Depends on the recording and what type of music. Sounds stupid but it really does. Anything but maybe clasical, 16 bit 44 is fine. When we record in the studio, we use 24 bit and maybe 96k to help with the processing. Basically we are talking about noise floor when it comes to dynamic range. The mic you use, the enviroment, the mic preamp, how soft the sound is you are recording. That is why we use 24 bit in the studio. When the mix is done, 16 bit is fine. Most music doesn't have more than 6 db of dynamic range these days, which is horrible. In the past the most was about 15 db. Max was 30db. So, really 16 bit is just fine for playback. Also, sample rate is much the same. When we mix music and use plugins, it causes major problems. Higher the sample rate we use the better except, a lot of plugins won't go past 96k. Our saying is, if it sounds good it is good. Playback is a different thing. Just don't go to mp3s.
Exactly. This. A great amount of music I collect and listen to is classical music -specifically piano. I play piano myself. And as an instrument, piano isnt just one singular sound. Originally called 'piano-forte,' it literally means the 'soft-loud' machine. It's very essence is a play between the dynamics of bass, and treble. So, high res audio and recording become especially more important in that case, when you're talking about a 9 foot long grand piano (of 30tons, 230 strings, and 18tons of collective tension) there is a wide range of dynamic frequencies, and literally an infinite range of possible harmonics as sound from the instrument resonates and vibrates off of various materials and surfaces in the instrument, the room, and even mic internals -things like the wood species of the piano, hammer mechanism, key material, sympathetic/string resonance, pedal resonances, and all the reflections and reverberations happening continuously over in the music hall as new sound waves interact with past ones, all affect the character and impact of the sound. So, yes for most modern music, with its more limited dynamics and shorter range frequencies, maybe there isnt any practical advantage to 24 bit over 16 bit, or 192kHz over 44.1kHz. Maybe the options are too close for comparison. But I think many can see that for larger more complex, intricate instruments, played within larger acoustic structures, higher res audio and recording is important to match a live, real-world performance for the listener. That's why I buy 24 bit 192kHz+ whenever I can for those special piano performances and I feel like doing so is a must (or else I'm just naive and don't know enough of the science and physics behind sound and sound engineering), at least until my mind is changed by newer info.
We did a blind test the other day and definitely heard the difference especially with material that was more sparse. There was definitely more depth and space at 24. With more dense music it was harder to hear but still there. This was in a studio so you probably would miss these details in noisy environments.
Higher bit depth simply pushes the noise floor down. It has nothing to do with fidelity within the prescribed range. The 24-bit example simply has a lower noise floor than the 16-bit example. The 24-bit example is not capable of “higher resolution” in the volume domain. It doesn’t work that way. It simply has a lower noise floor, not “more resolution.” More bits, less noise. That’s it. That’s all there is to it. Higher sampling rate simply changes the highest reproducible frequency. It has nothing to do with fidelity within the prescribed range. Nor does it have anything to do with phasing, or time alignment, or any other nonsense. All those things are correctly reproduced. If you think they are not then you do not properly understand the system. 192 KHz sample rate does not produce a better copy of a 20 KHz signal than a 44.1 KHz sample rate. Literally nothing about the 20 KHz signal is reproduced better by the 192 KHz sample rate. Nothing. The 44.1 KHz example will be functionally the same in every single way. The only time the 192 KHz sample rate will be better is when the frequencies you’re trying to reproduce lie above the ceiling that the 44.1 KHz sample rate is capable of reproducing. Since 20 KHz lies below that both sample rates will do the job equally. The 192 KHz one will not be more detailed. It will not have “more resolution.” It doesn’t work that way. It either can reproduce the frequency or it can’t. One isn’t better at it than the other. It is either a yes or a no. Yes it can, or no it can’t. There is no degree of quality involved. Either your input signal lies below Nyquist or it doesn’t. That’s all that matters. If it is below Nyquist then it is represented perfectly. If it is above Nyquist it cannot be represented at all. “But the input signal could be timed in such a way so as to fall between the samples, and won’t be reproduced properly in this case as a result.” This is incorrect. That’s not how it works. This can be seen in action on an oscilloscope. “In between the samples” timing is properly reproduced. This case is not the gotcha some think it is. It is not handled better by a higher sample rate because it is already handled properly by a sample rate that adheres to Nyquist. Sorry, Paul, but your ignorance is showing. You do not understand digital audio as well as you think you do. “High resolution audio” is not even a thing because there is no resolution to digital audio. It either works or it doesn’t. It is sad that anyone in the industry is perpetuating this crap when they should know better. The snake oil never ends. Digital sampling and reproduction is not that hard to understand. As long as the characteristics of your input signal fall within the limitations of whatever bit depth and sample rate you are using it is capable of representing it perfectly. Any inaccuracy in that reproduction is the fault of something other than the digital recording side of things. You need enough bits to give a sufficient noise floor, and a sample rate that is more than twice the frequency of that you wish to reproduce. Using more bits or higher sample rate does not improve the results. If you disagree then watch this over and over until you understand why you should not disagree. ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-cIQ9IXSUzuM.html
One could say the same of below 24 bit, if you think about the human propensity to reject information that undermines someone's worldview. And it's probably much easier to tune out things you don't want to hear, than hallucinate things that aren't there.
