What Sample Rate Should You Record And Mix At? - RecordingRevolution.com
HTML-код
- Опубликовано: 28 авг 2024
- ►► Create radio-worthy songs from your bedroom. Download my FREE Radio Ready Guide and learn my 6 step process → RadioReadyGuide...
What is the best sample rate to record and mix your songs at? Is it 44.1k? 48k? 192k?!?! Which sounds better?
Here's another question: does sample rate even matter?
I want to break these questions down for you once and for all in today's video so you can make a smart decision and move on with your musical life!
A record producer of some reputation once told me; " If it doesnt sound good enough at 16 bit, 44.1 k; write a better song." He was exactly right IMO.
Words to live by
@@alanjamesh.zamorano1677 Thanks!
Excellent point. Thanks for sharing that.
This is true really, just in some cases you get some ugly aliasing with lower sample rate and added harmonics. Oversampling the plugins is what we need, 48k is enough if plugins do not create aliasing.
Never truer word said.
What these 70 year old ears have learned in the studio and in life:
Techno nerds and audiophiles love to talk about sample rates, resolution etc. But the vast majority of music consumers really have no interest in that stuff; they’re mostly interested in the music. I doubt very seriously that a typical music lover listens to a track and thinks “wow, that reverb really sits nicely behind the lead vocalist and the guitars are panned just right”! Hell, they are going to take your perfect mix and trash the hell out of it by cranking the bass up and who knows what. Everyone has their own idea of what the proper eq is and everyone's speakers or headphones (or earbuds) are different. Most people listen to music in their cars and the majority play MP3s which is a lousy format anyway. I have a customer who puts all of his stereo tracks through a THX surround decoder. Obviously not what the mix engineer (nor the artist) had in mind and what comes out bears no resemblance to the original mix. And that’s OK. Once someone buys a track it’s theirs to do whatever they want. And if you want to use the old 20-20,000 standard, well, people don’t hear that range anymore since, in our modern society, most folks ears have been damaged by the ambient noise that surrounds us. Is it important to record at high sample rates just so your mixing gets less destructive? Maybe. But if you submit files that need to be down sampled you have no control over what samples are lost and any time you down convert you introduce noise and distortion.
What is most important is what the music says and how you have served it. I have a lot of albums on vinyl from the ‘60’s and you can say a lot of it is poorly mixed (not to mention the severe compression involved just to get it to track correctly) but the music is still enjoyable.
Now, I’m not saying that one shouldn’t take care in trying to make the song sound great but that I think (and this is just the opinion of an old man) a lot of mix engineers spend too much time on making the track sonically perfect and not enough on what the song and the artist is trying to get across.
Again, just the musings of an old man.
I think you nailed it.
We stress over things that most listeners will never notice,
not even other producers/engineers.
Agreed. Get it to sound good and keep refining your craft and forget about sample rate, bit crushing and delivery medium. spend more time learning mic placement and eq and compression.Learn to understand gain staging, probably the least understood of all processes of amateur mixes.
Al Dawson I couldn’t have said this any better. Thank you for your wisdom.
Well said Al
Exactly my point. The only ppl worried about sample rates are audio fools. Ive heard mixes done at 44.1khz 16 bits ...and it sounds really good compared with 192khz 30 bit float and sounds like crap. Consumers dont care about sample rates or bits ...they just wanna play it and if they find sounds weird they reach for the equalizers and tone controls ...thats why radio companies put them there from the beginning . On the other hand Ive seen ppl that demand watching a movie at 7.1 on their Bose home cinema systems ....but the speakers are wrong placed or not even connected right. Others just gather all speakers in one corner. They associate higher numbers with quality . Best example....they go crazy about a 55" HDTV ignoring the specs. in the box that reads 720p resolution. Some consumers dont even know whats Bass treble and high frequencies...do you think they know whats a sample rate? or even care about it...? I dont think so.
44.1 khz recording, mixing, exporting
88.2/24 recording and 44.1/16 for delivery
@@joeblough3781 WRONG !
I have been recording music since the late 70s.
I have used every medium and tried every method.
Your words are 100% true.
BTW,
you speak in straight forward vigorous English.
You would make a great teacher.
Thanks you both, agree👍
I prefer recording and mastering at 48K, mainly because every song is ready for a music video at that point (easy to sync). Then when I'm uploading for digital distribution, I'll render down to 44.1.
You make more sense than the guy in the video.
is that still 16bit? or 24
Sync because 48k?
Why can’t you use 44.1k in the video ?
When uploading for digital distribution, do you also have to render down for 24 bit, at 44.1 khz? Or is it only for 16 or 24 bit at 48 khz? Thanks...SBN RESONATE
Great video, Graham! I prefer to record and mix my songs at 48 kHz/24 bit. The only reason for this is that I wanted a higher sample rate than the standard 44.1 kHz assuming this would give me a better sound quality when I started off. And I somehow got used to it and never questioned it again 😅
me too😅
I had to use 24/48 just because 44.1 is not supported by Creative SB Audigy 2ZS ASIO driver. Only 48 and 96 KHz.
Hey there Graham, you nailed the answer for whoever it was that wanted to know what's the best sample rate to record and mix your music in. I say this because I've tried all sample rates in my little 5 year's of mixing and discovered that it's really not about the sample rate how your music quality is. I found out in my time of recording and mixing that if you want high quality sound, to record it first with the best musicians you can find, or self, then balance everything, EQ, Compression and not a lot of plugins and you got it. Thank you Graham for this video.
