Music and Measurements: What Do you Think? (1 Viewer)

[Lee Scoggins]

Friends,

I just read an interesting comment in Stereophile by William Z. Johnson, designer and founder of Audio Research, where he comments on why audio is so difficult to scientifically measure. He says:

"hi-fi is one of the few industries where products are measured one way and used in another way...Every measurement that we're aware of falls into the realm of what we call repititive, or static, measurement. In the real world, the simplest musical signal has component signals one ten-thousandth the size of some of the other signals present, and at many, many frequencies at once. It simply defies the abilities of static circuitry measurement."

This IMHO gets to the point that it is very difficult to apply even wonderful modern scientific techniques to describe audio phenomena. In many discussions here on the board we often fall on two sides of this coin. Those who don't believe you can measure everything and those who do.

I am curious...what do you believe?

 

[John Watson]

In the early 80's, my early adapting brother wondered why I wasn't careful to only get DDD disks?
I replied that I was hearing the disks - whether ADD or (horrors) AAD, through analogue equipment anyway, ie, my ears

 

[Lee Scoggins]

I replied that I was hearing the disks - whether ADD or (horrors) AAD, through analogue equipment anyway, ie, my ears

Good point. Plus, a lot of DDD are still compressed and aweful sounding, they've just been mastered all digital. Even that though has been exposed on audio journals as often "gamed" by record labels for sales reasons...

 

[Ted Lee]

regarding DDD - didn't i just read somewhere that some of these discs were not "true" DDD, but more like DAD? something about the boards or mixing equipment were really analog?

anyway, regarding the article, i firmly believe that we cannot measure or capture everything.

music is full of dynamics, nuances and "essences" that simply cannot be analyzed. how in the world are you supposed to capture that kind of stuff?

besides all the technical stuff, it's literally impossible to measure the affect it has on a human being. this is what kills me about audio. if someone says they hear "it", then who am i to say that they aren't? and how am i supposed to measure that anyway?

if a piece of music really gets to me, then i'll overlook just about any issue. i certainly won't care that it wasn't recorded using the highest bitrate or that they didn't use original source masters or that the compression was too high.

if i like the music, then that's good enough for me.

 

[Lee Scoggins]

besides all the technical stuff, it's literally impossible to measure the affect it has on a human being. this is what kills me about audio. if someone says they hear "it", then who am i to say that they aren't? and how am i supposed to measure that anyway?

Exactly. There is an emotional connection that can occur.

Thanks for your comments.

 

[Tim Hoover]

if a piece of music really gets to me, then i'll overlook just about any issue. i certainly won't care that it wasn't recorded using the highest bitrate or that they didn't use original source masters or that the compression was too high.

Exactly. There is no measurement for something so subjective as emotional impact - and this emotional impact can affect the sound of anything. Take, for example, an overdriven guitar amplifier. Engineers were horrified that players would be using their amps in this fashion, but the musicians felt that it added a certain "something" to their tone that gave it added emotion. I've overdriven inputs on mixer channels (and horribly fried one as a result )several times for similar effects. How do you analyze and measure a subtle vibrato for emotional impact?...
Add to this the paltry understanding we have of the ear/brain connection in relation to hearing. We can make some generalizations that are somewhat accurate, but with the billions of humans and their unique brain wirings we will fail to achieve anything more.
There are also several camps of audio engineering, namely (the one that Lee falls into) accurate reproduction of live musicians and its polar opposite of using studio technology to sculpt and alter live performances. Which one is better? Objectively we cannot say. Some musicians may actively seek a manufactured sound as more representative of the sound they hear in their heads. In this case measurements may be totally irrelevant as the artists are trying to create something that doesn't exist in the natural world. In the accurate reproduction world, measurements can be helpful because you have an existing sound that you are trying to duplicate as closely as possible - to your ears.

 

[Phil A]

I've always maintained, I don't know how one would go about measuring soundstage depth, height or width with regard to a particular musical source. Also the difference from one room to another could be significant let along measuring stuff like the timbre of instruments or voices. Measurements are good for the sake of measurements and not very useful when it comes to evaluating music on a given system in a given listening space.

 

[Ted Lee]

to touch upon tim's post about overdriving the guitar amp.

that reminded me of (what i call) the "lo-fi" sound that's been pretty popular...especially in electronic music. it has that kind of hissy, distorted, muddy sound.

if you tried to measure that, i'm sure the numbers would be atrocious (sp?). but it sure can sound cool.

 

[John Watson]

Yeah, I once made a tape with my oldest 45's, and used a "mute" button to better absorb the scratches. And I wrote "lo-fi 45's" on the label

 

[Lee Scoggins]

Measurements are good for the sake of measurements and not very useful when it comes to evaluating music on a given system in a given listening space.