@@chipsnmydip It's all subjective. As simple as that. There's no factual evidence. Now if we were android and could hear beyond the limitations of 20 -20,000. Actually only infants can hear the full spectrum, as we get older it fades ever so gently. We don't notice unless we are tested. 16 bit covers it. But for those that need to feel like there's more than have at it.
We record EDM in the studio and we know how it sounds after hours and hours of repeating the same loops. We can definitely hear a difference between 16 and 24bit exports. In 24bit, the very low frequencies feel more ample and loose and this is lost in the 16bit exports though it will only be noticed if you can directly compare. Nonetheless, I buy all my music in HR if available. I agree with other arguments like the fact that some HR releases are a scam and that, apart from carefully recorded classical music, you rarely use the whole dynamic range but you cannot deny that some types of music sound better in 24bit and/or high resolution if you have good speakers or quality headphones or IEM.
I highly doubt that your statement here is correct. You might have heard something but I don't think it's got something to do with the bit depth. 16 bit shouldn't really do anything to your bass as that's usually the loudest element. All it does is limit how low in volume oyi cna go before you will get artifacts. In edm you have very little dynamics anyway so this shouldn't be an issue. That's why dithering exists. It will put up some noise floor and this will eliminate any artifacts produced by dropping under the 96db threshold. You can even clearly hear through the noise what's going on "down there" while without the noise it would just be bitcrushed noise itself.
@@StephanBuchin Could be something in how the converters deliver the analog signal? I have noticed a difference in the bass at 16, 24 and 32 bits. 32 bit seems to sound the most balanced/natural, but there can't possibly be an audible resolution difference between 24 and 32bit.
As a long time high res listener and former recording engineer, I will absolutely stick to my guns that 24 bit sounds more nuanced and open than 16 bit, esp for things like reverb. This is on both recording and playback ends. However, I ended up emailing a professor of digital audio in London about why 24 bit would sound different and his answer was: Lots of top engineers insist they hear a difference, but there is actually no conclusive proof or explanation why. That was interesting for sure. I still prefer 24/44.1 to 16/44.1 where it us available.
@@OrangeMicMusic I have, lots of times. To me, the difference is still there. His answer wasn't that various mastering engineers or listeners were wrong, but that there has been little research done specific to bit depth audibility, and there is as yet no definite explanation.
@@chipsnmydip don't mean to argue or anything. Right from the start, a good blind test is difficult to set, to eliminate all the loopholes (which can lead to identify the source), like who is setting the test because you obviously know the songs you're using, then the level is crucial, 0.5 dB louder can give the perception of "better", and so on. For example, easiest way to overcome this is to load into 4U+ BlindTest plugin, few songs each one rendered from the same source at 16 and 24 Bit and play them. The plugin is not revealing what you're playing until you finish the test. On the contrary, there's a lot of researches and studies about this that someone can find today. But if you're one of the gifted in the world, I can't contest it.
@@OrangeMicMusic @Michael Irons I think it is common sense to use files of the exact same volume and audio source. So I start with a 24 bit file and dither it down to 16 bit. I do think that some tests can get carried away, like if the differences in files are on par with the subconscious cue of the way someone clicks on the keyboard or breathes, then they probably are too small to be relevant. The "signal" of the actual sounds should be a good bit higher than the "noise" of test biases etc. More than this, it helps to know specifically what you are listening for, and then it isn't hard to notice it. My understanding is that there are no specific research studies related to bit depth, and those that study high res audio in general tend to be a mix.
I've found that higher bit depth makes more of a difference during production than during listening. Specifically, if you're using a chain of plugins, the higher bit depth produces less artifacts. This is assuming you're at 192 kHz or below, since *most* plugins don't work above that sample rate. Best practices change dramatically when you get to 352 or higher.
But, this becomes the more theoretical so far that we don't make the music ourselves. But what if you are going to buy an album with a certain artist? So can't buy an album that is recorded with higher quality than what that artist's record label has recorded the sound with
The first CD players with 16 bits apparently did not provide 16 bit resolution. It was well type maximum 15.5 . the theory is that a stressed DA converter makes a worse result
Not sure about streaming, but many of us RECORD music at 24-bit/48KHz because 24-bit can lower the noise floor considerably. I can record where the "average" levels are -12dB, enjoy that headroom, and not have to worry about clipping and keep the noise floor out of the picture. Dynamic range is a good thing to have. Poor Chris. He'll come in on Monday and think "Jeez, I'm doing something wrong. This subwoofer isn't working at all like I designed"...Well, nope. Not with Paul dropping it...LOL...
