Is -140 the limit in measurements?

[stalepie]

https://www.audiosciencereview.com/...ith-topping-d10-jitter-measurement-png.12285/

what does it mean? I see the limit being -140 a lot. Why not lower?

 

[RayDunzl]

Sometimes lower shows up.

-140dB = 0.0000002 volt, if 0dB == 2 volts (decibel differences are a ratio, not absolutes)

That would be in the area of the last bit of 24 bits.

If it's just the end of the measurement range, then I would think resolution limitation of the measuring device.

 

[stalepie]

Another question. Why is it commas are not used after a decibel point, like this? 0.000,000,2
It would make it easier to read.

 

[RayDunzl]

Another question. Why is it commas are not used after a decibel point, like this? 0.000,000,2

Now I have a question.

What's a "decibel point"?

 

[stalepie]

That's like a decimal point, but when written out by someone airheaded like stalepie.

 

[RayDunzl]

I can haz airhead too:

Decibels in Distance:

If the top of Mt Everest is 0dB, then -140dB = 0.035 inches (above sea level)

 

[stalepie]

If the earth is an oblate spheroid it seems like the sea level would differ in some parts (where the earth is "fatter"). Unless that is just the atmosphere.

 

[March Audio]

https://www.audiosciencereview.com/...ith-topping-d10-jitter-measurement-png.12285/

what does it mean? I see the limit being -140 a lot. Why not lower?

-140dB isn't a measurement limit per-se. Remember dB is just an expression of a ratio between two numbers. So, as mentioned, with dac outputs the maximum signal is typically 2 volts rms. However this is not always the case so it's better to express it as relative to full scale output and quote that voltage figure. dBFS.

With 16bit encoding we can express a dynamic range of about 96dB. 24 bit encoding its about 144dB.

However by using FFT processing we can look at signals lower than this. Noise is spread across the entire measurement bandwidth. By splitting the bandwidth into smaller chunks the total noise in that chunk is less than across the entire bandwidth. Amir typically produces FFT plots with a resolution of 32k bins (chunks). The higher the resolution the lower the noise floor will go. I will produce some plots later to demonstrate this effect which is also called FFT gain. So with very high resolution ffts we can look well into the noise floor. There is of course a limit to the noise levels of the measurement equipment itself and thermal noise in components which is basically the noise created by electrons movement which is relative to temperature.

Sorry about my poor explanation

 

[stalepie]

Well is perhaps this thermal noise just below 140 or so?

Why can't we tell these electrons to shut up? They could be taught to pipe down or be sedated somehow. Maybe scatter them so there can be some peace and quiet down there. Or you say temperature, so do the measurements inside a volcano or in Antarctica so that they might be -180 dB.

 

[March Audio]

Well is perhaps this thermal noise just below 140 or so?

Why can't we tell these electrons to shut up? They could be taught to pipe down or be sedated somehow. Maybe scatter them so there can be some peace and quiet down there. Or you say temperature, so do the measurements inside a volcano or in Antarctica so that they might be -180 dB.

We can indeed make the electrons shut up. Just reduce the temperature to absolute zero. This is indeed done in some equipment such as radio telescopes.

 

[RayDunzl]

Well is perhaps this thermal noise just below 140 or so?

This examines Room Temperature and levels as dBm - relative to a milliwatt of power

https://en.wikipedia.org/wiki/Johnson–Nyquist_noise#Noise_power_in_decibels

For audio range:

 

[stalepie]

OK thank you. It is far over my head, of course, but maybe someone else reading finds it interesting.

 

[amirm]

Besides FFT gain, another issue is the test signal. With full amplitude input (or near it) as I use, the noise level goes up in both the DAC and analyzer ADC. With much lower amplitude signals we can do better such as this measurement I showed of RME ADI-2 DAC:

See that we are going below -150 dB. To go any further, I would need to increase the FFT length. Unfortunately my unit stops at 32K so that is the best I can show.

 

[Soniclife]

What would you hear if you played the -144 1kHz tone back and turned up the volume, noise or 1kHz tone plus noise?

 

[March Audio]

What would you hear if you played the -144 1kHz tone back and turned up the volume, noise or 1kHz tone plus noise?

I would suggest noise. Your ear also has a resolution bandwidth, like a limit to the fft resolution. So the surrounding noise will "overtake" the signal at 1kHz. I will try and demonstrate this when Im back home by showing that signal at different fft resolutions.

 

[Soniclife]

If someone can post the test file, or tell me how to make it I'll give it a try tonight.

 

[Frank Dernie]

This is intriguing but are there many power amplifiers with a SNR anywhere near this, so wouldn't a listening test be dominated by amp noise?

 

[Wombat]

What would you hear if you played the -144 1kHz tone back and turned up the volume, noise or 1kHz tone plus noise?

Why are you concerned about this(-144dB?)? Surely it is not something that will disturb your hearing experience.

I am interested.

 

[RayDunzl]

I made a -90dB file, and it plays very softly with the preamp full on (an unbearable level with normal source).

I don't think there would be anything to hear another 50dB down here.

 

[Soniclife]

This is intriguing but are there many power amplifiers with a SNR anywhere near this, so wouldn't a listening test be dominated by amp noise?

Probably, but your amps, and mine, are very quiet, so I was interested in what it sounds like, if anything. I don't consider it relevant to normal playback but I'm interested in how the ear + brain might might compare to the analyser + FFT.