Understanding Audio Dynamic Range / SNR (Part 1)

[@aynsley544]

Amir, you're a born teacher. Thank you for taking the time to share your profound knowledge and expertise.

[@AudioScienceReview]

That's very kind of you. Thanks for watching. Motivates me to do more.

[@fcracer88]

You’re doing a great public service sir! Many thanks for sharing your knowledge of audio electronics with us in such a clear, concise, and compelling way.

[@luckyupyours]

These types of videos are much needed in the "audiophile" community. Many people do not understand that audio gear are electronic devices that can be measured and tested for level of performance and are not some magical things that defy the laws of physics. Seems like a lot of self-proclaimed audiophiles barely understand electrical theory let alone electronics, so these videos should be very helpful for a lot of people.

[@AudioScienceReview]

Tru dat. :) I can do a lot more of these.

[@DougNoOnions]

Hi Amir, this kind of videos showing how to understand those "names" are amazing, thanks

[@CreamyBone]

This channel is the best antidote to audiophile snake oil on youtube 😜👍

[@RowdyRockyboy]

Oh man, Really wanted this video explanation from you Amir! Thank you :)

[@anthonylewis8382]

Class in session. Another great vid.

[@kevinchoi8076]

Hi, I recently found your video. It is fantastic because I didn’t understand important graphs and numbers. Well, I studied the business administration but I am a music lover, kind of audiophile for 25 years. Therefore, I wanted to understand what the numbers of my gear mean. Thank you so much. I will click every tutorials on basic.

[@jazzfreek54]

Thank you, I love these vids you are doing on measurements, much needed. They will be a valuable resource when I forget in a year and need to review it all again.

[@AudioScienceReview]

Super. Thanks for watching.

[@joserafaelhernandezcarucci1324]

Great video Amir, keep up the fantastic work!

[@AudioScienceReview]

Will do!

[@BadGuyGoodAudioReviews]

Another excellent video.

[@labalo5]

You’re really pumping out content. Thank you.

[@AudioScienceReview]

Thanks for watching them!

[@zyghom]

PERFECT brother! thank you for my "3rd age education" from your channel ;-)

[@TTykwer]

I've gone back to school. And I love it! Thank you.

[@Soundfreeek]

Dear Amir, I have found your channel a couple of days ago and I am blown by the clarity and quality of your contents. You're the first channel, among the hundreds I have subscribed to, for which I turned the notifications on ! :)

[@joshua43214]

I like these much more than your "debunking" vids.

[@AudioScienceReview]

Good feedback. i worry that these are harder for people to follow.

[@willbrink]

Another useful vid minus the usual nonsense. Win win.

[@normanbott]

Fascinating. Just discovered the channel last week and the content is superb. I have an engineering background, so I still regard proper test and specification data as important. Just looked at the spec. booklet for my vintage Quad 33 preamp and it DOES give s/n ratio relative to input level. E.g. RADIO : 85dB for 100 mV input level for 0.5 V main output. This tutorial helps me appreciate the data.

[@AudioScienceReview]

Super. Thanks for watching.

[@pnaubry]

Great video.... looking forward to more like it ... Well done

[@zzz2496]

Wow, finally this topic !!! Amazing video, Amir !! Keep making videos !!

[@the_wau_]

have you ever considered purchasing a vat of liquid nitrogen, submerging a dac or amplifier within it, and then observing the effect on measurements? :) sounds like a fun / interesting experiment to me even if it isn't practical

[@AudioScienceReview]

It would be quite cool, literally. :D

[@SwirlingDragonMist]

