The original question was how digital bits resolve dynamic range. Each sample in PCM is a value of the amplitude of the sound wave. This is represented by a number of bits. In CD this is 16 bits. In blu-ray this is up to 24 bits. The amplitude is sampled at the sampling rate. For CDs this is 44,100 times per second, or 44.1 kHz. PCM is typically sampled at 44.1 kHz, 48 kHz, 96 kHz, etc. For the sample, each additional bits doubles the number of values you can represent. With a single bit you can only represent 0 or 1, or two values. With two bits you can represent 00, 01, 10, and 11, or four values, etc. By adding a single bit to the sample width you double the number of possible sample values. In sound pressure a doubling of volume is 6 dB. So, for each bit you add to the sample width you add 6 dB to the maximum dynamic range. 16*6 = 96 dB of dynamic range for CD quality 16 bit / 44.1 kHz. 24*6 = 144 dB of dynamic range for 24 bit / 96 kHz PCM.
Antonio actually has it nailed! The "db"s used to promote digital audio are absurd and almost meaningless. Although you should have started by explaining that decibels are a ratio, a relationship between TWO measurements. They are not a specific single measurement point like volts, amps or watts. They are the ratio/ relationship between voltage measurement A and B, or Wattage between A and B. going from 5W - 10W is 3db just as going from 10W - 20W is 3db. 100W - 200W is 3db. Then there are dbSPL. That is a specific single measurement. 0 dB SPL, or Sound Pressure Level, is equal to the minimum sound pressure that the human ear can perceive. This is also known as the threshold of hearing and is equivalent to a sound pressure of 20 micro Pascals (µPa). Threshold of pain where ears bleed and hearing damage occurs is around 120dbSPL. OSHA can require hearing protection for work environments with constant 85dbSPL to protect against hearing loss. The average home listening environment is around 45dbSPL. Dolby Atmos specs a minimum of 85db dynamic range. So home theater tries to get down to 30dbSPL which keeps peaks just below ToP. 30dbSPL is almost uncomfortably dead quiet and expensive to achieve. Vinyl is spec'd around 75db s/n. So if your living room is a good 40dbSPL noise floor, vinyl will give you peaks up to 115dbSPL, almost ToP. Some digital file systems give +120db s/n? It can't be fully used in any actual listening environment. Plus most recordings are so heavily compressed they stay within 20db - 30db dynamic range regardless of delivery format.
@@TomasGarza-b5d Correct. db is a ratio of one measurement to another. Watt is a measure of energy/ power/ heat. There is the "acoustic watt", Sound Power Wattage. A symphony orchestra generates roughly one acoustic watt of energy. Literally how much heat the sound waves would generate against a given size surface area. For applications involving human hearing the dbSPL is the common unit of measure. being db yes it is still a ratio, relationship between two measurements. 0 dBSPL is the lowest sound pressure level that the average human ear can perceive, or the threshold of hearing. It corresponds to a sound pressure of 20 micro Pascals (µPa) at a frequency of 1000 Hz.
dB is NOT a unit of measurement of anything. It does not express any particular physical quantity. dB has no units . dB is just a way of expressing (logarithmic) magnitude relationships between different amounts of the same physical quantity. dB can be used to express magnitude relationships between any quantities. Decibels just tell you that this quantity is X times greater / smaller than that quantity. Thus, db of Sound Pressure Level is a different thing than db of Voltage or db of power or energy or of any other quantity. When you discuss dB relationships you have to state which physical quantity you refer to unless it is really obvious by context. In the case of this question dB's of Sound Pressure Level are confused with dB's of digitally represented signal amplitude (Volts). Paul's answer obfuscates the issue even more.
@@Spractral No because in the context of this video, db's are used to express magnitude relationships of at least 2 (SPL, Voltage) or even more different physical quantities. I can't be bothered to dissect how many different physical quantities are involved in the question or in Paul's convoluted answer.
You are correct. Yet again, Paul gets tied up in knots vainly attempting to express a concept in everyday experience. I admire his attempts, which are better than I can manage.