Graham, you just took me to school, I was mislead into thinking I HAD to mix and record at 192, and my pc choked all the time, to the point in which I thought about giving up on the whole thing because I couldn´t afford a better PC. Thanks a lot, man! Greetings from Argentina!
As a note for sample rate.... The Nyquist shannon sampling therom states that:
A bandlimited continuous-time signal can be sampled and perfectly reconstructed from its samples if the waveform is sampled over twice as fast as it's highest frequency component.
The highest frequency we record is 20kHz so a sampling rate of 40 Khz is all that is necessary to recreate the original waveform PERFECTLY - that is PERFECTLY. NASA probes - supercomputers - Data Centers - and all other sorts of mission critical data processes rely on the Nyquist therm for their processing. If you think you can hear a difference with your ears and NASA can not detect it then I would wager its all in your head. According the theory, you can sample at 2x recreate the wave - then re sample at 4x and recreate the wave and the results will be exactly the same. Audiophiles have been cheated out of money for decades on things that don't make any difference at all.
When Graham is talking about finer steps, he really means bit depth. 16, 24, or 32 bits. (these "steps" are smoothed out any way by a capacitor - resistor circuit when converted back into analog) Standard CD quality is 16 bits and blind sound tests reveal that no one can tell the difference between 16 bits and anything higher. However.... during recording or processing, if you levels are too low, you loose bits. If you record at half volume, you have just lost one bit. Now you have only a 15 bit recording. It gets worse through every plugin you don't set right. If you record at proper levels, then convert to 24 or 32 bits (a lot of daws do this automatically if you choose that feature in options) the processing will have a minimal impact and you will loose almost nothing when converted back to 16 bits. If you record in 24 bits to start with, you don't have to worry about your recording levels much but if you are relatively adept at setting your levels, 16 bit should be fine.
One other argument is that some current distributors and more in the future will want files submitted to them at higher sample rates and bit depths. However.... Its really easy to convert up to their standards and the human ear can't tell the difference anyway so it's not a big deal.
@Joel Schlecht Thank you sir/mam.This helped a lot.🙌🙏
I love your reasoning. The same argument can be made for guitar stomps and processors. The player has to know how to play the instrument before they can begin playing with the stomps and processors. To your point, once a songwriter has learned how to craft a good song with a solid arrangement and melody that can cling to the listener's ears, the mixing process faux pas won't be as glaringly obvious. Too often we want to focus on the after thought tools rather than developing the core skills that the sin of neglecting the acquisition of these core skills cannot be covered or compensated by anything in the mixing or mastering stages.
I’m a first time viewer, and you handled the topic very nicely - subbed.
That was CRAZY COOL information. Thank you Graham for making things a little easier for this newbie to recording and mixing. Glad to be in your VIP group, and great reminder to revisit the 6 Steps PDF. You're the (music) man!
Very helpful....guys like you have made RUclips such a nice play to get answers to almost anything. My best wishes.....I am Diptiman Chatterjee from India. :-)
480p video in focus looks way better than 1080p out of focus :D ;)
My man. Firstly, you are super brave to put yourself in front of the camera and teach people what they want to know. Bravo!
Secondly, as a video content producer, let me help you with the autofocusing. The autofocus on this camera struggled a lot to stay with you, and sometimes went for the computer in the background. This is really distracting for anyone just trying to get the info.
Since you are standing in the same place, you don't need the camera to be set to autofocus. For this type of stuff, I would set the camera to manual focus, place a light stand or tall object in your place, set the focus, tape the floor, and then don't move front to back while recording. And, if you go that far, you might as well set the lens to 4-5.6 f-stop to make sure your whole head is in focus, not just your nose, back of head, etc.
Thanks for the good explanation on 44.1 vs 48 and everything else.
The analogy with a painter is perfect Graham! I agree with you that the sample rate is not responsible for a good or bad mix. Anyway I record and mix at 48 kHz just to have that placebo effect and because my computer still can process everything at 48 kHz. I tried once to mix at 96 kHz, but it was super CPU demanding.. I really wonder what kind of beast is needed to work at 192 kHz.
It's like tetris...
When you double the height of the screen, you'll have a better view of where to place your blocks. Then, when you are done placing your blocks, you return to the standard screen size and distribute your product. Same for increasing sample rates in audio.
Say you mix at 44.1, stretch or pitch your audio, master, and release it as 44.1. Your DAW has only one reference point per half-cycle of each frequency to bring about stretching and pitch change. So your results will remain the same. But if you "double" your sample rate to 96, your DAW now has two reference points per half-cycle of each frequency when stretching or pitching audio, resulting in a cleaner stretch or pitch. Now, when converting down from 96 to 44.1, you are removing that second point (or deconstructing the audio) on a cleaner pitch or stretch change. This will leave a scar, but a good one!
Thank you,, I'm a super new GarageBand user, and I just bought a new 2/2 Scarlett 3rd generation so now I completely understand..!! Thanks bro
My problem with mixing in 44.1 or 48 is the foldback distortion from my favorite analog-modeled plugins and saturation plugins that have no anti-aliasing or oversampling built into it (i.e. Waves, Brainworx, Soundtoys, etc.). I often find that my mixes sound pretty good, but really lack clarity in the upper-mid and high frequencies due to the foldback at Nyquist of all the harmonics of the higher frequencies I boost on vocals, cymbals, etc.