Thanks for the effort! Dynamics is so hard to explain and so misunderstood. Actual room noise levels that hide much of any of even close to 90db. Actual recordings that are released. That most recordings are compressed to maybe 12-24db of actual dynamic range. Or that when the main performer switches from singer to lead guitar, that guitar is not going to hit levels any higher than those peaked out by the singer, or horns to follow, or... the max Zero peak established is the max zero peak. Which I find interesting as a Maggie owner when people talk about it having a lack of dynamics, when all they do is show the lack of dynamics on the recordings. But give them something with actual additional dbSPL and you hear it! now about digital volume controls...
A particular advantage of 24bit recordings comes when requiring to match extreme level differences when mixing. I had a live recording in 24bit where the first section was significantly quieter than the rest of the recording (due to an operator error in setting the initial levels). Because there is so much room to lift the level before coming anywhere near the noise floor it was the simplest job to raise the level of the quiet section without introducing any noise. When I compare that to back in the analogue days it was like magic! Back then you really wouldn't have a hope of matching extreme level differences without introducing very noticeable background noise. I completely agree that on a finished mix for general distribution 24bit is unnecessary because (as explained in this video) there is more than enough dynamic range available even using 16bit.
Well done Paul one of your best!! perhaps now folks will understand the loudness issue better... Compression used in this manner is like pissing in the soup...
There is a Big difference between Available dynamic range and the Actual dynamic range of any digital recording. One benefit of CD and better-than-CD formats is that they Allow the Possibility of wide dynamic range with vanishingly low levels of distortion and unwanted noise. Analog tape doesn't get Close to such performance, and records can't do as well as tape.
Objectively speaking, 24 bit (dithered from far higher bit depth during mixing or mastering processing - *regardless* of the sample rate) preserves more detail and depth of the source than 16 bit. *Subjectively* speaking, if properly dithered to 16 bit there should be very little difference beyond low level dither noise / hiss (which also accumulates at a lower level than truncation distortion from *not* dithering). This works because signal detail and depth of the source is also preserved *within* that low level noise. Higher sample rates afford gentler low pass filters, which helps preserve the HF phase response of the source.
24 bit really isn’t just about the overall dynamic range. There are lots of things to keep in mind here. In many cases, the album master IS the 24/44.1 file - the 16/44.1 is derived from it, and is thus a degraded iteration of the master. You can’t have higher fidelity to the master than the master itself. 24 bit audio also more faithfully reproduces subtle changes in level within the overall dynamic range. And 24 bit audio also provides more data for less destructive DSP, which is commonly used by all sorts of playback hardware/software. I do not find 24/44.1 a waste of money at all.
@@richardpeace4724 On QOBUS you should also be able to buy albums without Flac. now there are also other websites that offer it too With wav or AIFF there should be no compression whatsoever
@@Andersljungberg flac is lossless audio.. it has better data compression.. not audio compression.. it's a more efficient codec that can hold more metadata
I use Apple Music formally iTunes. I have been part of that original ecosystem since iTunes first came out when it was all 128 DRM. Current Apple Music has all the settings I need including EQ all for one price per month. Apple Music incorporates Shazam for free which I use cross referencing music from RU-vid to Apple Music. Apple Music keeps everything organized for me the way I like it. I have a very large music library collection in Apple Music including Google’s Android music that was all ported over to Apple Music when Google shut down that service. I have Apple computer, iPad, iPhone. It all works together seamlessly. The latest Album I ported over from RU-vid using Shazam to Apple Music was the album “With the Wolves” by Dirtwater.
Great ❤ video, Paul. I couldn’t agree more on your answer. 16-bit gives a dynamic range of 96 dB. 24-bit gives a dynamic range of a staggering 144 dB! And 32-bit float has 192 dB of dynamic range. Like you said, the sample rate is also very important when it comes to dynamic music. Although I love listening to high resolution lossless audio (24-bit/192kHz for example), I rarely use such high sample rates when recording. Vocal recordings on 48kHz 24-bit does the job just fine for my main genre (Trance music production).
I have always been a sampling rate person ! I record in 176 because the 192 in my system has a glitch !I love the high sampling rates because to me it equals resolution . I have a friend who has a very nice recording studio and he records at CD level all the time because he says it will end up there anyway . I disagree , even after a higher resolution recording is dumbed down I can still hear a better sound ! I never thought about the phase alignments and that makes a lot of sense , Thank You!
I think when making music it's good to record in higher rates as you'll be mixing a lot of stuff together. But for general listening since it's one track, probably doesn't matter so much.
The extra 8 bits would be helpful with something like a solo violin or cymbal tap that fades to black. The cadenza in the Tchaikovsky violin concerto is a good example. More steps as the level approaches zero would be audible. But that assumes a system and listening room with VERY low background noise. There is very little music recorded at 44.1/24. The USB version of the Beatles box from 2012 was 44.1/24, but that was obviously an analog recording, and supposedly has extra compression 🙁
A pure player that resembles a CD player is unfortunately probably better than a computer. because a computer is probably not made to be a hifi product
There is a difference between analog dynamic compression and digital dynamic compression. And today, obviously, the dynamics can be significantly more compressed So that everything almost sounds the same loud in the worst case
I've got a much shorter question: "Before listening to your products and recordings does the listener get offered an audiology test to allow them to get some idea how their hearing is performing?"