Measuring sound volume in the air is different than measuring recorded values in a storage medium, though you can use the same units of measure in both cases. In a storage medium, a recording medium, decibels describes the dynamic range that you have to work with. It describes the difference between the strongest signal you can store and where the weakest signal you can store, which you can think of as the noise floor. In tape, tape hiss is random noise that exists on the tape that is of a certain signal strength itself. The difference between the energy level of that tape hiss and the largest signal the tape is capable of storing is the dynamic range that is available on that tape. If you’re lucky, the difference between the loudest sound signal it can store and there hiss might be 48 dB. How loud the loudest sound stored on it will sound to you when you play it back via a stereo system and some speakers all depends on how far you turn up the volume knob. If you play it back with the volume knob set so it measures as a 80 dB sound in the air when playing that loudest sound then when it plays a “silent” portion of the tape you will hear that tape hiss at a volume of 80-48=32 dB. No matter how loud you play that tape, the tape hiss that is then noise floor for that tape will be 48 dB quieter than the loudest parts. Having a 48 dB dynamic range does not mean the loudest parts will necessarily be measured as sound in the air at 48 dB loudness. How loud it is is controlled by your amplifier and the control you use to set how much amplification you want it to do. It could be 48 dB, but it could also be 36 dB or 90 dB. That’s separate from the dynamic range available in storing a signal on the tape. A CD has a dynamic range of 96 dB, if we ignore dithering. Including dithering in the discussion just complicates things further, as dithering increases the usable dynamic range over what is possible without dithering, so we will just set dithering aside for this discussion. The situation with the tape is not that much different from the situation with the CD, except the CD’s dynamic range is much larger at 96 dB. That means its noise floor is 96 dB below the loudest signal it can store. How loud it plays back through the stereo is limited in the same way as the tape example. The loudness is controlled by how you set the volume knob. If you set it so the loudest sounds are playing back at 80 dB as we did in the tape example then the loudest sounds will obviously be just as loud as with the tape, 80 dB. But the “silent” parts of the recording that let you hear tape hiss at a 32 dB loudness level will now be pushed down below the noise floor of the stereo electronics itself. We have the same 80 dB volume level for the loudest parts, but subtracting 96 from 80 puts us 16 dB below zero. Since playing sounds blow zero is not possible we are left with however much noise the electronics itself is producing. That won’t be zero, but it might not be very much, either. It may or may not be audible. But since we are playing at a volume level that is a smaller value above zero than the CD’s dynamic range of 96 dB this means the storage medium itself is not responsible for the noise floor in this case. Our volume level must surpass the storage medium’s noise floor in order for its own noise floor to enter the picture. So you have to play a CD louder than 96 dB before its noise floor becomes an issue. This is why we really don’t need a dynamic range larger than a CD can provide. Situations where you listen to anything louder than 96 dB are likely to be few and far between. A “silent” room might still actually have a sound pressure level of perhaps 30 dB. So you’d have to play more then 96 dB above that for a CD’s own noise floor to become an issue. Not many people listening to music at louder than 126 dB, I’d wager. And why we don’t need more than 16 bits for our final playback storage. 24 bits is really only useful in the studio while you’re still mixing things. After you’ve adjusted volume levels and mixed your music just-so, rendering it into 16 bits is more than sufficient. 24-bit storage media for consumer playback makes no sense because of this. Nobody plays things loud enough for it to matter. 24-bit doesn’t store a “higher resolution” signal. It isn’t better in any way other than having a larger dynamic range. The CD’s 16 bits is more than enough on the consumer playback side of things.
@@5starmaniac That has nothing to do with the sample depth and everything to do with all the other electronics before it got turned into digital. If you understood digital sampling properly you’d know this. I suggest you go watch Digital Show And Tell by Monty Montgomery.
@@toltecstrings1 They’re different devices with different electronic circuits. Digital sampling depth and/or rate has nothing to do with the differences you hear between those three devices. It is the quality of the rest of the audio chain in them that you are hearing. Literally none of the differences you hear are due to the difference in sampling rate or sampling frequency. It all comes from the analog stages not being the same circuit design and not being built of the same quality of parts, and the quality of the analog to digital converters. 16/44.1 isn’t going to sound any different to you than 24/48 if everything else is equal. 16/44.1 isn’t less detailed than 24/48. A 20 kHz sine wave that is -10 dB below full scale will be exactly the same with 16/44.1 and with 24/48. Exactly. There will be literally no difference between the two signals. 20 kHz is less than half the sampling frequency in both cases, so both will reproduce it perfectly. Change the sine wave to 23 kHz and the 44.1 sampling frequency version will no longer reproduce it properly because this is above half the sampling frequency, but 48 will continue to reproduce it perfectly. And -10 dB below full scale is well above the noise floor of both sampling depths, so both will reproduce it perfectly. Change it to -100 dB below full scale and the 16-bit version will no longer contain your signal because it is below the 16-bit noise floor, which is -96 dB below full scale. But 24-bit will still reproduce it perfectly because that is still well above its noise floor. As long as your signal is above 16-bit’s noise floor, so greater than -96 dB below full scale, and below half frequency of 44.1 kHz’s sampling frequency, so less than 22.05 kHz, then both 16/44.1 and 24/48 will sound exactly the same. Exactly. No differences whatsoever. Same goes for 16/44.1 vs. 24/96, and 16/44.1 vs. 24/192. If your signal fits into 16/44.1’s reproducible parameters it will sound exactly the same with those other higher sampling depths and sampling rates. Higher sampling rates/depths don’t give you a more detailed version of the same signal.
@@5starmaniac No, it doesn’t. Higher bit depths don’t give you more detail, only a lower noise floor. Higher sampling frequencies don’t give you more detail, only a higher maximum frequency. Digital reproduces your signal perfectly if it is within its reproduction window. As I just said in my reply to the other person. There is no resolution with digital audio. It either works perfectly or it doesn’t work at all. If your signal is within the operating window of 16/44.1 then 16/44.1 can reproduce it perfectly. This means that 24/48 will also be able to reproduce it perfectly. This does NOT mean that 24/48 produce an even better copy with more detail. It doesn’t work like that. If 16/44.1 can already reproduce your signal perfectly it isn’t going to get “even more perfect” with 24/48. Perfect is perfect. Again, there is no resolution with digital audio. It either works because it is within its operating window or it doesn’t work because it is outside of its operating window. 44.1 can reproduce a 20 kHz signal perfectly, and so can 48. But 16 can’t reproduce 23 kHz because that’s above its 22.05 kHz operating window, but 48 can because that’s still within its 24 kHz operating window.