What I've done lately is take my recordings in 48/24, mix at 48/24, upconvert my Cubase session to 96 once the mix is approved by the client (if my computer can handle the CPU load from the plugins), and bounce back to 44/24 and 44/16. It works pretty well, clears the nastiness right up. Cubase, like pretty much all DAWs, has a great anti-aliasing filter in their export algorithm, so any random harmonics above 22k get filtered out perfectly. It also automatically resamples the individual audio files in the session when you change the project sample rate, which is a huge time saver.
Thanks for sharing your expert opinions!
That is very true. And it's pretty much the only problem people will have when mixing at lower sample rates. One thing Graham says in this video is a common misunderstanding of the way conversion works and is very false. That is, if you record with a higher sample rate, you have a higher quality sound because you have more samples and the digital steps are smaller. This is very very wrong !
Actually in Nyquist theory, if you sample at double of the highest frequency you want to record, you obtain the EXACT thing you had before. Now that might not be quite perfect in the real world when actually doing it...
Also, the digital waveform isn't made of steps...there are many interpollation algorithms that predicts where the next sample will be and "smooths" the staircase.
I am not saying we should record at 44.1 KHz, I just say that is more complex than what many think, and some people will tell you oversampling is pointless and others that 44.1 is killing the audio. Wether sampling rate is noticeable or not, it depends on who you ask.
Marcos Brian Klatt A digital waveform is made of steps--that's what the sample rate is. An interpolation algorithm is a layer of processing, but the original steps are still there.
Just making sure, have you ever run a sine wave through an analog plug-in at 44.1 kHz project sample rate and then at 96 and listened to the difference?
At the lower rate, they'll introduce foldback frequencies - and not just in the highs, also in the mids and lows. While silent (starting at -80 dBFS) on their own, if you have multiple emulations running, the foldback harmonics may become something noticeable. A staggering number of gear will avoid foldback, but gear emulation plug-ins often come without oversampling.
In my experience: *Just* recording at 44.1 kHz? No problem. Do it.
Running recorded data through analog plug-ins at 44.1 kHz? Not without an oversampling function.
Even if you export at 96 and let's say for argument's sake you *must* resample to 44.1 kHz. This doesn't mean that your work at 96 kHz was futile. On the contrary, while the high frequencies will be cut off according to Nyquist, but what about the foldback harmonics that may or may not be sullying your mix? Say you have a 10 kHz sine. Then the foldback will be at ~14 kHz, 4 kHz, ~2 kHz, 500 Hz, 100 Hz.
Every analog harmonic modelling plugin without oversampling *will* introduce aliasing, i.e. foldback.
If you know all this and are still content with all your analog work at 44.1 kHz, I won't stop you and more power to you.
@@bonchbonch A digital waveform is not a staircase. It is a collection of points from wich a "line" can be mathematically constructed, also between the points but without the details over the Nyquist frequency.
@@mankepoot9440 Those points are obviously the steps I was referring to. I didn't use the term "staircase." I don't know what you think you're correcting.
@ No, I wasn't describing sample and hold. I was describing the discrete values or steps of a digital waveform which differ from a continuous signal.
I found more benefit in tracking individual tracks at 24 as opposed to 16 bit in my early digital days 20+ years ago. It's noticed mostly when sent to effects or if you have to mix the track at lower than optimal levels to fit correctly in the mix. There is still enough detail to avoid getting that "grainy" low bit count sound. Once the tracks are mixed and mastered in the 24 bit (or greater now) realm, dithering down to a final 16 bit product is virtually indistinguishable from the original....especially for streaming and casual purposes.
I have also found no substantial benefit to higher sample rate recording. Better recording techniques, better recording space, better instruments, quality mics and pres, better musicians, a better arrangement and a better song make ALL the difference.
Hi Graham, most of what you have said is incorrect with regard to sample rates (there are no stair steps). As long as the sample rate is above Nyquist, the input is the only possible output. When converted back to analogue (DAC), the waveform is continuous (smooth), like it was when going in.
I was going to buy a 192khz interface, but by the sounds of it, it's not worth it. Cheers for the info, bud!
every interface is 192 lmao
@@silhz9201 not every interface is 192khz , you should research a little on audio interfaces .
Buy something that has good clean converters and gives you the options you need. RME, Presonus, UA, Apogee... there are others, but that’s where I would start.
Exactly the info I needed! Thanks so much for taking the time to help so many of us who've wondered about this bro!
I think I'll bounce all my individual instrument tracks at 96 sr x 24 bit, and then on the final create two versions! One for the 96 sr x 24 bit and one for todays standards. That way it's already set to go if we ever get to using and streaming those higher def versions. Thanks man!
Liked your comparisons on painting at the end as well!
ive been mixing for over ten years ,you are absolutely right !I always argue that mixing at 44,1 is best.
Shayzar1 then maybe you should check out how plugins work and learn what aliasing is.
I appreciate that it doesn't really matter at the end of the day, and as you say, the mixing skill is far more important. But sample rate for audio and frame rate for video are definitely not interchangeable.
Yes, there are more samples taken, but it doesn't give us a 'smoother' perception of the audio, like frame rate would for video. It seems that way from looking at a stair-stepping wave, but according to the Nyquist theorem, as long as the sample rate is twice that of the audible frequency range (up to 20,000Hz), the audio will be perfectly represented digitally.
Unless you are a dog, sample rates higher than 44,100 Hz (more than double the audible frequency range) will not sound smoother or higher quality. As you say, it's a placebo effect.