You can pickup dry ice at many grocery stores. Just gotta be careful about the condensation on the really cold gear. Maybe put it in a plastic bag when you're all done and returning it to ambient temperature, as being cold in the open air will condense allot of moisture on the internal components. There will even be frost forming on the dry ice, leaving a puddle behind when all the ice is evaporated, you don't want that puddle on your stuff ha ha. Best to keep the whole experiment sealed from air, maybe you put the gear in a sealed vacumed bag before you even start, but the tradeoff there is if you dip it, you don't get the liquid nitrogen directly on the components. Which probably also need to be somewhat warm to work effectively ha ha ha. Capacitor electrolytes can gell up around 15 degrees Fahrenheit, and most parts are optimized for ambient temperatures, suffering if above of below. But theres an interesting phenomenon where electrical resistance creates heat, and then the added heat increases the resistance, which in creates more heat, and more resistance there's a term for that which I can't remember. (anyone know) For me it's compelling that by getting everything cold it would really boost efficiency, and probably lower the noise floor from less thermal noise, but some of the parts might get really wonky, or even break :/ I think to make it sing, the components need to be designed to work at low temperatures, then you would get those resistance thermal runaway advantages, and of course turning your whole hifi into a superconductor is obviously the way to go ha ha.

[@the_wau_]

@@SwirlingDragonMist but dry ice is only at around -80C! take a look at the techniques used in preventing condensation by those overclocking computers with LN2; if you cover the PCB in Vaseline and shop towel you'll be okay for a good while--definitely enough to take some measurements.

[@toneysunny9283]

The BEST AUDIO ch... PERIOD!

[@AudioScienceReview]

You are very kind.

[@eddiejennings5262]

Thank you, Amir. Great review.

[@harryrupam01]

Excellent video Amir, it would be great if you could make a playlist of these educational videos about audio.

[@SwirlingDragonMist]

Very cool man, I know our hearing is it's most sensitive at 1khz and there's this whole convention around it's standardized usage, but wouldn't we ~also~ want to measure at full bandwidth? Maybe not every time for convenience sake, but at least as part of the final delivery. Especially since you've been tasked with debunking, I personally find myself left wondering if there's some emergent phenomenon somewhere else in the bandwidth that might explain people's "night and day" experiences. Not even bothering to look for something just because we know it's impossible to be there, reminds me of many historical moments when established science was so sure of it's self that it rejected the newer and truer theories because they were impossible. I don't know if the analyzer is really setup for simultaneous full bandwidth, or what that would look like graphed, but maybe just another sample at high frequencies, and another at low frequencies. 40hz 80hz 120hz all seem like special candidates for the low, as most stereo tower speakers are capable of getting at least 40hz, maybe 60hz for bookshelf users, so it's a very relatable bass frequency that most everyone would be hearing and perceiving, 80hz because of it's localization characteristics, and all the other reasons it was chosen as an ideal crossover frequency, and 120hz as it's the upper range of the LFE channel. I'm less familiar with what would be a good measurement point for "high frequencies" I feel maybe 10k would be good, or maybe something that is in the range of violins and flutes. Because people always say about their gear, "wow the highs just openend up" or "the Bass flowed like velvet" So I feel we gotta at least throw a dart out there and say look they're not just identical at 1k, the lows are identical, and the highs are identical too. I feel it's an important balance to what other reviewers are saying about products, to have measurements that also consider highs and lows. 0-100khz would be great, because then someone like myself can ruminate over some subsonic or ultrasonic anomaly that may be causing some other audible emergent phenomenon. but 20-20khz seems perfectly reasonable as well. Really glad you're on youtube :D I'm loving your videos.

[@BruceCross]

Yes, I've wondered the same thing, but testing at full bandwidth would be too much work. Testing the select frequencies you mention is possible, but it gets complicated because our hearing varies with frequency. The standards would be different for every frequency being tested. The standard for 1 kHz is 0 dB. Did you notice when Amir mentions our threshold of hearing is 10 dB below zero at 2:21? He's referring to our most sensitive range of hearing at 2-5 kHz (see the Wikipedia article on "Hearing range," Chart #2).

[@SwirlingDragonMist]

@@BruceCross I'm not sure it is anymore difficult to do 1 vs multiple frequencies, he sometimes does a whole spread of like 32 frequencies, anyway on the off chance that there is a non-linear behavior elsewhere on the spectrum we can't see it unless we look, and to add to the link you gave me, scale weighting and equal loudness curves are worth considering too.