A vinyl record can store only so much information, in the form of etchings in its groove. That is what limits how many decibels a stylus can read via rubbing those etchings to create a voltage, by way of the cartridge's coils) which eventually makes it to your speakers. The electrical current starts at the cartridge's coils. But the information is derived from the stylus rubbing in the groove that then shakes the coils. With digital, your DAC (or CD player, which contains a DAC), creates the voltage that eventually makes it to your speakers. But in the case of digital, the way the information is stored in a file (mp3, flac, wav, etc, that uses PCM coding), describing how much voltage a given byte of datum represents. Your DAC reads that information, and produces a corresponding voltage. So for digital, the files containing the voltage information can contain just about anything. The DAC will create the described voltage, but within its own limits (it cannot create unlimited voltage, and that is what determines its dynamic range limits). If your DAC were to feed too much voltage to your pre-amp, then your pre-amp would choke -- probably spit out ungodly noise. For analog, the voltage begins when the stylus vibrates the cartridge's coils. It has a physical limit to how much voltage it can generate, as well as how much the record's groove can vibrate a stylus to then vibrate the cartridge's coils. For digital, a DAC can be engineered to generate just about any level of voltage based on the information in the file that contains the information about the voltage. A DAC is akin to the coils in a cartridge, in that both the DAC and the coils create a voltage from thin air (in a manner of speaking). The DAC acts on what it is sees written in the music file, and the coil simply shakes based on vibrations from the stylus. @2:48 "A needle." There are no needles in a turntable / tone-arm / cartridge. It is a stylus.
dB is a relative logarithmic measurement. For recording media it is relative to min/max signal (the resolution of the media). For sound pressure (dB SPL) it is relative to 20μPa which is basically considered the lowest sound pressure level a human can possibly hear. Similar for radio sensitivity (dBm) it is relative to 1mW of received power.
Paul needed to explain how the loudness is represented in the digital file. In other words, bit depth. That is what the guy is really asking about. Bit depth = dynamic range.
dB is actually a number measurement of analog signals, so digital is already one step closer to the dB number. As to the log scale of decibels, this is really because hearing is relativistic. A doubling of sound pressure starting at 50 dB (+~6 dB=56) will sound like the same increase as a doubling of 100 dB (106 dB). The average person can reliably identify a 1 dB change, this is the same if that is 50 to 51 or 90 to 91, even though 90 dB is 100 times the sound pressure as 50. However there are also limits to the useful range, from what one can hear at all to the loudest tolerable, so this can overstate the scalability of hearing.
Sounds like after decades of experience working in the audio industry, Paul is still as confused about dBs as that listener was :) dBs are not a unit of measurement. They just express the log of *ratios* between two numbers. In case of digital they express the ratio between the smallest and largest signal that can be stored on the medium. In case of SPL measurements (e.g. jet engine), they express the ratio of the sound pressure between that of a jet engine and 2e-5 Pascals. 2e-5 pascals is a reference point obtained by measuring the quietest sound pressure that average humans can detect.
dB is the logarithmic ratio of something to a standard. (Some EE classes in my past.) A common thing in electronics, and some Bell Labs nerd applied it to sound pressure 100 years ago when coming up with Volume Units ("VU"), setting "0" to the lowest lowest sound pressure most people could hear. As others have noted, in digital, it is signal-to-noise ratio. A good example is when you experience low dB, say, a 5th generation cassette recording. When you hear low dB, you can then appreciate higher dB. Even something like 40dB can be demonstrated--when played at low volume, you don't hear the loss of signal, but crank up the volume and it becomes apparent. (Maybe the 40dB sound-pressure/volume-control is when it starts to become apparent.) I would guess many people won't hear a difference until you get to 70dB or so, but when you are working with a multi-stage audio chain (from the microphone picking up the sound, though the mixing console, A-to-D, audio storage, D-to-A, amplifier, speaker), all the limitations of each stage multiply together, so a 90dB range on a CD is the reasonable minimum for non-audiophile listening. (And the Richter Scale is a measure in dB when it comes to earth movement.)
My understanding of dB measurement is that of a ratio of gain, so get the slide rule out to measure amplitudes of sound as a ratio. My experience of dB usage comes from understanding Radio Antenna gain.
Any representation of a level difference between audio signals (sound wave air pressure, speaker cone position, cartridge needle position, amp voltage, CD PCM level numerical value etc.) can use the decibel scale, which is logarithmic. Why use a logarithmic scale? It is because the human ear functions very much in a logarithmic fashion with the threshold of perception occurring at a doubling of level (e.g. speaker power), which is a 3dB difference. This means that the logarithmic scale is a great way to represent the wide dynamic range of human hearing. A digital PCM value represents a voltage level and every bit more adds a doubling of voltage or 6dB increase of sound level. This also means 16 bits (CD quality) can give you 16x6 = 96dB of maximum dynamic range (level difference between lowest and highest level).
The size of the "wiggles" in the case of a vinyl record have little to nothing to do with "loudness". The size or width of the grooves ("wiggles if you want), have more to do with frequency than with decibels. That said, it is also true that certain frequencies can be louder than others. It is also true that there is cause and effect for different frequencies.
@@Paulmcgowanpsaudio Sorry, but you may be confusing frequency for loudness. Yes, it is true that certain frequencies are louder than others as I stated, but the "wiggles" are directly proportional to said frequencies, not volume control. Just look at a classical record or better still, one of those old stereo test records they used to sell and apply a basic knowledge of sound. (I went with traditional study as well, but that is just me and not a requirement). We will have to agree to disagree on that point.
This time, Paul, it is explained too complicatedly and unfortunately misses the point. "Dynamic range" is the key word. The others explain it in more detail below ;-) Greetings from Germany Johannes
Overly long/complicated explanation for what was a simple question. The real answer: in audio, decibels are primarily used to describe the volume DIFFERENCE between two sounds. 0 dB does not mean the absence of sound. It means the same volume as whatever the reference level is, i.e. it’s context-dependent; you just need to know the context. For pretty much any audio recording format, either analog or digital, 0 dB is the loudest possible signal that the system can handle, and any audio that you actually record will be at negative dB numbers. Bigger negative numbers = softer. For 16-bit CD, their 96 dB S/N just means that the noise level inherent in the digitization is ridiculously soft. You’ll run into other noise sources long before that, e.g. the hiss from analog electronics, and room background noise. That’s very differerent from the absolute dB numbers you get from a SPL meter, where they’re tied to a standard reference level, and the measured dB numbers are positive.