Yup...I've been doing recording & mixing for the last 39 years since analog to digital....now I only record & mix @ 44.1Khz, 24 bit......that's good enough. Yes off cause sometimes somebody gives me @48Khz, 24 or even 44.1Khz, 30 bit. I agree 100% on what you have said in this video.....
THANK YOU for clear talk about these issues! I'm trying to get my own home studio setup going, and your channel has been a MASSIVE HELP with all these confusing topics!
yes, Ive also noticed that most of these FREE videos are NOT free they are going to sell you something at the end.... they give you some free advice but its only a marketing video to sell a product that really is NOT free.
Thank you Graham! This was a very good and still simple explanation about sample rates. I appreciate it!
Just sharing.. 24 Bit - 44.1kHz. As basic as possible, but not limiting myself.. It’s a good Standard and is better than CD quality. 🙏
I've been a drummer all my life, playing in tempo is #1 rule for good music!
What does this have to do with sample rates though?
@@Luksenburg00 Same question in my mind...
I totally agree with everything you said. And I am cool with 44.1 for that extra power. Do what you got to do. The mixing and power is the priority. But with a higher sample rate, you are MIXING a more detailed collaboration of files before it gets dumbed down. It isn’t as simple as dumbing down a file that is a same sample rate, the details are getting less detailed. So my question is, does a finished mix of 60 tracks at 44.1k sound different than the same mix of 60 tracks at 96k? I would bet there is a difference, but I would also bet that, if you don’t have the gobs of money to make the difference, that recording at 44.1 is going to make every person on the earth happy. Just maybe not quite as easy of a job. That’s my thoughts at least. Cheers.
I'm new to recording digital audio but I come to it with experience in digital photography where higher resolution and color depth are important if you're going to manipulate the image. Stretching the final gradation in a photo introduces color banding, and enlarging the photo excessively reduces image clarity. No one argues against that. When we apply equalization, compression and other filters to an audio file the digital info gets stretched to the point that artifacts are introduced. If we start with at least double the sample rate that we need in the finished product it can tolerate some stretching. If we start with even more than we need it can tolerate when more manipulation. Weather we have higher than necessary bit depth or sample rates there will be fewer artifacts introduced with filters and affects since we have more data to take samples from instead of relying solely on interpolation to make up for missing data. I think it would make sense to record at the highest depth and rate your hardware can handle, run your filters and affects, and then sample down the file to what you need for distribution.
This is absolutely the BEST explanation that I've been able to find! Thank you for the clarification - while I understood what the technical terms meant, I didn't fully understand the real-world impact/value or how to determine what to choose. Thank you!
My take on this is that, recording @ 24bit/48k then mixing it down at 32bitfloat/48k will definitely resolve the sound problem when uploading your music in social media such as youtube or FB that is known to compress the audio file and distort the sound quality...
Graham. I share your point of view at 90%....BUT...I made some experiences. My conclusion is the following : With reverb plugins, the difference is obvious. Make a trial with something like Vallhalla vintage verb: The top end and the low mid range is totally different between 44.1khz; 96Khz and 192khz. The reverb is simply better at higher sample rate ! More alive, with more air and less cloudy/muddy things in the low mids. My conclusion after this experience was the following: On a minimal project with just for exemple only a guitar and a singer, I prefer to use higher sample rate because of reverb quality increase and because of high quality of reverb needed in this kind of songs ! Better reverb quality = less effort and issues during the mix. So less plugins needed. And with less tracks, High sample becomes possible with a budget computer.
It's the best compromise I found. Thanks for your videos. :p. That's great !
"less cloudy/muddy things in the low mids"
Mh, I find that suspicious.
The higher the sample rate, the more definition..which is actually better for high frequencies and should be less relevant the lower you go.
A sine wave at 40 hz is indistinguishable at 22.050 or 196.000, but since i believe you and trust what you hear, I would ask myself if maybe that particular plugin was designed, in some way, to act differently according to the sample rate of the source.
@@ChristianIce I was suspicious too. This topic is very polemical. I hear everything and the opposite on this point. That's why I tried to make my own opinion. This topic is comparable to "analog summing" topic. Very polemical topic :p. Sample rate and analog summing are the worst polemical topics you find in mixing discussions.
@@romanwinkler3414
Maybe you can solve it up for all of us (My system won't work at 196.000 even if I wanted to :)
If you take a short burst of white noise, and apply the same reverb in a session at 44.1 and in another at 196, then converting the second mixdown at 44.1 and invert the phase, putting them on top of each other and see if there's a perfect phase cancellation.
If there's not, whatever is left is how different the samples were treated.
@@ChristianIce Great idea ! I'll try...
@@romanwinkler3414Do let us know what happens if you go ahead with it. 🤘
The part you're missing is what you just explained is that recording are captured more accurate at higher rates. You don't lose that when downsampling from 88.2 to 44.1 if using a conversion tool like Ozone.
Awesome video Graham ! I'm new to mixing and this demystifies such a seemingly complicated yet simple concept of sample rate! Thanks dude! Keep up the great work 👍
The average listener doesn't know what an audio sample rate is, but it never deters them from loving music.
I use 44.1 mix sounds awesome. I reference with radio/album songs as I mix... thanks to you lol. Preciate the advice. Info is awesome
That's actually a really good point about referencing. What's the point of recording in 96 if we're trying to make our mix sound like a streamed/cd quality of 44.1!?
