?

I suspect that the results would be different 🙂

Tony.

 

[Philosophil]

But let me remind folks that you are now talking about which driver is flattest and sounds the most like the source. The definition of a studio monitor - that requirement can easily be found by looking at the freqiency response curve for flatness and balance. From the previous thread, we all know which driver was flattest and would give the closest sound to the original source material. But this test here is for what you enjoyed listening to the most, so the flattest measuring driver may not be the winner - or so we shall see after all is revealed.

Good point. Many of the claims here are about what people should or ought to like (as a kind of normative expectation or demand), whereas this test is more about what people actually claim to like (as an expressed matter of fact). How the two relate to one another psychologically, physiologically, normatively, and so on, is an interesting question, but one that is probably best left to a separate thread.

Also, thanks for the link, wintermute. I'll be sure to give it a read.

 

[irribeo]

And just full range driver without 2 or more drivers added for bass to hide the fullrange tweeters in this test🙂

 

[AndrewT]

I listened to the first set of 7clips:
I could not tell a difference.

But I did notice crippling levels of noise and a "weird" moment around 26 or 27 seconds.

I wonder if the other posters are right in that the system between the source and the listener is limiting what can be discerned?

 

[xrk971]

If you do not want influenced voting the thread should have been locked after posting the soundclips.

Or use something like Surveymonkey with blind progress.

 

[xrk971]

I listened to the first set of 7clips:
I could not tell a difference.

But I did notice crippling levels of noise and a "weird" moment around 26 or 27 seconds.

I wonder if the other posters are right in that the system between the source and the listener is limiting what can be discerned?

Are you using headphones or listening through your cell phone's speaker? 🙂
Trust me, there is a huge difference if you are using good headphones. Which clip are you talking about an anomaly? Specifically give file name.
There is no clipping in the recording as I monitored levels carefully to keep them at least 10dB below 0dB FS. Then in the final output I adjusted all levels to -1dB peak value so that volume of playback is not a source of difference.