Side Note: 🎵 I Have Noticed Back When Some Of The Top Recording Engineers Back in The 1970’s and 1980’s During The Final Mixing Process on Tape That Used Let’s Say a Studer A80/RC Tape Machine They Remove The Record Head and The Erase Head And Just Leave The Play Head On The Block During The Final Mix Which Removes Even More Potential Tape Friction or Hiss 🐍 Which Can Be Audible At Even Low db Levels 😮 This B.S. Has Been Floating In The Rolodex of My Mind Since The 1970’s 😮 Thanks Paul For This Unsolicited Brain 🧠 Fart 💨 How Much Do I Owe You….?
@@julesc8054The fact that you can move between values anywhere within that range. Any way of storing a signal you can imagine will have a strongest signal it can store and a weakest signal it can store. The range in between those extremes is the dynamic range you can work within. You can move anywhere within that range, but you can’t move outside that range.
@@julesc8054 In short, the bit depth of the audio file or stream dictates the potential for quantitative dynamics. The higher the bit depth, the more accurate the original sound can be reproduced in terms of amplitude swings. For instance, a 24-bit file/stream would be superior to a 16-bit one in terms of dynamic range... with 24-bit yielding a greater difference between the loudest peak and the quietest part of the reproduced audio/music.
@@mattrismatt Sorry just trolling a bit. I think its just range. Dynamic refers to the audio itself, music, as it has a time element. Here's my understanding. I don't know why we call it dynamic range as dynamic implying over time where the format itself is referring to the softest vs loudest of a single sample, no time element. The dynamic of program content is infact dynamic and has a range of soft vs loud that changes over time. The loudness wars is what chops off anything below 60hz and through a series of dynamics compressors / expanders and limiters reduces the dynamic range till we get a loud sossage looking waveform. The issue that you may well know is high dynamic range music sounds soft even if it isn't. So reducing the dynamic range of music basically just gives quieter samples more energy compared to the louder samples and a limiter rounds off the peaks. The result is more energy where it matters.
Is the question are more about the relation in between dynamic range in db and the bitrate 16, 24 or even 32 in new recorder with no dynamic range limits?
Yes, I understood it as a dynamic range question- and the vinyl only has 96db dynamic range. We’re going through this right now with TV technology - the new Sony Bravia 9 is an LCD TV that just gets insanely bright. Which means you have more HDR “ high dynamic range” capability than other TVs. But right now it doesn’t matter as much because the movie content is currently not recorded with as much dynamic range as that TV can produce- It still does matter some though because even if you’re not going to get all of the ratio that the TV is capable of, in a very bright room it’s still brighter than other TVs, which means the entire dynamic range can be brighter than other TVs… which is sort of like listening to stereo in a noisy New York City apartment- you can have two systems both producing the same dynamic range, but you want the one that gets louder (with the softest parts of the music also louder) to overcome the background noise- you need a big amplifier. Or you can get headphones that black out the background noise just the same as with the TV you can get blinds that block out the sun so you don’t need so bright a TV.
As a longtime watcher of Paul’s videos, I cringed when I saw the video title. While I generally believe Paul understands things like this, simplifying explanations of these kinds of topics is not in Paul’s wheelhouse. As the video played out, his attempt to keep it simple unfortunately butchered the answer. Maybe Paul should have started with the explanation of decibels in Wikipedia, basically a widely-used ratio of two quantities (discussion of logarithmic scale in this case could be omitted here for simplicity), and then explained some common usages of decibels in the world of audio, like for expressing loudness, dynamic range, etc.
This is probably the worst explanation of decibels I have ever heard. First let's be clear the 144 dB is not in SPL but in dynamic range. This is completely different and totally unrelated to the 125 or so dB SPL that a jet engine makes at take off full throttle. Next it should be clearly explained that the Decibel scale is not linear and a decibel has absolutely no meaning without a reference next to it. Decibels can be used to measure sound pressure levels, voltages, or in the case of digital audio a scale in relation to maximum level, which is fully dependent on the specifications of the digital scale. So this explanation of decibels is extremely inadequate!
Paul Has Serendipitously Stumbled Over a New Trade of Endeavor if This Audio Gig Doesn’t Work Out By Helping Old Timers ⏱️ Like Myself Release Brain Fart Gas….. Like a Festering Boil Paul Has Helped Me Expel Old Brain Gases That Were Not Well Chewed 😮and Had Been Re/Conjured Through Reverse Osmosis Paul Has Once Again Helped Me Remember Nonsense That At My Age I Really Thought I Forgot About Over 45 Years Ago Thanks Again Paul You Are Truly a Wise Guy and a Genius 😮
I believe the question revolves around dynamic range, CD(16 bits) 96 db dynamic range, bluray(24 bits) 144 db dynamic range.
Correct
The original question was how digital bits resolve dynamic range. Each sample in PCM is a value of the amplitude of the sound wave. This is represented by a number of bits. In CD this is 16 bits. In blu-ray this is up to 24 bits. The amplitude is sampled at the sampling rate. For CDs this is 44,100 times per second, or 44.1 kHz. PCM is typically sampled at 44.1 kHz, 48 kHz, 96 kHz, etc. For the sample, each additional bits doubles the number of values you can represent. With a single bit you can only represent 0 or 1, or two values. With two bits you can represent 00, 01, 10, and 11, or four values, etc. By adding a single bit to the sample width you double the number of possible sample values. In sound pressure a doubling of volume is 6 dB. So, for each bit you add to the sample width you add 6 dB to the maximum dynamic range. 16*6 = 96 dB of dynamic range for CD quality 16 bit / 44.1 kHz. 24*6 = 144 dB of dynamic range for 24 bit / 96 kHz PCM.