I didn't think of that.
@@slavesforging5361 yeah man. Saves time too
Higher sample rates are for making better decisions in the mixing and mastering processes for a better end result after mixdown/dithering, not for end user listening. So I thought.
Agreed! It’s like buying expensive studio monitors and a nice pair of cans most people are listening to the music on wired apple earbuds or on their tv’s speakers. The high quality in studio is for detailed work
@@davidsawyer3945 - Yeah, but why not start with high quality marble for the sculpture regardless of how/where the final product will be displayed?
having more "steps" doesn't make it more real. You don't playback digital audio, it always gets converted to analog. I don't say 88 isn't better than 44, but that's NOT why. Great video as always!
Its the same with editing photos for web, converting the image to 144dpi or 300dpi so that you can clean up banding, just to get a smoother crisp image so you can render it back to 72dpi
Coming at this as a words editor, I do know you can kill the original spirit and concept of a piece by over-editing or focusing too much on editing. Editing is essential, but how much to obsess depends a lot on the original. No substitute for having an amazing original piece and keeping the intention of it as intact as you can. The sample rate seems to be a smaller factor in the whole audio process, so I'll take your suggestion, as a beginner (with a good ear), and go with 44.1 to start with. Thanks for the in-depth info, which I found really helpful. Also important to be aware that everything is hidebound by our own ears in the end (and same for every listener).
I say 44.1 foreva! :) Thumbs up for placebo effect. You are so right about it.
John Horn it's not placebo, it's real. You need to get your hearing checked.
@@axelfoley1768 it's not real. Objective tests had proven that human hearing can not reliably distinguish this difference. Try a blind test and you'll see for yourself.
@@covoxer You are so full of crap. Rubert's Neve's research shows you are completely wrong. It is measurable in brainwaves. ruclips.net/video/SsgQcyXyzss/видео.html&t=6221
I record and mix at 48Khz/24bit
Mr.Graham you have been an eye opener for many people including myself.Thanks a lot.
I have a file from an award winning mix engineer. I have the track stems to mix against. I have mixed the song inn 44.1 and with almost all standard/free plugins and my mix sounds just as good and even better to some listeners when they are given the two to compare. I focus on the "big wins" and the tracks were recorded and performed very well.
I record (I call it Capture) at 96k. With my DAW, like most DAWs, this is a setting default. I have plenty of computing power (CPU speed, RAM, SSD drives etc.) so it is not an issue there. Then when I get a stellar performance from my artist, I've got it captured at 96k. Mix downs and cutting .mp3 & .wav are at the normal 44.1 & 48 for now. But If I get the opportunity to have a project considered for movie, TV, film etc, I've got 96k tracks available to send to the production studio in that realm. The last thing I want to do is find myself saying "dang, that turned out to be an awesome production. I wish I'd captured it at a higher sample rate".
I used to record at 96. But, I’ve bought a UAD interface and 96 uses a lot more processing from UAD, what means that I can load fewer YAD plugins in my session. So, instead of buying more UAD processors, I’ve chosen to record at 44
Thanks man, this cleared up a lot of confusion as I'm getting ready to release single #2 today!! Gots a long way to go.... but grateful we have every resource at our fingers.
Thank you... Ive used 192 and thought it was better because the interface goes to 192. I think it boils down to preference and CPU power. I have many songs in both 44.1 and 192 and I really don't know a difference.
Quite right because any difference is outside of the audio spectrum. You can't hear it.
@@buddhistsympathizer1136 sample rate is not the same as pitch did you watch the video at all?
I notice a major difference. one of my synthesized kickdrums changed very much when I tried to go from 44,1khz to 192khz for the first time.
It just shows you more detail.
And different output devices bring these different details in your track up. So if you mix in higher sample rate, higher detail, it will be easier to notice all those details and make the track sound like you want.
@@tormendor8585 It does show more detail, but the detail you are referring to is LITERALLY OUTSIDE THE RANGE OF HUMAN HEARING - You cannot hear it!
Fine if you wish to argue with science. Make a paper on it and win a Nobel Prize.
If you hear any difference it will not be as a direct result of using a higher sample rate - It will be something else that has happened.
did you even read the damn comment
if this was about video, you would be referring to the frequency of light.
(humans cant see high frequency light like UV)
and i would be referring to framerate
(frequency of light has absolutely nothing to do with that)
I personally love listening to music at a higher sample rate. I'll go out of my way to buy those higher end audio files if they are available.
It's a lot like recording your gameplay at 4K. RUclips compression and streaming kills that quality completely. There is really no reason to record at such a high quality if you are going to those platforms. You can record at 1080 and then scale up and the bitrate would match what you had originally.
It's not so much about what others will get when you post it out. It's more about your personal archive and enjoying that high fidelity.
A lot of platforms are switching to this "HD" music. So who knows. Maybe here real soon the higher sample rates will make it to others.
Well, after years of researching (and implementing) the subject, my recommendation would be to record in 24/48. I have found that to be the optimal "compromise", all things considered.
I use 24/48 too, and I think many others do. Perfect for video, and resampling to 44.1 for CD at the end of the chain, where needed, is not a problem.
Same here. I think about movie editing programs and record everything at 24bit/48k. The MP3’s get output at 44.1 and everything is balanced in the world.
And if possible, record at 32-bit, since if you clip past 0dB you can recover the lost signal
Close, but higher sample rates means less aliasing for some plugins and that doesn’t get “dumbed back down” when you covert the wav file down to 44.1.
