Massive Master, this one's for you

[SouthSIDE Glen]

I guess I'll start out by asking what tonality, dynamics, detail soundstage and clarity mean. What do they mean?

I won't pretend to be able to provide quality detailed definitions of each of those offf the top of my head at 7:52am, but they each have one very important property that pink noise does not; they hold specific properties or flavors of information that have meaning to the human ear. Pink noise conatins no such information, and therefore carries no such meaning to the human ear.

The argument can continue from there (though I won't claim it to be all that quantitative of an argument) that distortions in signal that caries detailed information can me more audible to the human ear than equally measurable distortions in a signal that carries very limited information (e.g. a simple sine wave) or one that carries no such information (noise). We can ear changes in information quality - i.e. changes in a signal with meaning to us - easier than we can can hear changes in randomized noise that has no meaning to us.

As an analogy: consider a piece of paper printed with letters of the alphabet versus another printed with random line segements. Now slightly distort the size and shape of the print on each sheet. Give each sheet to your average reader and on which one will they be able to recognize the distortion faster? The one with the real letters.

This is a main basis for the argument that "number measurements" alone cannot have a fixed meaning, and must be taken within the context of the signal type and quality for the measurement to tell us anything meaningful.

G.

 

[bozmillar]

Thanks Glen, all that makes sense. I would say that words like tonality, clarity etc are words that people use to describe the way something sounds, not what is physically happening to the signal.

So let's take a simple example: clipping. To someone who doesn't know what clipping is, they will call it something else, like distorted or harsh or even loud.

With music, the physical effects are subtle and generally a certain sound is a combination of effects. So with one amp, describing it as harsh may mean a lot more than just clipped. It may be clipped, but it might also have some high end noise, low slew rate and an exaggerated frequency response. All of these are physical and measurable characteristics, but all of them work together to describe a components character. Character itself is not a measurable quality, but it is a combination of measurable qualities.

There is a very technical reason why an amplifier may sound better at 0dBu as opposed to +20dBu, even if it is not clipping at +20dBu. There is a very technical reason why it will sound clearer, or have better stage or better imaging, or whatever word choice you have to describe what you hear, but all of these characteristics are derived from easily measurable properties.

It's like describing the weather. You can't know by temperature alone how comfortable the weather will be. it may be 75 degrees outside, but that doesn't describe the humidity, the air pressure, how much pollen is in the air, how much dirt is in the air or even if the sun is shining or not. All of these are very quantifiable and measurable metrics, but combined create days that can be described as nice or comfortable or muggy or chilly or smoggy.

Along those same lines, words like tonality, dynamics, detail soundstage and clarity are words that describe how something sounds, not how something is. All of these words are just there to describe how combinations of frequency response, noise, distortion, etc (which are all quantifiable and measurable) combine to create the sound of a component.

So in that sense, yes you can measure the tonality etc of a device with test signals.

 

[SouthSIDE Glen]

Yep, I agree with your assessment as well. The problem with such adjectives as "tonality", "harshness", "warmth", "clarity", etc. is that the definitions are not necessarily the same from person to person. One person's "clarity" may be another's "coldness", and a third person's "warmth", may be a fourth person's "distortion", and so forth. Then when you add the fact that there is never total agreement as to the "goodness" or "badness" of any such given property - even if they agree on what "clarity" means, they may not always agree on whether that's a good thing or not - and it all becomes so subjective as to be impossible to reach any real agreement as to what it all really means.

Measurement by numbers does give us a comfortable sense of certainty. I may find that a room temperature of 75° is way too warm whereas my mother thinks it's just right, but at least we can both agree that it's 75°. And that *does* indeed tell us something worthwhile. The problem is that the meaning of exactly what it tells each of us differs from person to person.

So yeah, you can test for things like THD and IM and other types of distortions, and test for how such distortions very easily using less complex test signals like sines and weighted noise and so forth, and they will yield certain numbers that do have a real value and an intrinsic value. No question. But the interpretation of the meaning or even the actual usefulness of those values can be far from accurate in real life situations, and can in fact actually differ from ear to ear.

G.

 

[Tim Gillett]

I agree the relationship between "the specs" and "the sound to my ears" ought to be viewed as pretty close.