Antonio actually has it nailed! The "db"s used to promote digital audio are absurd and almost meaningless. Although you should have started by explaining that decibels are a ratio, a relationship between TWO measurements. They are not a specific single measurement point like volts, amps or watts. They are the ratio/ relationship between voltage measurement A and B, or Wattage between A and B. going from 5W - 10W is 3db just as going from 10W - 20W is 3db. 100W - 200W is 3db.
Then there are dbSPL. That is a specific single measurement. 0 dB SPL, or Sound Pressure Level, is equal to the minimum sound pressure that the human ear can perceive. This is also known as the threshold of hearing and is equivalent to a sound pressure of 20 micro Pascals (µPa). Threshold of pain where ears bleed and hearing damage occurs is around 120dbSPL. OSHA can require hearing protection for work environments with constant 85dbSPL to protect against hearing loss.
The average home listening environment is around 45dbSPL. Dolby Atmos specs a minimum of 85db dynamic range. So home theater tries to get down to 30dbSPL which keeps peaks just below ToP. 30dbSPL is almost uncomfortably dead quiet and expensive to achieve. Vinyl is spec'd around 75db s/n. So if your living room is a good 40dbSPL noise floor, vinyl will give you peaks up to 115dbSPL, almost ToP. Some digital file systems give +120db s/n? It can't be fully used in any actual listening environment. Plus most recordings are so heavily compressed they stay within 20db - 30db dynamic range regardless of delivery format.
then decibels are noy a measure of loudness and watts arent a measure of loudness. Is there a specific measure of loudness?
@@TomasGarza-b5d Correct. db is a ratio of one measurement to another. Watt is a measure of energy/ power/ heat. There is the "acoustic watt", Sound Power Wattage. A symphony orchestra generates roughly one acoustic watt of energy. Literally how much heat the sound waves would generate against a given size surface area. For applications involving human hearing the dbSPL is the common unit of measure. being db yes it is still a ratio, relationship between two measurements. 0 dBSPL is the lowest sound pressure level that the average human ear can perceive, or the threshold of hearing. It corresponds to a sound pressure of 20 micro Pascals (µPa) at a frequency of 1000 Hz.
dB is NOT a unit of measurement of anything. It does not express any particular physical quantity. dB has no units . dB is just a way of expressing (logarithmic) magnitude relationships between different amounts of the same physical quantity. dB can be used to express magnitude relationships between any quantities. Decibels just tell you that this quantity is X times greater / smaller than that quantity. Thus, db of Sound Pressure Level is a different thing than db of Voltage or db of power or energy or of any other quantity. When you discuss dB relationships you have to state which physical quantity you refer to unless it is really obvious by context. In the case of this question dB's of Sound Pressure Level are confused with dB's of digitally represented signal amplitude (Volts). Paul's answer obfuscates the issue even more.
Correct.
dB is a relative unit of measurement equal to one tenth of a bel.
Yes correct, but in this context it IS measurement of something... Pedantically no; in context yes.
@@Spractral No because in the context of this video, db's are used to express magnitude relationships of at least 2 (SPL, Voltage) or even more different physical quantities. I can't be bothered to dissect how many different physical quantities are involved in the question or in Paul's convoluted answer.
You are correct. Yet again, Paul gets tied up in knots vainly attempting to express a concept in everyday experience. I admire his attempts, which are better than I can manage.
Measuring sound volume in the air is different than measuring recorded values in a storage medium, though you can use the same units of measure in both cases. In a storage medium, a recording medium, decibels describes the dynamic range that you have to work with. It describes the difference between the strongest signal you can store and where the weakest signal you can store, which you can think of as the noise floor.
In tape, tape hiss is random noise that exists on the tape that is of a certain signal strength itself. The difference between the energy level of that tape hiss and the largest signal the tape is capable of storing is the dynamic range that is available on that tape. If you’re lucky, the difference between the loudest sound signal it can store and there hiss might be 48 dB. How loud the loudest sound stored on it will sound to you when you play it back via a stereo system and some speakers all depends on how far you turn up the volume knob. If you play it back with the volume knob set so it measures as a 80 dB sound in the air when playing that loudest sound then when it plays a “silent” portion of the tape you will hear that tape hiss at a volume of 80-48=32 dB. No matter how loud you play that tape, the tape hiss that is then noise floor for that tape will be 48 dB quieter than the loudest parts. Having a 48 dB dynamic range does not mean the loudest parts will necessarily be measured as sound in the air at 48 dB loudness. How loud it is is controlled by your amplifier and the control you use to set how much amplification you want it to do. It could be 48 dB, but it could also be 36 dB or 90 dB. That’s separate from the dynamic range available in storing a signal on the tape.
A CD has a dynamic range of 96 dB, if we ignore dithering. Including dithering in the discussion just complicates things further, as dithering increases the usable dynamic range over what is possible without dithering, so we will just set dithering aside for this discussion.