[@BruceCross]

@@SwirlingDragonMist Yes, I guess it depends on how automated the process is.

[@Mrch33ky]

Excellent video. Excellent content.

[@TheColeyzzz]

Right On! 😊 I love it!

[@bahathir_]

SNR is a ratio of signal level compared when there is no signal. BTW.. Noise has many types. I wish you can look into this topic too. Thank you..

[@EK-em2zv]

Thanks for making the videos and explaining everything at an understandable level 😁

[@iowaudioreviews]

I've also read the brain reduces dynamic range as things get louder. Thats why we get acclimated to loud music and want to keep turning it up.

[@DustinUnderwoodMKE]

I was just wishing you’d cover this topic. I was trying to teach myself about why ADC results in a noise floor, dependent on the bit depth, and how to characterize that - random, frequency-dependent, controllable by the engineer. The way this all is relevant to normal listening volumes is also interesting to me. I look forward to future videos!

[@AudioScienceReview]

I am glad it was a timely topic!

[@Joonsik_e]

Thank you so much!! Love from South Korea!

[@TomasPiliponis]

Thank you, this is very interesting. And very well explained. Amazing content!

[@aweidenhammer]

Love learning this stuff. You’re a great teacher.

[@AudioScienceReview]

Thank you Andrew. You are very kind.

[@curtgozaydin922]

Amir, good stuff! Explaining concepts like “SNR”, or SINAD or “noise floor” is so important! Explaining the # of bits of dynamic range equivalence thing, woah, I just forget that they came with a name (ENOB) for it but that also a good concept. One slight improvement to your lecture would be where you could’ve tried to show that it’s a logarithmic ratio thing. Or equating the Dynamic Range to experiences like audio in-air acoustics the quietest value of noise you can hear (threshold of hearing, say a pin-drop in the quiet library) versus the loudest jet engine noise when at the airport; this is usually something like on the order of 120 to 134 dB (in-air acoustic sound intensities). Yes, maybe a slightly more advanced concept but possibly useful.

[@AudioScienceReview]

Thanks. Part 2 will be about the second part of your post. There is a lot of complexity there and I wanted to make sure we covered the basics first.

[@dan-nutu]

@@AudioScienceReview was part 2 ever published? I just searched the channel and this one seems to be the only one (for dynamic range/snr). thanks

[@dagnisnierlins188]

I found your old KZbin channel, bring those videos here or the topic of high resolution files.

[@AudioScienceReview]

I thought about that but forgot. Thanks for the reminder. Will do.

[@LynnXternal]

At 2:20 you say that our threshold of hearing is -10 dB, but this is a misleading statement at best and not really true. 0 dB SPL is defined as 20 μPA, and in some situations we can hear below that, but that's only for the most sensitive frequencies (around 3.5 kHz), and even then it's only for young, healthy ears adapted in a pin-drop environment and more like -6 dB SPL, not -10 db SPL. For most people, 0 dB SPL is the threshold of hearing even at the most sensitive frequencies, if not higher. Even in a quiet room, you'd be hard-pressed to hear sounds fainter than 10 dB SPL if the room didn't have acoustic damping to further reduce the noise floor of "quiet environment".

[@miroslawkaras7710]

For The Head phones you shoul keep that 4 volts just attenuate the output of the original signal to 50 mV. That will shift dow signal and noise keeping the dynamic range.

[@hypnoz7871]

Another great video. Thanks Amir :)

[@BruceCross]

Very informative, although I wish it covered A-weighting.

[@ClosDeLaRoche87]

FYI in car audio, 2v RCA is considered low. 4v is good and there are systems with with 5v and 6v RCAs

[@AudioScienceReview]

Good point. There are a lot of sources of interference so higher voltage definitely helps.

[@davidpetersen7091]

A "great" clear discussion. Thanks! Mr. Pete--------> aging hippie

[@petertreyde3212]

Thank you. An excellent presentation.

[@AudioScienceReview]

My please Peter. Thanks for watching.

[@ferossan]

Thank you SO much for this.