Hey Graham, I dispute your "no-one's going to hear it" statement. You won't hear it on streaming services, but anybody with a good sound system should be downloading HiRes audio files to play back locally on their system. This means good quality personal audio players, local server streaming networks, computer /external DAC setups etc. A lot of my collection is 24bit 192KHz. Producers and engineers need to publish good quality music as their default. Sites like HDTracks have a pretty good range of stuff, but it's not the norm. MP3s were great in the days of dial-up modems, but most of us have progressed from there, so the music formats should be staying ahead of public demand. You should be able to get a HiRes copy of anything you like. And for those that are happy with low quality MP3s, I'm sure you can keep on streaming downsampled versions from sites who don't really care about quality audio.
Thank you Graham for this great video from Cali Colombia !.
Thanks alot Graham!! Finally someone who's on the same page as me.😃😃😃😃
Thank you for covering this topic Graham! You explained it very well. Great video!
Thank you Graham. This was very useful.
Graham, I like the video in general. One misconception is that digital audio is stepped - it's not. It's only stored that way as a series of timed samples. When it's reproduced back into analogue, it doesn't step. It just drops (or rises) to complete the next sample point. To all intensive purposes it's the same as the input source (no steps). Whether people can hear a difference between sample rates is entirely another can of worms.
Great video , I do agree with all of you , but 48k for daw and video ready sound like a winner 🏆
Hi friend, I record and mix and everything at 48k because I can't make my hardware record at 44k. I mean, I could in the past but now, for some strange reason, my DAW won't record unless it's 48k. I must also say this only happens on a 15 year old, many times rebuilt PC. I have a much better computer but I am emotionally attatched to this one, so I feel that placebo effect you talk about, when recording on it. I'll do my recordings the same way til my old friend dies. I really don't like my newer computers so...
Greatings from mexico city! Great show!
Higher sample rates doesn’t give you higher quality. 1k 10k 20k sine waves are going to sound exactly the same on both 48k and 96k or any other sample rate. Any difference you might hear has to do with how good your converters are at that sample rate (also the aliasing filter at 44.1k). Do a null test and find out yourself.
As long as you aren't trying to preserve frequencies above 22 kHz (which is beyond human hearing), the math involved means that a 44.1 kHz sample rate can perfectly represent that
Pretty much always at 48k, not a huge difference in CPU usage but then its ready for video or RUclips if need be.
24 bit is a no brainer until 32 becomes a viable solution.
Exact
More mega pixel does not mean you get a better picture.
With higher sampling rate you record way above anyone can hear and where mostly noise is located.
So you need more cpu for nothing. This is mostly marketing.
As already commented: The wave does not need to look nicer.
Audio can get perfectly reconstructed into analog if your sampling rate is a bit more than twice the highest frequency.
44.1k and 48k are perfectly fine.
I’ve been working at the digital domain since 2000, and I have always recorded at 44.1. We have great converters now so things just sound better these days. I use an Apollo system and I have no complaints. Well done.
Apollo Range should be the Standard All time to home studios....They are amazing....But very expensive for a little home studio, thats why focusrite win with those cute red soundcards...:)...Anyway, i own MADI Interfaces Rednet and ANtelope, but i love my Apollo quad Firewire with my Sonnet Firewire card...:)..Cheers..
As Dan Lavry claims, the problem isn't the sample rate. Its how we get there. Listening to an acoustic guitar in a room, we hear and feel the instrument, its subsonics and supersonics. The mic we use to capture that instrument tries to capture as much info as possible. Then the interface chops off any information above the set sample rate. Good quality interfaces do this well. Cheaper interfaces reflect all kinds of sonic garbage back into the audible range. It can sound bad. Double the sample rate, which in reality is only an octave, and leave lots of headroom and it can really help. Like the best always say, "trust your ears."
Thanks so much, I was going crazy trying to answer this question. This makes total sens. Thanks a lot!
I'm totally new to this and you clarified things so much. Thank you!
This video is so on point, you’re highly talented man!
There's an important advantage of recording at higher sample rates. If you are using vst instruments or amp sims, you will have much less latency for the same buffer size.
But if your computer is able to e.g. process your audio with 128 samples buffer size at 96 kHz, it will be able to process the audio with 64 samples buffer size at 48 kHz, because this is the same delay (measured in milliseconds). So you gain nothing and will double CPU load when running at 96 kHz
There are so many great sounding productions from the 90s and early 2000s, that have been recorded on crappy digital equipment (in comparison to today's technology) with 44.1kHz and even only 16bit.
Modern technology and higher sample rates do not replace knowledge and experience. Apparently, many people still think that.
In my opinion for recording you need only 44.1k / 24bit (or 48k for video) in 99.9% of all cases. For final format 16bit.
you are great! Love youre honesty and knowledge put simply and clear! Thanks ! :)
Graham is 100% right about recording/mixing technique mattering more than sample rate. That said, when crunching analog to digital, whether it's audio or photographic data, if you capture at a higher resolution, you get more detail by starting with a higher quality capture, even if you reduce the sample rate for mixing. Strike a sensible balance between storage and processing capacity.
Really nice explanation Graham. The difference in sample rate sound like night and day me. Its hard on the computer but it gives me a piece of mind that I got the best capture of the individual elements of the mix. I am in love with 96K !!!!!
This was such a helpful and comprehensive video. Thank you.