Anybody here did a double blind test to see for himself? I don't care about showing me speakers, talk about bit depth and whatnot. Just get a plug-in (like 4U+Blind Test) or ask a friend to play for you 30 songs at various bit depths. You'll be disappointed about your ears, knowledge or equipment if you think it makes a difference
Prior to starting collecting vinyl again 2 years ago, I was heavy into FLAC files and there is a chasm between 24 bit and 16 bit, if played on a dedicated FLAC player over a stereo. I had both files and removed all 16 bit files and kept only 24 bit!
Listening and comparing Peter Gabriel and Pink Floyd albums that I know are 16 and 24 bits equals I clearly hear a snappier and punchier sound at 24 bits, drums specificly. Problems is that you rarely have the same recording & mastering available at both bit depth. Often you get a re-master when the hi res is released.
Oooh! That sub does look sah'weet! Cast frame, dual spiders, nice! My assumption on the goal here being, combine a pair of the upcoming subs with the FR-30 and you got yourself a PS Audio IRS V killer.
Yup...I was expecting something much larger honestly if that is what Paul was describing for a near field sub previously...but you are probably more correct with your assumption. ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-aP_Z3s8Zas0.html ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-TWZrBL7S6C0.html
So what is best for me on a Computer PC. i have a sound card that can play 16-bits or 24-bits and betwen 44.1 Khz - 48 Khz - 88.2 Khz - 96 Khz - 176.4 Khz - 192 Khz ... Witch one is best do you think? Thx for a good video. / Sweden!
About dynamic range, I have yet too experience more of that than Mike Oldfields "Amarok". And that was on CD in the early 90´s. Most hires I have heard didn´t even try to go nearly as far as what Oldfield did back in the day of Amarok.
I hear more improvement in sound going from 16 bite to 24 than i do going from 44.1 to 88.2. Only incremental minor increases going above 48K. Maybe it’s the equipment, but we did a test with some nice studio equipment and found that to be the case…
@OctaveRecords00 thanks for your teaching sessions here! As audiophiles, we can be fooled by sales gimmicks like stones that make the sound better, we can go on and on…
agree, I live in a home in the suburb with double glass windows and no nearby traffic yet the noise floor would not go significantly below 30 db and I would never listen to significantly more than 90 db (with the db meter on ‘C’ position)
Current issues are poor recording more than the bits !!! but... in some cases, artist who care to record in high samples does care about good recording and good sound engineer... not always ... but... more often!
It's not about listening, I just want to have the best available source as a start. I downsample to 48k and convert to lossy (usually Apple AAC) depending on target device anyway - because I won't ever hear a difference anyway.
If the correct equipment isn't utilized, a 24 bit file will be played at a lower bit rate as a laptop, for example, will likely play a 24 bit file but the onboard DAC won't present a decent representation of it. For instance, I often use an audio interface to record at 32 bit Float, 96kHz. But I'll edit that recording on a laptop WITHOUT the audio interface. Apps like Adobe Audition can handle the file just fine, but I know the built-in audio DAC on my laptop ISN'T going to give me that 32bit 96k in my standard headphone jack. It's an on-the-fly downsampled representation. Did the author of this question list the gear he used?
I hope everyone that knows these differences don't seriously use the laptop dac and headphone jack for listening, indeed these don't do the job. Use an external dac amp for this, even a cheaper model (like Ifi) will be better.
@@robinr5787 @richardcruzjr Modern MacBook Pros provide the ability to match the laptop headphone output to the source file either 44.1, 48, 88.2, 96kHz and bit rate to 32bit float.
the thing a lot of people miss when discussing 16 vs 24 bit recordings is that the extra 8 bits are not tacked on top of the 16, so you gain new headroom ... they are in fact tacked onto the bottom so that you gain more depth of range. That is... it's not letting you record louder stuff... it is letting your record quieter stuff... Quieter stuff that nobody can hear. Most mixing and mastering now toes the line with the normalization standards of streaming services. This means you are recording for an average level (0 VU) at -16dbfs which is 16db below digital clipping ... this effectively gives you 16 db of headroom for snare shots, special effects, etc. no matter the bit depth of the samples or the actual recording levels you used.
@@nicksterj PCM, and all digital audio for that matter, has a hard limit at 0dbfs which is when every bit in the sample is a 1. You ain't going beyond that no matter what you do.
I've found that higher bit depths are useful in the recording and production process, if you want to manipulate the files with processes like time-stretching etc... Also can be handy for things like field recording... But, for the end user, the listener, it's completely unnecessary and pointless...
It is all about resolution first then dynamic range & compression (cutting out and throwing away digital audio information on the floor, and audio compression/limiting) & whether you "yourself" with good? average? bad ears? can hear differences in any, or all of those things. Do people today prefer to watch video in 480p, 720p, 1080p, 4K or 8K (today people mainly listen to 240p-like 320kbps mp3's when 1987 CD/Wav files are the video equivalent of 4k today WTF)??? It is really that simple. It is mainly about the clarity & life-likeness of the video image or audio image. Whether specifically "you" have the ears to appreciate/differentiate is something for each individual. The physics of highest resolution possible digital audio by the numbers relating to better audio quality is sound.