Agreed, but there are some that even to use your ears you need to do double blind listening tests. But these tests have their own issues and are not very practical for everyday evaluations.

 

[Kevin C Brown]

I personally think it's a lot more likely that a component that measures well, will sound good too. And that a component that measures poorly will also sound poorly.
For example, I won't buy speakers unless I can find a freq response curve somewhere.
Stereophile (which is ironic in terms of where the quote comes from) and www.audiovideoreviews.com (go to Loudspeakers, and then look for the reviews with thre red NRC logo next to them) are my 2 favorite places. The NRC is one of the pre-eminant speaker R&D facilities in the world, just over yonder in Canada. S&V is OK and Home Theater mag is just barely acceptable (it's obvious their graphs don't have the same resolution as the 1st 2).

 

[Lee Scoggins]

I personally think it's a lot more likely that a component that measures well, will sound good too. And that a component that measures poorly will also sound poorly.

I wish I could believe this Kevin, but I have seen really low THD and other specs on amps which led me to audition them and I found out that sonically they were a mess because they are using lots of negative feedback to boost the metrics and lots of negative feedback is not a good thing usually.
I think you have to look at both the objective and subjective, but rely on the subjective at the end of the day.
By the way, the new MLSSA "waterfall" plots are currently the state of the art for speakers as the frequency response curve has issues in and of itself.
The problem is there are many dimensions to audio - an estimated five key parameters - so metrics can only do 2-3 variables at a time.
As much as I love and use science on a daily basis, there are worlds of innovation needed before the stereo music experience is fully captured.

 

[John Watson]

Lee, the use and abuse of statistics and measurements has filled many books.
I think the comment in Spinal Tap, "as long as its got dubbly noise reduction" aptly sums up the consumer's awareness of what all the techno jargon and data means

 

[Lee Scoggins]

"as long as its got dubbly noise reduction"

So true!

 

[Kevin C Brown]

Lee- I agree with your example of negative feedback and amps. But I truly don't know of any other cases.

But I also somewhat agree: if I am comparing 2 components, and they both subjectively sound the same, but one measures better than the other, I go with the one that measures better.

I do look at measurements more for speakers. It just amazes me when a speaker with a poor freq response response gets a good review. And you cannot argue with the fact that the NRC and Floyd O'Toole (I think that's his name) have done many studies examining what people say sounds good, vs the measurements. "Spiky" or "divoted" curves are bad. Rolling hills are OK. Flat still sounds the best to listeners. (Article in the latest Audio Critic with example spectra and listening test comparisons.)

 

[Lee Scoggins]

Earlier pricey Wilson Audio speakers often had frequency response problems.
Fortunately I never liked them until the Sophia and the much improved Watt Puppy 7.

 

[Zane Charron]

One thing we have to remember is that the main component of listening to music is the listener themselves. Everyone not only has various dips and boosts in their hearing for whatever reason, but each listener has PREFERENCES to specific frequency ranges, possibly due to the aforementioned dips and boosts. Some like their highs boosted a bit or the lows or mids dropped or the sound generally EQ'ed differently. It might sound different or poor to someone who has a different freq response in their hearing or who prefers a different EQ curve.

So it's very true that measuring audio phenomenon is a very subjective thing, at least in how it relates to listening to music and what we think sounds good or bad. Which is why you should never buy audio equipment without first LISTENING to it (if you're an appreciator of good sound anyway).

 

[Kevin C Brown]

Zane- You might want to take a look at the latest Audio Critic. They have the interesting article I mentioned above where they *correlate* listener preferences to different examples of freq spectra. There are basically no surprises in that poor (choppy) looking response curves are *not* preferred by the listeners, and *smooth* curves being preferred.
This isn't some hi fi mag (or Consumer Reports ) doing the tests either. It's...
http://www.nrc.ca/corporate/english/index.html

 

[Zane Charron]

Hmmm, interesting. So much for using logic!

I don't have access to many home theater/hi-fi mags over here (unless I want to pay $10-12 an issue), but I'll take your word on it. Maybe all things (peoples' hearing) being equal, a generally 'flat' freq curve is preferable to one with various dips and peaks. The law of averages, maybe?

 

[Ted Lee]

kevin - i clicked your link, but couldn't find the specific article. can you help me navigate?

also...

There are basically no surprises in that poor (choppy) looking response curves are *not* preferred by the listeners, and *smooth* curves being preferred.

i wonder. we all know most speaker graphs are pretty choppy. i've yet to ever see a really decently "flat" response. so what does a graph of a human voice sound like? it is always smooth or does it also appear choppy?

if it's smooth then your quote would make sense to me. perhaps someone who always attends an opera would naturally find a smoother curve more pleasing.

ultimately we'll never be able to account for (or measure) an individuals preference.