Hey Graham! How are u? Very good video! Very well explained. I would like to ask a question: It would be interesting to record, mix and master the song at 48Khz and use this file in the music video, as this sample rate is better for videos, and then make the conversion to 44.1 Khz to distribute the music. for digital platforms, for example? Or that's wrong? And does this conversion from 48 to 44.1 spoil the audio because it's not multiple numbers? Thank you very much!
Like he said you have to know when to record at higher sample rate for example if you're recording vocals then you shouldn't record at an extremely high rate that's when you'll pick up sound you don't want to be there such as popping noises from your signing or rapping however if you're recording acoustics such as a guitar then a higher sample rate will be better to hear all the little details of the guitar. Keep in mind that each sample rate is used for a different task. Just like you wouldn't use a hand brush when you can use a roller or use a roller when you can use a paint sprayer
When it comes to synthesizer design I start and 3/2 ratio or preferably 88200SR. The quality is sooooo much better but more cpu stringent. A lot of people don't realise even though their DAW is running at 44.1 a lot of their tools are not.
Tidal Masters already offers 24 bit/96KHz streams. The difference is amazing in terms of quality.
There is no doubt that there is more high-end detail in higher sample rate recording. When you do a null test, all of the similar content cancels, and you can hear the higher frequencies from the higher sample rate recording. This is because recording at higher sample rates allows you to capture higher frequencies than recording at lower sample rates (see the Nyquist Theorem). In my opinion, it is always best to record at the highest sample rate that you can manage, because it does allow you to work with higher fidelity audio, and that is never a bad thing.
I do agree that it will not make your mixes sound any better if you are not a seasoned mixer, but for a seasoned mixer, there is no reason NOT to record at the highest sample rate that you and your hardware can handle. It will only make a good mix sound its best with extended high frequency content. Yes, some people won't be able to hear the difference, but some people will. There is also the argument about the psychoacoustic effect of listening to audio that has content outside of the hearing range, but we can save that for another time. :)
Also, imagine this: imagine if you could go back in time and re-record the movie "Back to the Future" in 4k. At the time, most people would not have had the ability to display the movie in 4k in their home. Fast forward to 2019, when many people have TVs and playback devices that can display 4k. Wouldn't it be nice to be able to see many of those old, classic movies in true, native 4k? Even if you prefer the feel of the older video standard, it is never bad to have the option to see the same content at higher resolution. Well, by recording audio in the highest sample rate possible, you are preserving this same option for your audio. No one will ever complain that you gave them the option of hearing the audio at a higher sample rate.
Robert Simpson exactly. Everything you said here seems like common sense to me. It’s not a matter of if we’re capable of streaming higher quality audio. It’s a matter of when. That’s basic Moore’s law. In fact, 5G could easily change streaming standards.
The biggest reason for recording at a higher resolution is the harmonic overtones that are inherent in any acoustic source. While the human ear doesn’t hear those overtones, they still exist and are known scientifically to alter the perceived “color” of the root frequencies. So with that said, only the first few repetitions of the overtone series are still perceived at any functional level and ONLY when unmodified. Classical recordings and maybe some more traditional jazz is really the only music this type of thing applies to as electronically created music will only exist at its native output sample rate.
There is always that argument that the combination of all of these overtones in a mix result in more detail even when down sampled, but when you smash the fuck out of it with compression then dynamic limiting at the mastering phase, it’s a moot point. So yes while both sides are correct to some extent, unless you are recording a friggin orchestra or something that won’t be touched by compression, recording for the assumed medium of release is ALWAYS best practice. Or just record at 1 bit 12mhz DSD so no one will ever listen to it like that. 99% of the recordings made will never be heard by more than a few people, and yes 44.1 is rendered as an exact replica of the original electric signal anyways...
And what about the aliasing? What about the colouring that it is adding to the audio processing?
At higher sample rate, less aliasing in audio processing and more clarity to the mix. Thats what i can hear, at least.
I Have My Speakers at 48k I Swicht to 44k Dammm I heard you So Much Better Now .. 44 is the ONEEEE GUYS
Archive highest sample, export a 44.1 version then mix and master that. Recording high samples shouldn't kill your computer but the mixing and plugin processing will. Processing power will improve 1000x in the next 20 years.
The problem with lower sample rate is aliasing as many plugins use saturation these days, also dynamic processing plugins create harmonics as well. No problem in recording at 44.1, however mixing could be an issue with some sources when mixing if aliasing gets noticeable. The best way to solve it is to use oversampling for the plugins and keep the project sample rate at 48k, however many of them are still without this feature unfortunately. With that being said aliasing is rarely a problem, also using anything higher than 44.1 for a simple playback of the end product is nonsensical.
while you're absolutely right, that no one's mixes "suck" because of the sample rate they choose. it's ALSO true that these days hard drive space is cheap and computers are powerful. And so everyone should probably be working at 96k unless there is a compelling reason (like lack of processing power in their computers) not to.