Upsampled PCM may not improve the recording, but makes DAC easier and with better results. Also 24 bit doesn't enhance music itself as dynamic range isn't that big, but improves noise floor due to dithering.
OMG you're ow so right about dynamics!!! Why o why are all recordings compressed... Only a few recordings use the dynamic range of a CD. And it's great to experience that. What a waste of dynamics. Dynamics are very important in music.
Hi Paul. We've both been around this industry longer than I like to think about.... I agree with you on 24 bit 44.1 not being worthwhile. Qobuz labels it as hi res but I really consider that a bit of a misnomer. I disagree with you on the true hi res PCM issue. I'm not a big SACD fan, largely because of the issue with being limited, for the most part, to one piece players. I also find some SACD's to be overly smooth. A really good hi res, say 24/192 recording is well worth while plus you don't have the incredible limitations of no post product being practically possible. I will always choose great music over great recordings of music I don't care about. I do rip my favorite, mostly classical, LP's at 24/192 with an Ayre A/D convertor. The area of SACD/DSD that interests me most is great analog recordings of classical music, the early Deccas, RCA Living Stereo, Mercury Living Presence, etc. Still have the issue of being stuck with the onboard DAC, though. Best to you and your lovely and very pleasant wife.
Paul is correct about the bit depth but then veers into pseudoscience regarding the sample rate. There’s no significant musical information above 20 kHz, most microphones won’t capture it even if there was, and no human can hear it anyways. It’s just ultrasonic noise. The Redbook (CD) engineers knew what they were doing when they created the standard.
You may be surprised to know that electronic music from many artists via Bandcamp is released as 24-bit 44.1KHz files and I have to assume that is how they were recorded.. In fact I've even downloaded some albums from there that offer a mixture of both 16-bit 44.1kHz and 24-bit 44.1KHz files and, believe it or not, I generally find the higher bit-rate files sound cleaner, and, judged from the music, I don't think a wider dynamic range enters the minds of those doing those recordings. I believe these are all "home" recordings and the sound is in most instances tolerable, and sometimes a little better.
@@chipsnmydip I like many of the artists who are offering electronic music on Bandcamp, and the only alternative to file downloading is buying CD-Rs and I've had a number of bad experiences in the past with them (muting, skipping, noise) so I avoid them like the plague now.
I don't understand what all the fuss is about. Qobuz resolutions run all the way to 24/192 and a graphic appears on the player/library page saying so. I have Spotify premium and a Qobuz account. $9.95 vs. $12.95 per month respectfully. Nothing to be concerned with there if money is the issue. It's cheap enough for what you can get within an instant! Just my experiences.
Nice to see you here! What do you think of the high Res snobbery? As far as I can tell in ABX testing, backed up by every study I can find, that nobody can hear a difference between 16 and 24 bit.
24 bit and high sample rates are very useful in the recording studio. More dynamic range and lower noise floor, but the big advantage is that they give the effects plug-ins in the DAW more data to work with when the mix is being created. If you want to record an unblemished live acoustic performance, you would use two high quality microphones, no compression, equalization, reverb, delay etc. at all, and record directly to 44.1 kHz at 16 bits just like it is going to be on the CD. Pure and almost untouched by anything other than the microphones, mixer and the analog to digital converter themselves and right onto the digital file that goes onto the actual CD. Once you use the plug-ins of the DAW you are recording with on a 24 bit 192 kHz recording you have completely changed the recording at a bit to bit level and have done that many times in course of creating the final mix and in mix down to 44.1 kHz 16 bit CD master. You tell me which method would actually give you a more accurate final recording.
Best final recording would be to record in 24 bit, 48khz should do. You then dither the file to 16 bit giving the 16 bit file it's maximum theoretical resolution. If you record to 16 bit, you have to back off the levels a bit and you might end up with 12-14 bit resolution.
@@lundsweden Not that I really insist that a particular recording be done straight off the mic into the ADC you're using, but when you dither from the theoretical 24 bit 48khz to the final 16 bit 44.1 khz file, the dithering algorithm is making an assumption about what the final file data is. The direct to 16 bit 44.1 khz conversion that the ADC does will not involve an additional step in the process that dithering entails. Recording a 24 bit 96 khz original capture is most likely going to yield better results in the final mix especially if you use any sort of effects (Equalizer, compressor, gain control, etc) in the DAW to create the sonic signature that is desired. My thought experiment is which recording would be most accurate, if the only modification was setting the gain level on the mixer (Tone controls and Aux sends bypassed) to avoid clipping or the one dithered down to 16 bits at 44.1 after being recorded the same way with higher settings. Just being picky I guess.
Oh dear Paul. Bit depth does indeed control the dynamic range, but it has nothing to do with resolution. You should know that. Higher sampling rates and bit depths can be useful in the recording process so data can be processed without impacting the audible range, but it is wasted in domestic playback systems. Try Mark Waldrep’s Hi Res Challenge . . .