People not versed in electronics and the performance of amplifiers, or audio production might say "what's that harsh, tearing sound". Another person sitting next to them hearing the same sound will just reply matter-of-factly, "oh that's clipping, ".

Or they might say, "I dont like that gurgling sound when the bass comes in". Another will say matter-of-factly, "oh, that's IM distortion".

I find people are much better at describing objectively say a colour photo whose colour balance is off. "It has a pinky tinge." "It has a bluish tinge" etc. Easy. Even a child can describe that objectively.

But sounds are much more abstract. We often resort to metaphors, images.

And the subjective severity of a certain type of audio distortion varies depending on how it affects our overall enjoyment. The other night I viewed a great movie as part of our city's annual Festival. Dialogue was crisp and intelligible. But whenever there was program below about 100hz it modulated the mids (vocals) badly. (IM distortion)

But because the IM distortion mostly didnt affect the intelligibility or the emotion of the voices, and it was a very dialogue based movie, no one else in the audience really seemed to be bothered (even if at first it drove me crazy!)

And to some extend we can choose to focus on or ignore audio faults.The effect was apparent within the first seconds of the movie but I had to tell myself, "Put it out of your mind. This is shaping up be a great film. The dialogue will stand up and you'll be able to follow the movie." And that's what happened. After a while the IM distortion almost seemed to not be there once I made a conscious decision not to let it ruin the movie for me. I got engrossed in the story. One of the best I've seen in the past year (Bride Flight). Well that was my estimate of it. You might have hated it.

 

[SouthSIDE Glen]

I find people are much better at describing objectively say a colour photo whose colour balance is off. "It has a pinky tinge." "It has a bluish tinge" etc. Easy. Even a child can describe that objectively.

True, But a big part of that is because everybody has a base reference to compare it to; real life. It's easy to tell when the color is filtered differently from the way things are in sunlit or candlelit atmosphere because we all pretty mush know what those look like. Also, we are al used to using our eyes in very detail-discriminating ways.

but there is no analog in real life to what the Black Album or Dark Side of the Moon or Aja (or any other typical album one may wish to cite) sound like, it is a very artificial sound in many ways. It is our brain that fills in the gaps between that artificiality and reality. There is nothing natural or commonplace about the sound of a guitar amplifier, it just doesn't occur in everyday experience unless one is an engineer. I won't even include player, because they rarely ever actually hear how the amplifier actually sounds to the intended audience.

Yes, you have life performances that one can compare to, but objectively, unless one is talking an acoustic trio or something like that, there is only the slightest superficial resemblance between the way your typical live rock/pop band sounds in person and the way a recorded album sounds.

And finally, we are not whales or bats; we are just not wired in common everyday experience to apply that much conscious detail to what we hear the way we are with our sight. The seeing-impaired might be, their brains make up for their deficiency in sight, but for most of us, we have to learn how to listen at that level of detail, and even then most never actually get there.

For the non-engineer, who wouldn't recognize IM distortion from an audience fart, the numbers on a sine wave are close to meaningless, because they have no real reference; i.e. the "information" is not relevant to their experience. Throw that distortion on the sound of their baby crying or the sound of their telephone ringing, or something they are familiar with, then it suddenly becomes audible to them.

IMHO, YMMV, HDTV, PDQ

G.
G.

 

[Bguzaldo]

I agree the relationship between "the specs" and "the sound to my ears" ought to be viewed as pretty close.

So would you buy a guitar or a drum kit or an amp or even a microphone soley based on the specs? Would you say that the specs tell you enough about the sound to the point where you know you'll like it or not? Sure the specs help point you in the right direction, but IMO you have to HEAR it to decide.

Just my 0.02,

I think this is a pretty cool discussion
-Barrett

 

[bozmillar]

So would you buy a guitar or a drum kit or an amp or even a microphone soley based on the specs? Would you say that the specs tell you enough about the sound to the point where you know you'll like it or not? Sure the specs help point you in the right direction, but IMO you have to HEAR it to decide.

Just my 0.02,

I think this is a pretty cool discussion
-Barrett

In a sense, yes (I understand fully what you are saying here, but hear me out). The problem with looking at most published specs, you get very little information from it. THD+N alone does not tell you what the distortion sounds like, it only tells you at what level it is, so two products with the same THD+N spec can have two different flavors of distortion.