The situation with the tape is not that much different from the situation with the CD, except the CD’s dynamic range is much larger at 96 dB. That means its noise floor is 96 dB below the loudest signal it can store. How loud it plays back through the stereo is limited in the same way as the tape example. The loudness is controlled by how you set the volume knob. If you set it so the loudest sounds are playing back at 80 dB as we did in the tape example then the loudest sounds will obviously be just as loud as with the tape, 80 dB. But the “silent” parts of the recording that let you hear tape hiss at a 32 dB loudness level will now be pushed down below the noise floor of the stereo electronics itself. We have the same 80 dB volume level for the loudest parts, but subtracting 96 from 80 puts us 16 dB below zero. Since playing sounds blow zero is not possible we are left with however much noise the electronics itself is producing. That won’t be zero, but it might not be very much, either. It may or may not be audible. But since we are playing at a volume level that is a smaller value above zero than the CD’s dynamic range of 96 dB this means the storage medium itself is not responsible for the noise floor in this case. Our volume level must surpass the storage medium’s noise floor in order for its own noise floor to enter the picture. So you have to play a CD louder than 96 dB before its noise floor becomes an issue.
This is why we really don’t need a dynamic range larger than a CD can provide. Situations where you listen to anything louder than 96 dB are likely to be few and far between. A “silent” room might still actually have a sound pressure level of perhaps 30 dB. So you’d have to play more then 96 dB above that for a CD’s own noise floor to become an issue. Not many people listening to music at louder than 126 dB, I’d wager. And why we don’t need more than 16 bits for our final playback storage. 24 bits is really only useful in the studio while you’re still mixing things. After you’ve adjusted volume levels and mixed your music just-so, rendering it into 16 bits is more than sufficient. 24-bit storage media for consumer playback makes no sense because of this. Nobody plays things loud enough for it to matter. 24-bit doesn’t store a “higher resolution” signal. It isn’t better in any way other than having a larger dynamic range. The CD’s 16 bits is more than enough on the consumer playback side of things.
Yet, a 24 bit digital sample most often sounds a hell of a lot better than a 16 bit
@@5starmaniac That has nothing to do with the sample depth and everything to do with all the other electronics before it got turned into digital. If you understood digital sampling properly you’d know this. I suggest you go watch Digital Show And Tell by Monty Montgomery.
@@ClaytonMacleod Actually, it has to do with both !
@@toltecstrings1 They’re different devices with different electronic circuits. Digital sampling depth and/or rate has nothing to do with the differences you hear between those three devices. It is the quality of the rest of the audio chain in them that you are hearing. Literally none of the differences you hear are due to the difference in sampling rate or sampling frequency. It all comes from the analog stages not being the same circuit design and not being built of the same quality of parts, and the quality of the analog to digital converters. 16/44.1 isn’t going to sound any different to you than 24/48 if everything else is equal. 16/44.1 isn’t less detailed than 24/48.
A 20 kHz sine wave that is -10 dB below full scale will be exactly the same with 16/44.1 and with 24/48. Exactly. There will be literally no difference between the two signals. 20 kHz is less than half the sampling frequency in both cases, so both will reproduce it perfectly. Change the sine wave to 23 kHz and the 44.1 sampling frequency version will no longer reproduce it properly because this is above half the sampling frequency, but 48 will continue to reproduce it perfectly.
And -10 dB below full scale is well above the noise floor of both sampling depths, so both will reproduce it perfectly. Change it to -100 dB below full scale and the 16-bit version will no longer contain your signal because it is below the 16-bit noise floor, which is -96 dB below full scale. But 24-bit will still reproduce it perfectly because that is still well above its noise floor.
As long as your signal is above 16-bit’s noise floor, so greater than -96 dB below full scale, and below half frequency of 44.1 kHz’s sampling frequency, so less than 22.05 kHz, then both 16/44.1 and 24/48 will sound exactly the same. Exactly. No differences whatsoever. Same goes for 16/44.1 vs. 24/96, and 16/44.1 vs. 24/192. If your signal fits into 16/44.1’s reproducible parameters it will sound exactly the same with those other higher sampling depths and sampling rates. Higher sampling rates/depths don’t give you a more detailed version of the same signal.
@@5starmaniac No, it doesn’t. Higher bit depths don’t give you more detail, only a lower noise floor. Higher sampling frequencies don’t give you more detail, only a higher maximum frequency. Digital reproduces your signal perfectly if it is within its reproduction window. As I just said in my reply to the other person. There is no resolution with digital audio. It either works perfectly or it doesn’t work at all. If your signal is within the operating window of 16/44.1 then 16/44.1 can reproduce it perfectly. This means that 24/48 will also be able to reproduce it perfectly. This does NOT mean that 24/48 produce an even better copy with more detail. It doesn’t work like that. If 16/44.1 can already reproduce your signal perfectly it isn’t going to get “even more perfect” with 24/48. Perfect is perfect. Again, there is no resolution with digital audio. It either works because it is within its operating window or it doesn’t work because it is outside of its operating window. 44.1 can reproduce a 20 kHz signal perfectly, and so can 48. But 16 can’t reproduce 23 kHz because that’s above its 22.05 kHz operating window, but 48 can because that’s still within its 24 kHz operating window.
A speaker’s Peak Transient Composure as it propagates HOW WELL SIGNALS are recorded, is one way of illustrating COMPARATIVE QUALITIES among brands.
A vinyl record can store only so much information, in the form of etchings in its groove. That is what limits how many decibels a stylus can read via rubbing those etchings to create a voltage, by way of the cartridge's coils) which eventually makes it to your speakers. The electrical current starts at the cartridge's coils. But the information is derived from the stylus rubbing in the groove that then shakes the coils.
With digital, your DAC (or CD player, which contains a DAC), creates the voltage that eventually makes it to your speakers.
But in the case of digital, the way the information is stored in a file (mp3, flac, wav, etc, that uses PCM coding), describing how much voltage a given byte of datum represents. Your DAC reads that information, and produces a corresponding voltage.