In the sense of the word you're referring to you are correct., But think of it this way. Imagine a 4 bit word (as opposed to 16 or 24). The "resolution" of the level steps generating the output waveform would be quite ragged. That's the resolution I am referring to. Sorry for being unclear.
Paul, the "level steps" neither exist nor "generate" the output waveform. Bit depth only determines the noise floor and resultant signal to-noise, _it has nothing to do with a real or perceived "higher resolution" or change of detail in the resulting sound._ This is explained and demonstrated from 3:35 to around 10:30 here: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-cIQ9IXSUzuM.html
"Audiophile" special recordings don't sound like they are as challenging as recent hi-res "large" classical recordings - eg Mahler 2nd and 8th and also organ recordings. What is your view re specific advantage using hi res for such music
Then you must have a good DAC and headphones or speakers for the computer if it is you who you are going to use as a listening source, the opinion seems to be that it should be an external DAC not an internal DAC. Then you must also check that the sound setting is correct for the best sound quality
So once you get the sampling rate just high enough that there are no phase alignment issues in the top end what differences do people hear when going beyond that?
A good place to start would be, The Steven Wilson remixes, note not remastered remixed, he has full access to the original tapes, he then makes several mixes, CD quality, a surround sound version an original mix and a 24 96 mix. I can tell you the 24 96 sounds fantastic, your question is 24 44.1 a waste only if you intend to buy it twice, CD 16 44.1 are fine. having said that the best recordings I have are all HIGH REZ SACD, blu ray and so on.
A high end recording will at least be 24bits, 96kHz. If one offers downloads for sale with 24bits, 44.1khz I wonder why they did not offer the full 96kHz. This would make me suspicious if it is real 24bit and not up sampled 16bit. And for me, for listening, CD quality is all I will never need.
Everyone records at 48khz / 24bit we mix and deliver at 24bit - so no one should charge a premium. I want everyone hearing the mix I made, without dither.
We hear the very best in the frequencies of 500 to 4000 kHz . That's probably why Sony made CDs with super bit mapping. That would provide 20 bit quality . 16 bit was considered not enough
Yah, this topic has been covered to death. Why Paul decided to cover it is odd, considering the sheer amount of articles, testing, results, etc., available at the touch of a Google.
Having a greater dynamic range should allow us to here the soft notes more easily while at the same time allowing the loudest notes to be present as well ...shouldn't it?
No. Perhaps this chart will help: en.wikipedia.org/wiki/Hearing_range#/media/File:H%C3%B6rfl%C3%A4che.svg 2 limits not shown: quiet room background noise: 25-30dBA and hearing damage 85dBA. Unless listening to music in an isolated rural cellar or noise cancelling headphones, faintest musical passages unmasked by ambient noise cannot be heard without risking hearing damage from uncompressed peaks.
Considering that all modern music is mastered at the “loudness war” level, 24 bit won’t make a difference. With an old school master of a classical recording you could tell the difference.
An example the highest quality you might buy an album with Lauren Daigle is 24 bit and 44.1 KHZ In technical terms then She Is of course on vinyl too and apparently also on MQA
Paul, I am a fan of your RU-vid videos, but I have to disagree with you on the benefits of 24 bit. Years ago I burned the same song on a CD at 24bit/96KHz, 16/44.1, 320Kbps mp3 and 160Kbps mp3. I first listened to the 24bit recording. There was a huge sound stage that extended beyond my room. However, that sound stage collapsed to a much smaller sound stage when I switched to the 16bit recording and got worse and worse as I move to the mp3 files. My wife heard the same thing and she is not a critical listener.
Snake oil alert...I hear a great difference like day and night between 24/16 bit...at producing my own music, most open, natural, spacious, less hard cold "grey-white" highs, more dynamic. That is the most in the DAW before export. The 16 bit mixdown is always somewhat disappointing and sounds "flat" and more "hard". Here comes the snake oil alert...I did spend LOTS of time in the past building my own speakers and did some ~ 50 changes in the 2 way passive cossover alone...the most eye opener was to use a "tin foil condenser" for the tweeter which sounds much more detailed than any other foil type, all other do some colouring/muddy stuff in the upper mids/lower highs...the tin foil is 10 times thicker, may be that`s the trick.... build my own amplifier which works in full Class A in the lower power range ( I never listen loud...) with relative less negative feedback...and no Class B crossover distortion...spent weeks to try ~ 20 input capacitors and ended with a russian teflon military type parallel with a styroflex... ...got some older cheap Yamaha Consumer Stereo Amp one day, and first thought this thing has a problem, cause the hard, grey, flat unpleasant sound... did some tests and at low power there was an up to 19th + harmonics distortion fence at ~ 0.15% THD ...as a general problem in class B, the lower the power the higher the distortion...the low THD is at FULL power. So...If anybody can hear a difference between 16 and 24 bit depends largely on the used speakers and the analog hardware/amplifier, which does a "lot" of distortion and colouring/muddy things that outweighs the bit difference and sound somewhat similar..."hard" and "flat". The consumer stereo amps I had, one was an older Luxman, did sound grey, flat and muddy as hell in comparison. It`s impossible to hear a 16/24 bit difference with those... And to sum up further, If I would not have changed the speaker capacitors first, I probably could have not not heard the changes in coupling capacitors at all or just not significant enough etc....