This does not mean however that you can't measure the distortion of each one and see how they are different. You don't get that information in a THD+N spec, but you do get it by looking at the frequency and time content of different test tones. Whether you know how to interpret those tones and graphs is a different question, but all the information is there.

In fact, if you know how to interpret it, you can know exactly what a device sounds like. It's the interpreting part that most people don't know how to do.

Same with a guitar. If you know exactly the shape of the neck, the weight of the body, the type of wood, the size of the frets etc, you can know exactly how the guitar feels and sounds. It's just a matter of you knowing how to translate the numbers into "feel."

 

[Bguzaldo]

Hahah all good points, I think the problem is some of us lean towards the subjective and some lean towards the objective.

 

[noisewreck]

Same with a guitar. If you know exactly the shape of the neck, the weight of the body, the type of wood, the size of the frets etc, you can know exactly how the guitar feels and sounds. It's just a matter of you knowing how to translate the numbers into "feel."

Many have tried to copy the Stradivarius violins, by using the same proportins, the same wood and materials... None have succeeded.

 

[bozmillar]

Many have tried to copy the Stradivarius violins, by using the same proportins, the same wood and materials... None have succeeded.

Good point. But doesn't that mean that we are physically missing something? If he could do it, and he could do it more than once, it means there was some physical method to what he did. We just don't know what that was.

And we can measure it. Recreating is another issue altogether.

 

[Bristol Posse]

This thread really got me thinking about this stuff so I thank you for that and I so I went out and got analog VU meters and signal generators to to some testing on my gear and now know precisely where to aim on the DAW metering to get good signals but I gues now I can just stick the VU meters into the chain when level setting I probably won't ever really worry about my DAW meters again

I was also able to confirm that the gain on the preamps was pretty linear on the preamps until a certain point. My AI built ins were more or less linear up to around +8 dBu but the remaining 8 dB after that radically raised the noise floor exponentially baddly over the last 8dB. The preamp self noise at maximum gain barely caused a flicker on the VU meter however registered at -54dBFS self noise on the DAW meters and much more voltage on pin 2 of an XLR cable using a voltmeter . That would be barely perceptable when listening but as soon as you start adding effects like EQ, compression that boost levels you would be getting audible levels of noise/distortion in the tracks

My more boutique pre faired better but the last couple of db of Gain added disproportionate amount of noise/distortion.

**EDIT**

thinking about it some more since the AI and converters are integrated into the same unit I have no way of knowing if it is the converters or the pres that are begining to distort above +8 dbu but either way I know where to keep the levels on that unit. I suppose I could try using the after market pres which I know performed better through after market converters and see if I get thet the same distortion at above +8dBU to get an idea if it is the converters.
****

However now I know how to gain stage properly I know will never have to go into the areas of noise on the pres

Also I know I will never clip my converters.

On the downside I now have no equipment to blame if I can't get my recordings to sound good in the future

 

[bozmillar]

Huh, that's cool. I'll have to try that out on mine too.

I actually read a thread on another forum where someone ran a test tone through his API preamp into a few different A/D converters, and from what he found, the 0dBVU mapped to all different levels on different devices. I don't remember off the top of my head, but on his API A2D it maped to about -18dBFS, but on his cheaper A/D converters it mapped to 0dBFS.

What that tells me is that on certain A/D converters, hotter is better, which goes back to my original post in that I think it is true for most high end and/or vintage analog gear, but that -18dB cannot be assumed a sweet spot for some of today's interfaces. So for those of us who do use cheap interfaces, me being one of them, shooting for -18dB may be doing more damage than good.

 

[Bristol Posse]

on my M-Audio Profire 0 dBVU maps to -14 dBFS with a 1khz test tone. Since most don't state AIs don't state it I guess you have to test each piece of gear in you chain

 

[bozmillar]

lukily for me, my chain is pretty short. mic->preamp->A/D

 

[Massive Master]

So for those of us who do use cheap interfaces, me being one of them, shooting for -18dB may be doing more damage than good.

No interface has 1.23VRMS equating to full-scale. That would be beyond broken.

And for those using "cheap" interfaces, lower signals BY FAR would be doing more good than harm. Converters are ANALOG devices up until the converter chip. The "cheesier" units can distort on their own well short of full-scale.