So for digital, the files containing the voltage information can contain just about anything. The DAC will create the described voltage, but within its own limits (it cannot create unlimited voltage, and that is what determines its dynamic range limits). If your DAC were to feed too much voltage to your pre-amp, then your pre-amp would choke -- probably spit out ungodly noise.
For analog, the voltage begins when the stylus vibrates the cartridge's coils. It has a physical limit to how much voltage it can generate, as well as how much the record's groove can vibrate a stylus to then vibrate the cartridge's coils.
For digital, a DAC can be engineered to generate just about any level of voltage based on the information in the file that contains the information about the voltage.
A DAC is akin to the coils in a cartridge, in that both the DAC and the coils create a voltage from thin air (in a manner of speaking). The DAC acts on what it is sees written in the music file, and the coil simply shakes based on vibrations from the stylus.
@2:48 "A needle."
There are no needles in a turntable / tone-arm / cartridge. It is a stylus.
its called a needle. And the dac doesnt create the voltage, its the amplifier. The dac and coils dont create voltage you need electric power.
dB is a relative logarithmic measurement. For recording media it is relative to min/max signal (the resolution of the media). For sound pressure (dB SPL) it is relative to 20μPa which is basically considered the lowest sound pressure level a human can possibly hear. Similar for radio sensitivity (dBm) it is relative to 1mW of received power.
Paul needed to explain how the loudness is represented in the digital file. In other words, bit depth. That is what the guy is really asking about. Bit depth = dynamic range.
dB is actually a number measurement of analog signals, so digital is already one step closer to the dB number.
As to the log scale of decibels, this is really because hearing is relativistic. A doubling of sound pressure starting at 50 dB (+~6 dB=56) will sound like the same increase as a doubling of 100 dB (106 dB). The average person can reliably identify a 1 dB change, this is the same if that is 50 to 51 or 90 to 91, even though 90 dB is 100 times the sound pressure as 50. However there are also limits to the useful range, from what one can hear at all to the loudest tolerable, so this can overstate the scalability of hearing.
Sounds like after decades of experience working in the audio industry, Paul is still as confused about dBs as that listener was :)
dBs are not a unit of measurement. They just express the log of *ratios* between two numbers.
In case of digital they express the ratio between the smallest and largest signal that can be stored on the medium.
In case of SPL measurements (e.g. jet engine), they express the ratio of the sound pressure between that of a jet engine and 2e-5 Pascals. 2e-5 pascals is a reference point obtained by measuring the quietest sound pressure that average humans can detect.
dB is the logarithmic ratio of something to a standard. (Some EE classes in my past.) A common thing in electronics, and some Bell Labs nerd applied it to sound pressure 100 years ago when coming up with Volume Units ("VU"), setting "0" to the lowest lowest sound pressure most people could hear. As others have noted, in digital, it is signal-to-noise ratio. A good example is when you experience low dB, say, a 5th generation cassette recording. When you hear low dB, you can then appreciate higher dB. Even something like 40dB can be demonstrated--when played at low volume, you don't hear the loss of signal, but crank up the volume and it becomes apparent. (Maybe the 40dB sound-pressure/volume-control is when it starts to become apparent.) I would guess many people won't hear a difference until you get to 70dB or so, but when you are working with a multi-stage audio chain (from the microphone picking up the sound, though the mixing console, A-to-D, audio storage, D-to-A, amplifier, speaker), all the limitations of each stage multiply together, so a 90dB range on a CD is the reasonable minimum for non-audiophile listening. (And the Richter Scale is a measure in dB when it comes to earth movement.)
I think the two examples he cited were signal to noise ratio measurements
My understanding of dB measurement is that of a ratio of gain, so get the slide rule out to measure amplitudes of sound as a ratio. My experience of dB usage comes from understanding Radio Antenna gain.
Any representation of a level difference between audio signals (sound wave air pressure, speaker cone position, cartridge needle position, amp voltage, CD PCM level numerical value etc.) can use the decibel scale, which is logarithmic. Why use a logarithmic scale? It is because the human ear functions very much in a logarithmic fashion with the threshold of perception occurring at a doubling of level (e.g. speaker power), which is a 3dB difference. This means that the logarithmic scale is a great way to represent the wide dynamic range of human hearing. A digital PCM value represents a voltage level and every bit more adds a doubling of voltage or 6dB increase of sound level. This also means 16 bits (CD quality) can give you 16x6 = 96dB of maximum dynamic range (level difference between lowest and highest level).
The size of the "wiggles" in the case of a vinyl record have little to nothing to do with "loudness". The size or width of the grooves ("wiggles if you want), have more to do with frequency than with decibels. That said, it is also true that certain frequencies can be louder than others. It is also true that there is cause and effect for different frequencies.
Sorry, but this is incorrect. The size of the wiggles is directly proportional to the level.
@@Paulmcgowanpsaudio Sorry, but you may be confusing frequency for loudness. Yes, it is true that certain frequencies are louder than others as I stated, but the "wiggles" are directly proportional to said frequencies, not volume control. Just look at a classical record or better still, one of those old stereo test records they used to sell and apply a basic knowledge of sound. (I went with traditional study as well, but that is just me and not a requirement). We will have to agree to disagree on that point.
dB SPL is what you're talking about. db and dB SPL is not the same.
This time, Paul, it is explained too complicatedly and unfortunately misses the point.