[@osasglo]

Hi Amir, that was very enlightening, many thanks for these videos. Please could you in the future do a video on how the volume knob level and gain switch on a headphone amplifier say the thx 789 for example affects the output voltage, dynamic range, and fidelity of the signal, I think a lot of people as well as myself with learn something very valuable from this.

[@glenncurry3041]

Nice job on a difficult subject. I hope you plan on mentioning Threshold of Pain in your next one? You did cover Threshold of Hearing. It might help them to know somewhere up at those upper levels their ears will start bleeding. Before someone decides the 90db of dynamic range they have in a modest noise room needs to go up a bit more! :-)

[@RealHIFIHelp]

Interesting.

[@ChristianGoergen]

You never mentioned, that even the most quiet room has a noise floor around 30 to 35 dBA. I think it‘s useful to know. Is it irrelevant in this context? When you take the pain threshold of hearing in regard, we are far from 115 dB useable dynamic range. Usual level while talking in a room: 60 dB, large office : 75 dB, door slamming :90 dB, from now on : stress and pain and damage. Excellent recordings with commercial distribution (no ambient, noise, industrial, test records etc.) go up to 34 dB. And doubling power, loudness or voltage results in different changes of dB. Doubling loudness is 10 dB more. Doubling voltage is 6 dB more. And doubling power is 3 dB more.

[@AudioScienceReview]

That will be all in the second part. This one was the intro to measurements and details behind that.

[@Sergio83Hi-resAudio]

this data of 30-35 dB is just a lie, it’s just much simpler... in fact, 30-35 dB of noise relative to complete silence is a lot... during the day in a noisy city, if there is traffic outside the window, I’ll still believe in 30-35 dB of thought in room, but at night or in the evening in a quiet place there will not be such a level of noise in the room in real life...

[@ChristianGoergen]

@@Sergio83Hi-resAudio You need no belief. Measure it.

[@Sergio83Hi-resAudio]

@@ChristianGoergen in order to measure this, the microphone must allow measurements from 0 onwards, but they are initially programmed from 30 dB onwards... Do you catch the difference?)) how can you measure something that is not in the measurement range of the sound level meter?)) you can measure it with a phone without any problems there are even indicated values ​​of 0 dB - complete silence, 5-20 breathing, wind, 20-30 - library, 30 and above dialogue.... I measured it more than once, you’re just a theorist or you believe everything that they write and sell you)))

[@ChristianGoergen]

@@Sergio83Hi-resAudio The microphones have been calibrated in an anechoic chamber. No need for your arousal.

[@Eric-xx3mb]

Awesome video Amir! Thank you for educating us all. I'm sorry if I'm a slow learner, but am I right in saying that as you turn the volume down the dynamic range is getting lower so that if you're usually listening to music softly there is less need for equipment/electronics with a high (120dB/20bit) dynamic range? Thank you for your expertise!

[@59seank]

Thanks, Amir!

[@bubbabubba100]

Very interesting! When I listen to music with headphones I’ll typically have my source (whatever is streaming Qobuz to my rpi) turned all the way up but the headphone amp volume knob will be fairly low. Will I get a better signal to noise if I turn down the source but turn up the headphone amp? Thanks!

[@MrBlitzpunk]

Im interested to know what a group delay is, you frequently include that on your headphones review, some are messy some are clean. What does that mean in real life listening?

[@Elazarko]

Hey Amir, I'm looking for the best closed back headphone my money can buy, a do it all kind of headphone but more so for the main use with a digital piano / Rhodes. I'll be trying out the AKG K371 in a couple days but it seems to me they're very flimsy.. (and not the most reliable) a second pair I'm considering is the Meze 99 Nior with Brainwavz oval micro suedes pads that help control/tame some of the the lower frequencies to my understanding. My question is, what pair I should get? Please recommend me a closed back pair that you can think of I can bring my budget closer to $350 if it's really worth it.. (I currently own a "Yamaha Clavinova" digital piano but am soon upgrading to tge "Korg SV2" electric piano/Rhodes. I much prefer the look and quality of the Meze tho I'm concerned the Meze 99 might not fit my ultimate needs. Any other recommendations? Thank you very much.

