Understanding Audio Measurements

[Richard Marsh]

Hahaha I got kicked off DIYAudio.com I also bought a 2722 fully loaded when it first came out.

I think I am going to like being here.

Thx - RNMarsh

 

[Astrozombie]

Just FYI there's a typo (missing y) by the PC Testing section. I found it hilarious Amir got banned from AVS

 

[DVinylfan]

Thanks to @amirm for the excellent and succinct summary of measurements.
Not wishing to subvert this venerable thread, but (after a thread titles search) and without a better place to share it, I have the following from Reddit.com:

TonAMGT4
•10mo ago
Can you provide which measurement graph that allows you to reliably tell the different between an electric guitar and a violin playing the exact same music notes?
No, you can’t."

Is this correct?

Because if I purchased an item of hifi gear that measured SOTA and came highly recommended, only to find listening to Frank Zappa sounding like he was playing a violin, I'd be...concerned.

Any opinions, or preferably data, on this?

 

[Kal Rubinson]

Can you provide which measurement graph that allows you to reliably tell the different between an electric guitar and a violin playing the exact same music notes?

Magnetoencephalography (MEG).

 

[kemmler3D]

Can you provide which measurement graph that allows you to reliably tell the different between an electric guitar and a violin playing the exact same music notes?
No, you can’t."

Wow, what? Any bozo with a freeware spectrum analyzer or spectrogram can do that. This is maybe the most "I don't know how to do this and haven't googled it therefore it doesn't exist" comment ever.

 

[DVinylfan]

Nicely these opinions.
The data waits

 

[kemmler3D]

Nicely these opinions.
The data waits

Download Audacity and look for yourself, it's not subtle. It's like asking if cameras can differentiate zebras and horses... The answer is yes, and I can't believe someone is asking for proof.

 

[DVinylfan]

With the utmost respect, no, I won't download Audacity, or this or that app, to settle an argument.
I prefer to do it the old fashioned way, using reason and rationale.

 

[kemmler3D]

With the utmost respect, no, I won't download Audacity, or this or that app, to settle an argument.
I prefer to do it the old fashioned way, using reason and rationale.

"You can't tell the difference between a guitar and a violin by looking"

"Actually you can, it's easy, try it yourself"

"No."

Reason and rationale, everyone.

Fine, I'll just tell you. The harmonics will look radically different in a spectrogram, to the exact extent that they sound different. Pretty trivial to demonstrate. Anyone with recordings of those things on hand can do it. You can find exactly what you need for comparison on freesound.org . Not sure what refusing to do it yourself proves.

 

[danadam]

Looks like harmonics level in 5k-10k range could be a good indicator. Or how long the 3rd and higher harmonics are sustained. Violin at the top, guitar at the bottom (audio files in attachment):

 

[amirm]

Because if I purchased an item of hifi gear that measured SOTA and came highly recommended, only to find listening to Frank Zappa sounding like he was playing a violin, I'd be...concerned.

Since even a $9 headphone dongle doesn't do that, you have your answer. Person making that comment doesn't know which end is up. As noted, we can trivially show the spectrum of different notes being different. Indeed, this is how a synthesizer is able to create these differing sounds in electronic domain.

 

[Travis]

Thanks to @amirm for the excellent and succinct summary of measurements.
Not wishing to subvert this venerable thread, but (after a thread titles search) and without a better place to share it, I have the following from Reddit.com:

TonAMGT4
•10mo ago
Can you provide which measurement graph that allows you to reliably tell the different between an electric guitar and a violin playing the exact same music notes?
No, you can’t."

Is this correct?

Because if I purchased an item of hifi gear that measured SOTA and came highly recommended, only to find listening to Frank Zappa sounding like he was playing a violin, I'd be...concerned.

Any opinions, or preferably data, on this?

4 instruments playing the same note in the 4 spectrograms below using Audicity. Anyone can see the difference. The question quoted seems to ask what measurements would produce a graph showing different instruments playing the same note.

The same can be done with different brands of strings for a string instrument, clear differences. There are published papers on this.

It’s, as others mentioned, easy to tell the difference between instruments looking at spectrograms playing from two different instruments playing the same note. In terms of identification, it takes a lot of experience in looking at spectrograms like these to be able to look as say this instrument is a stringed instrument, a woodwind, a horn, piano. It takes even more experience to be able to look and say what specific instrument it is.

 

[amirm]

The general talking point that we can't measure everything is of course very wrong. I have measured the most minute difference in every audio tweak. People confuse that, with ability to interpret the difference. Many times we don't need the interpretation. If the change with the audio tweak is hugely below threshold of hearing, it doesn't matter what the difference is. It is inaudible.

People also just repeat these points without understanding them. I have engaged them and showed them measurement after measurement yet they just get angry and fight back.