"Dynamic range" is the key word. The others explain it in more detail below ;-)
Greetings from Germany
Johannes
Overly long/complicated explanation for what was a simple question. The real answer: in audio, decibels are primarily used to describe the volume DIFFERENCE between two sounds. 0 dB does not mean the absence of sound. It means the same volume as whatever the reference level is, i.e. it’s context-dependent; you just need to know the context. For pretty much any audio recording format, either analog or digital, 0 dB is the loudest possible signal that the system can handle, and any audio that you actually record will be at negative dB numbers. Bigger negative numbers = softer. For 16-bit CD, their 96 dB S/N just means that the noise level inherent in the digitization is ridiculously soft. You’ll run into other noise sources long before that, e.g. the hiss from analog electronics, and room background noise.
That’s very differerent from the absolute dB numbers you get from a SPL meter, where they’re tied to a standard reference level, and the measured dB numbers are positive.
Simple answer is: The measurements is taking after the signals conversation to analog.
Side Note: 🎵
I Have Noticed Back When Some Of The Top Recording Engineers Back in The 1970’s and 1980’s
During
The Final Mixing Process on Tape That Used Let’s Say a Studer A80/RC
Tape Machine They
Remove The Record
Head and The Erase Head
And Just Leave The Play
Head On The Block During The Final Mix
Which Removes Even More Potential Tape Friction or Hiss 🐍 Which Can Be Audible At Even Low db Levels 😮
This B.S. Has Been Floating In
The Rolodex of
My Mind Since
The 1970’s 😮
Thanks Paul For This
Unsolicited Brain 🧠
Fart 💨 How Much
Do I Owe You….?
In which Paul attempts math.
Now am really confused😂
In this case, the pertinent term is called _dynamic range,_ measured in decibels.
I'm curious, what makes it dynamic?
@@julesc8054The fact that you can move between values anywhere within that range. Any way of storing a signal you can imagine will have a strongest signal it can store and a weakest signal it can store. The range in between those extremes is the dynamic range you can work within. You can move anywhere within that range, but you can’t move outside that range.
@@julesc8054 In short, the bit depth of the audio file or stream dictates the potential for quantitative dynamics. The higher the bit depth, the more accurate the original sound can be reproduced in terms of amplitude swings.
For instance, a 24-bit file/stream would be superior to a 16-bit one in terms of dynamic range... with 24-bit yielding a greater difference between the loudest peak and the quietest part of the reproduced audio/music.
@@mattrismatt Sorry just trolling a bit. I think its just range. Dynamic refers to the audio itself, music, as it has a time element.
Here's my understanding.
I don't know why we call it dynamic range as dynamic implying over time where the format itself is referring to the softest vs loudest of a single sample, no time element.
The dynamic of program content is infact dynamic and has a range of soft vs loud that changes over time.
The loudness wars is what chops off anything below 60hz and through a series of dynamics compressors / expanders and limiters reduces the dynamic range till we get a loud sossage looking waveform. The issue that you may well know is high dynamic range music sounds soft even if it isn't.
So reducing the dynamic range of music basically just gives quieter samples more energy compared to the louder samples and a limiter rounds off the peaks. The result is more energy where it matters.
Is the question are more about the relation in between dynamic range in db and the bitrate 16, 24 or even 32 in new recorder with no dynamic range limits?
Yes, I understood it as a dynamic range question- and the vinyl only has 96db dynamic range.
We’re going through this right now with TV technology - the new Sony Bravia 9 is an LCD TV that just gets insanely bright. Which means you have more HDR “ high dynamic range” capability than other TVs. But right now it doesn’t matter as much because the movie content is currently not recorded with as much dynamic range as that TV can produce- It still does matter some though because even if you’re not going to get all of the ratio that the TV is capable of, in a very bright room it’s still brighter than other TVs, which means the entire dynamic range can be brighter than other TVs… which is sort of like listening to stereo in a noisy New York City apartment- you can have two systems both producing the same dynamic range, but you want the one that gets louder (with the softest parts of the music also louder) to overcome the background noise- you need a big amplifier. Or you can get headphones that black out the background noise just the same as with the TV you can get blinds that block out the sun so you don’t need so bright a TV.
As a longtime watcher of Paul’s videos, I cringed when I saw the video title. While I generally believe Paul understands things like this, simplifying explanations of these kinds of topics is not in Paul’s wheelhouse. As the video played out, his attempt to keep it simple unfortunately butchered the answer. Maybe Paul should have started with the explanation of decibels in Wikipedia, basically a widely-used ratio of two quantities (discussion of logarithmic scale in this case could be omitted here for simplicity), and then explained some common usages of decibels in the world of audio, like for expressing loudness, dynamic range, etc.
dB is ratio.
This is probably the worst explanation of decibels I have ever heard.
First let's be clear the 144 dB is not in SPL but in dynamic range.
This is completely different and totally unrelated to the 125 or so dB SPL that a jet engine makes at take off full throttle.
Next it should be clearly explained that the Decibel scale is not linear and a decibel has absolutely no meaning without a reference next to it.
Decibels can be used to measure sound pressure levels, voltages, or in the case of digital audio a scale in relation to maximum level, which is fully dependent on the specifications of the digital scale.
So this explanation of decibels is extremely inadequate!
A decibel is 1/10 of a bel.
First one here!! Lol. Thanks Paul
Paul Has Serendipitously
Stumbled Over a New
Trade of Endeavor if This Audio Gig Doesn’t Work Out
By Helping Old Timers ⏱️
Like Myself Release
Brain Fart Gas…..
Like a Festering Boil Paul Has Helped Me Expel
Old Brain Gases That Were Not Well Chewed 😮and Had Been
Re/Conjured
Through Reverse Osmosis
Paul Has Once Again
Helped Me Remember
Nonsense That At My Age I Really Thought I Forgot About Over 45 Years Ago
Thanks Again Paul
You Are Truly a Wise Guy
and a Genius 😮