[@pejbartolo2365]

We can not hear ultrasound but your dogs can up to 50 kHz and probably even beyond that. It might explain why your dogs would be howling to a music.

[@zogzog1063]

My dog loves Beethoven recorded at 24 bit 96 Khz. (no not really)

[@bwoody1090]

It also explains why jangling keys gets them every time, as (from what I've read) the energy level is highest at about 35kHz. I'd guess each key ring would be quite distinct to them!

[@vinnytube1001]

RE: IEMs taking small voltages, what do you think about devices that increase impedance for headphones? These are typically male/female adaptors that add resistors in-line.

[@AudioScienceReview]

They are useful when you have a channel mismatch. I have done a review of them.

[@berlyfredy7153]

Great video professor. Just wanna ask you a doubt. So if im using dac direct to integrated amp and using dac volume control, then is it better to fix integrated amp volume to a level where you can listen music at moderate levels with DAC volume modulated at the upper extremes. I use dacmagic plus and smsl su8 which both of them have been reviewed in ASR. Their volume controls are said to be DSP instead of bit variations.

[@berlyfredy7153]

@Douglas Blake i use asio out from pc to DAC. So pc volume is fixed while streaming bitperfect asio out

[@sebdhaese]

The "best" setup would be to have all the signal strengths at max and do the volume control at the end. The most practical I find, is to have the source (pc or tv) at 50%, the DAC at 100% and the amp tuned to average listening level. That way you can use the volume control of the pc or tv with pretty much no audible difference.

[@Sergio83Hi-resAudio]

@@sebdhaese How are you going to digitally adjust the volume on the DAC by 50% from the PC?))) and who told you that if you adjust the volume on the PC, the signal will not be digitally distorted?

[@Fatih-xl7yn]

is there any audible difference in 2 and 4 volts, if you only use IEM's? Or is it just important for loudness?

[@maxrockbin]

Would putting an attenuator between the amp & phones improve the signal to noise ratio esp for IEMs - since your amp output level would have to be higher?

[@nc3419]

Can I assume if a dac outputs between 2.5 to 3 volts into a the amplifier using rca to rca and you don't hear distortion that doing so should be ok as in safe for the equipment?

[@matthewpool494]

Amir- What about other amplifier specs like damping factor and current?

[@AudioScienceReview]

I will do an amplifier measurement video in the future and cover some of these concepts.

[@matthewpool494]

@@AudioScienceReview Thanks that will be excellent

[@michaeltuohy1249]

I have always wondered what the practical application of dynamic range is in room. If your room noise is 30db, what good is so much range? In order to "cover up" room noise, the volume will get very loud in order to exploit these very high dynamic ranges. Am I wrong in my thinking? As a tool to determine quality, sure. As for it making a difference in our listening? I'm unsure.

[@Sergio83Hi-resAudio]

a simple example, take a fan with a noise level of 30 dB and turn it on in your room, based on your logic you will not hear its noise against the background noise of the room. But I’m afraid that this is not the case and such a noisy fan will be perfectly audible against the background of complete silence... so this is nonsense about 30-35 dB in silence...

[@DEALUX]

I'd like to see a video on IMD as well.

[@alanross3661]

Very instructive. I’ve been watching all of your videos and have learned a lot. Thanks. Do you have any interest in performing tests on power conditioning equipment? I’d love an objective view on whether this stuff is worthwhile in a stereo setup.

[@sebdhaese]

He did a review a couple of days ago about something from PS audio with a lot of explaining about power conditioning. Short answer: not worth it

[@alanross3661]

@@sebdhaese That was the cheap plug in power device. I’d like to see him review a power regeneration unit and whether it makes any real difference.

[@AudioScienceReview]

Yes. I have done some test and plan to do more.

[@SirioAstarot]

What software are you using to perform these analyzes?