@@nicksterj that's ~ right If you have the "Endprodukt" but Not by producing/ mixing in 16 bit with dozens of plugins, eq, compressors, reverb, etc each one computing and each one producing quantizise noise, by keeping headroom far from 0 db and clipping...you work at lower Levels and add dozens of quantizise Errors >> grey/White hard, muddy...grainy...
Apple Music’s Spatial Audio is the future . After hearing it SACD disappoints. And it’s not about bit and bite quantity. It’s about Dolby ATMOS object based mixing.
and ATMOS (like Apple's AAC+) is lossy (the bit rate being 768kbps at most, sometimes less, ie: half that of CD), and lossy degradation is easily cumulative (especially via bluetooth systems or broadcast). Objectively speaking.
Maybe there were comports of the recording that were 44.1kHz, that is to say maybe a digital guitar effect pedal or a synth or even a loop running in the background. That’s before the ProTools plug-ins which only run at 44.1 or 48kHz which may have been used on significant parts or the whole mix. I suspect the recording engineer (assuming not a partial or wholly recorded at home) may have decided to use lowest common denominator rather than up-sampling. If a partial or whole home recording, I believe from memory that GarageBand, or certainly early versions are / were limited to 44.1kHz at 16 or 24 bit. I have had similar thoughts about 24/44.1. Hope this waffle makes sense! Happy to clarify if I haven’t made any sense.
Those master tapes for classic albums have nothing that requires 24 bits. If people are hearing anything at all it is either because the 24 bit version was mastered differently (could be a good or bad thing) or it is their imagination. I find with audiophiles the imagination runs wild.
Digital bit depth and tape noise floor are not stricy equivalent. I used to do tape to digital transfers and 24 bit was handily better than 16 bit, even though the tape has a noise from of -65db. The digital equivalent would be 13 bit or less, and if you've ever used low bit digital you know there is no way you can transparently record anything, including tape. Apples and oranges.
Correct. Even the highest-spec Studer running at 30ips of top$ formula through pristine companding/NR _still can't beat 13 bits of (mathematically sound, pun intended) dynamic range._ A typeIV cassette in a Nakamichi Dragon might approach 9 bits of S/N on it's best days!
16bits are 65536 «steps» in dynamic range. 24bit is 16777216. On paper that is a really really big improvement. I know that in reality on a finished record I know that its not that big of a deal. Just feel like you are telling people that 16 bits is 2/3 of the theoretical qality as 24. I am sorry for bringing this up.
I find it's hit and miss on those. Suspect from 16bits. However my HDCD at 20bit do sound better as they are noticeably more dynamic. Also made to be on the disk.
Dynamic range isn't the only reason for a higher bit rate. Lower quantization noise is probably more important. I've never heard of 44.1Khz 24bit. Isn't it typically 96Khz 24bit? I definitely can hear a difference between that and CD quality.
There are those who believe that DSD 64 is not so good that DSD 128 and DSD 256 are better. The man on the RU-vid channel Anadialog is, among other things, of that opinion . Note SACD supports maximum DSD 64 as far as I understand
Human hearing only goes to around 20kHz at best (more like 15-18kHz for most people), so anything above 16/44.1 or 16/48 should be plenty for most people to enjoy. 24/96 and 24/192 are pointless, unless all you want to do is argue with other audiophiles (who probably can't really tell the difference between a 16/44.1 CD quality FLAC and a 24/192 FLAC anyway. 😅
Ive seen this kind of comment a lot and I think its clear that a lot of people misunderstand what sample rate actually is. So sample rate has less to do with the frequencies the recording is capturing and more to do with how often it is capturing the sound frequencies. A sample rat of 92k doesn’t mean that there are frequencies being captured above 20k, for example, it just means that it is capturing all the frequencies of the recording at a rate of 92,000 cycles per second. Meaning that more information is being recorded or played back. Thats why it uses so much bandwidth, because there is more information therefore more fidelity to the sound. Thats why he’s saying that you can analyze phase issues better at a higher sample rate. Whether the average listener can hear a difference, well obviously thats being discussed.
@@Pinko_Band I've been mastering amd mixing my own music for years and have a very well trained ear.. even i could not tell the difference.. and even if I could hear a barely significant difference it would not take away a thing from my listening pleasure as fussy as I am.. And I have come across people who, tested blindfold, couldn't tell the difference between two masters they thought they were ''very different''
@@milanforever7014 Agreed. Im prone to believing the anecdotes about ppl giving themselves blind tests. There’s gonna be folks who say they can and folks that say they can’t tell a difference. How the test is conducted is obviously pretty important so that theres no immediate bias being injected into it. I dont give ppl a hard time for swingin either way, Im just trying to clear the air for a lot of folks who seem to have a misunderstanding about what sample rate does, is, and means. Like the phase alignment in the higher frequencies thing, if he says that this is an actual measurable phenomenon then Id tend to see why
"High resolution formats" have wider frequency response and dynamic range than standard CD, but unfortunately most people can't hear the sound in the range. lol That's why you can't tell the differences. If you can hear the differences, maybe because the music was remastered for HR formats or you strongly believe HR is better than CD.