[@monsieurVi]

Feel free to add “if you like this video hit the like button and subscribe to this channel for new content”. It’s obvious that this happens naturally due to quality of content but why not to “call for the click” this way? :) Also does this mean that lot of converters are simply not capable to get/reproduce even close to 24 bits content?

[@sebdhaese]

yes

[@AudioScienceReview]

It does. No converter in the world can reach 24 bits. The limit is 21, maybe 22. Fortunately that is all we need. On subscribe bit, appreciate the advice but I get so tired of hearing that on other people's videos that I decided to not mention it in mine. Hopefully people subscribe anyway. :)

[@jitrapornpha5104]

Very useful

[@berkut6313]

Crystal clear

[@sebdhaese]

Depends on the DAC

[@anafabulakz696]

Would it be hard to measure the performance of internal power amp of active monitors? Even pink panther certified monitors, for instance the adam t5v (they're inexpensive though), have quite audible tweeter hiss in nearfield, does it implies high noise floor of amp thus not likely for them to reproduce CD quality sound? Also, could the measured THD be a result of both driver performance and amp noise from active monitors?

[@AudioScienceReview]

It is not hard but I would have to rip them apart and tap into the right bits. Can't do this with loaner speakers so have to wait until I can open one of mine. I suspect they are not very good and hence the reason they are noisy.

[@anafabulakz696]

@@AudioScienceReview Thank you for the response, guess there're still benefit to getting passive speaker+expensive amps

[@idray986]

Wow already 7k subs!

[@SteelBlueVision]

Someone really needs to make a video demonstrating 16-bit or even 24-bit music vs a properly dithered reduction of the music down to 10-bits. Anyone who hears what the extra 6-14 bits does for practically all music one would listen to (i.e., basically, absolutely nothing) would never pay big money for high-end audio again (and would never again question the perfectly acceptable practice of reducing volume in the digital domain vs an analog stereo pot). For most modern music, even 9 bits is overkill, 10 bits leaves a safety margin for the most dynamic recordings listened to at high volumes and with the false assumption that there is no such thing as ear fatigue and adaptation at these loud volumes.

[@Sergio83Hi-resAudio]

you are confusing the concept of loudness with the concept of dynamic range... dynamic range is needed to correctly reproduce quiet and loud sounds in a recording, so that everything does not merge into one loud scream. this is tantamount to saying that you need to artificially make all people the same and then everything will be perfect...

[@Thoughtflux]

Wait, does it imply that lower volumes have lower bit depth and higher volumes have more? Did i understand this wrong?

[@JeremyHansenblue2kid3]

Any AVR ever actually able to reproduce 24 bit (gaming and bluray bit depth) also what is a 44.1k and 48k ect sample rate? 🤔I'm loving the content but I'm confused please more examples like cd 16 bit that was easy to understand.

[@AudioScienceReview]

Nothing in the world can reproduce 24 bits. The limit is about 21/22 bits. There will be more future tutorial and at some point, it should all click. :)

[@iowaudioreviews]

Need one of these on 24bit vs 16bit. Even though 24bit/96khz audio isn't possible how come when people listen to it on their system that's not even capable of CD quality they think the 24bit sounds better... Is it in their head? A lot of music that is upsampled and remastered to 24bit, might it sound better because of the remaster not the 24bit nature? The industry often markets 24bit incorrectly on purpose. If it says 24bit on it all that means is it can DECODE 24bit audio, not actually convert it into an analog signal. I think 24bit has mostly been a marketing campaign and abused in the playback side of things. More light needs to be shed on whats really going on here.

[@sebdhaese]

Decoding 24 bit is easy, just put a 24 bit dac chip in there and done. Producing a very clean analog signal is something else.

[@msd5808]

Seems like dongles should be standardized at 1 volt

[@AudioScienceReview]

1 volt will not be enough to drive high-impedance headphones. Like to see at least 2 volts if not 4.

[@msd5808]

@@AudioScienceReview People should use low impedance, portable phones when they are out and about. Too much power from a dongle drains phone battery.

[@lextr3110]

your missing square wav rendering at 60hz, 300hz & crosstalk in your measurements also phase and time alignment for speakers