Question: I've been, as almost a habit, applying low pass EQ around 19k-20k (and high pass around 25) to nearly *every* individual track I record (in my bedroom) because I "heard" somewhere that no one will ever hear that anyway so you might as well save the space on your files, etc. At the same time, I record in 24bit at 48k. Am I doing something detrimental for the three people that listen to my songs?
Definitely not something I would recommend doing. When you low pass in that range you're causing audible phase shift in frequencies below that point where we can hear. We make a point of doing the opposite. In our electronics as well as our recordings we want them to go out to at least 50kHz to make sure there is no phase shift or unwanted artifacts below that.
@@Paulmcgowanpsaudio Thank you very much for the reply. I don't really give any thought to phase beyond being slightly aware of dual mic placement. I happily still have a lot to learn about recording.
@@shaft9000 Post (after) my original comment, I've gotten out of the habit of crashing everything south of 20k. Tell me more about this linear-phase shifting. And, I'm only slightly joking. Obv. I can look things up, but how do you put this into use? Also, thanks for the reply.
441Khz? I assume you mean 44,1KHz and that is to low, just by going up the 48Khz sampling frequency you get huge benefit in sound quality. 24 bit also helps but for that you can also use upsampling, I can not hear difference between 24bit recording and 16bit upsampled to 24bit in double blind test, but hear immidiately the diffrence between 16 and 24bit.
The number of bits don't matter just in dynamic range. The Nyquist theorem for DAC assumes that each measurement is exact. Sampling at 24 bits is 256 times as accurate as sampling at 16 bits.
I agree with this from experience, which is shared by other people I know. But it doesn't jive with the theory, and as yet there is no proof that 24 bit changes things above the 16 bit noise floor. But it does. ;)
@@chipsnmydip It seems to me that there is another factor in the number of bits sampled. Suppose you are sampling at 16 bits. Then sounds that are 8 bits down from the loudest sounds only have 8 bits of resolution themselves, making the lower sounds effectively having an 8-bit sample. Sampling at 24 bits means that these low sounds have a 16-bit sample.
@@judmcc That is true, but they would argue that you don't need that much dynamic range, esp when the noise floor of microphones etc is often not that low. I have my own theory that bit depth may impact the precision of the DAC's digital filter, but nobody knows for sure.
@@judmcc I do think you are on to something, because when I generally hear PCM digital resolving quiet sounds less well than DSD or analog. Intuitively it would seem like more bits, and you get a more complete reproduction of lower level information like reverb, textures, etc. But as far as I know, the text books say that there is no difference at typical listening levels.
How good are 16 bits and 24 bits respectively at reproducing weak sound details. perhaps that is the important question. And by that I mean not only in theory but in reality
I disagree with you. There could be a big difference between a CD player from 1990 and 1992 when it came to audio details and I mean CD players that cost the same amount. The number of bits is also about how many sound levels there are . Then the question is how does Flac affect the sound and affect the internet itself the sound
Mostly agree, but I do not think that bit depth is about dynamic range. Bit depth is the horizontal resolution and sampling rate is the vertical resolution of a sound. So with greater bit depth you gain simply more precise horizontal values of the input sound when sampling. This is the important part. However, 16 bits is usually high enough to not to really hear the difference when compared to higher bit depths. What this industry is hugely lacking is proper blind tests of all these myths and misconceptions. Probably nobody is brave enough to do it, and/or not properly educated for such a task.
@@nicksterj Obviously they do not have the same resolution. But a well done final 16 bit product might be indistinguishable from 24. However, when mixing, or processing audio in a consumer product, 24 bit is a huge advantage.
@@nicksterj What I do not understand is the feeling I get from people, kind of arguing against 24 bit. As if it matters, and the world would be such a better place without it. 24 is better. this is a physical fact. If you can have it, have it. This 16 vs 24 discussion is meaningless. Nowadays such amount of extra data is nothing.
@@nicksterj Well mathematically that is wrong, as it provides more precise "description" of the audio signal than 16 bits. So it sounds better, but humans probably can not hear the difference if both are well mastered. But perhaps a few people could... This is just maths.
@@nicksterj Sir, maths can not be wrong. You do not understand that article, or the theory is wrong. This is not even complicated. Your argument would be only correct, if 8 bits would be just as perfect as 16 or 24. But you do not think that either.
@@nicksterj Ok, I'll stop here, this discussion does not make sense anymore. 24 bit is better, it is a scientific fact. It is completely meaningless to argue against this fact, or any fact for that matter. Most if not all people can not hear if it is about a master, yes. But 24 is better regardless. This is all.