I do almost everything at 32 float/96k (except when it's specifically for video in which case 48k).
there is NO REASON for 88.2 to exist
Definitely been learning some things here and made my first bulk purchase to start assembling the creative space!👍🏾
Okay Graham, I love your channel, and it could be even better if you get your sample rate theory correct once and for all. Okay...the only real thing you get when you mix at super high sample rates are higher frequency ranges (see Nyquist theorem), and some added distortion caused by ultra sonic frequencies in the lower audible frequencies. So, if you are doing scientific work, or will be slowing down the audio substantially, nobody will hear the difference. The output is not in stair steps, the DA converter converts the digital into a smooth and accurate wave form. If you hear a difference, it's most likely because your I/O interface may be optimized sonically for that rate...a little bit of possible trickery by the manufacturer. By the way, your buddy Dave Pensado and a lot of the other engineers that are well past 40 years old, as good of mixers as they are, can't hear the upper frequencies everybody says are there. Most of the upper frequencies that are there are most likely a type of distortion. Have you watched Monty's video yet?...it doesn't sound like it to me. ruclips.net/video/cIQ9IXSUzuM/видео.html
In the end...it won't be the super high sample rates (88.2, 96, and 192 Khz) that wins...it's the talent and skill (you know this). Warren Huart mixes in 48 Khz, but says they will be moving to higher sample rates because of future proofing. This is necessary IMO for big timers to stay with the super rich boy snobbery. Okay...enough...thanks much for what you do!
Thank you ! loved that video thx xiph.org
I've been posting these two videos for a few years now on multiple sites.... www.xiph.org/video/vid1.shtml?time=678.1 and xiph.org/video/vid2.shtml
With what graham said.. I think it's misleading to compare an Audio's Sample Rate to a Video's Frame rate -the "Bit-Rate" however is a more accurate candidate for comparison, since the Bitrate IS the no. of bits transmitted or received per second. Higher sample rate simply means Higher Frequency Range -so it's pretty inaccurate to compare it to "FPS"
@@markgeuellim9384 bit rate is dependent on sample rate AND bit depth. The more bits used per sample (bit depth - 16,24,32) the higher the bit-rate in the stream. Also, the more samples taken (sample rate - 44.1, 48, 96 thousand samples per second) means higher bit-rate as well. I think it's fair to assume that one "frame" of audio would be one complete sample of the waveform, which would contain all the bits at that moment. Think of a bit as a pixel, and a sample as a frame.
Yep, Monty is THE reference. I saved that video in my YT-account categorized as "VERY important videos" some years ago, not for me (you know, once you know what the world looks like outside the cave you actually KNOW what it looks like) but rather for other people to open their eyes. I'm afraid as much as I like Graham, he is so busy with his...well...business that he doesn't have time to read our comments. So it is up to us to compensate the wrong message.
I love the way FAbfilter, Universal Audio and other manufacturers build in placebo effect oversampling option in their plugins just to drain my cpu and and it does nothing to the quality of the audio signal.. As manufacturers they of course know nothing about antialiasing, they just think its cool to have an oversampling option in some of their plugins ...
I typically record at 48; but I also never know when something is going to be used for video. It never taxes my system with the minor bump up and the mix doesn't sound distinguishable when converted to 44.1.
when recording and processing at higher sample rate and your eq and compressor sound better like you say, will also sound better on your final mix when converting to 44.1 even converted to mp3. When mixing at 44.1 and using lots of pluggins like compressor or maximizer will cause all kind of artifact that will remain there. using oversampling or higher sampling rate will fix the problems.
I've always been told to record at 44.1 but render at 48. Because I do Sync Licensing, for TV Film.
Graham, with all due respect, you're way off on some things. It's an objective fact that a lot of plugins alias and cramp at 44/48. This aliasing and cramping doesn't come back if you mix or master at 88/96 and convert to 44. A lot of plugins oversample at 44/48 and end up using just as much or more cpu than if you used them in an 88/96 session. Plus that oversampling isn't always good. Haven't you read the Ultimate Plugin Analysis thread at Gearslutz? You also end up needing to use more plugins at times working at 44/48 in an attempt to counteract the inherent problems. Computing power is cheaper than ever. All you need to work at 88/96 with a bigger session is something like a Ryzen 1600 provided you use bigger buffers. Even with smaller buffers, you can use external DSP by Waves or UAD to take off the processing power from the cpu. It's the only way to overcome the single core gridlocks that can happen when higher cpu plugins overload one core and cause cpu to run out altogether. If people bought less plugins and gear, they can easily afford an 8700k or 9900k system. A lot of the time more plugins and gear aren't anywhere as important as a cpu upgrade. Finally, the things they use to convert to 16-bit 44 are quite important. Big difference using Pro Tools dither compared to Oxford Limiter, for example.
More people should read this.
Recording at 44.1? I'd do it.
Running plug-ins, pitch-shift, and especially analog modelling plug-ins at 44.1? I'd do it if someone held a hatchet to my head, then I'd up the volume and kill the perpetrator with the dirty foldback frequencies.
I'm currently working in a song with a lot of virtual instruments and just out of curiosity I switched from 48 to 96 and then checked the task manager, I notice that now on 96 the project is consuming less cpu...
Dude wrote an essay. Me dig that. Dude dumb tho.
viciousblissvideos 192 récord will sound good even without plugins 🚀🚀🚀
Can you record at 48 kHz, then upsample to 96 kHz using, say, interpolation, and then do your plugins? Cause I can't seem to find any inexpensive ADC's that sample at anything other than 48 kHz. I already have a mixer, but it is analog, and I just want to be able to convert the stereo signal into digital. Most high quality ADC's are packaged in some fancy audio interface with a lot of bells and whistles, which I already have in my mixer.
Useful and Helpful information as always. Thank you !
Mastering engineer Mandy Parnell
Says it makes a difference.
I think so also but I use UAD for processing so I can record and mix at 96 otherwise my computer would crash.
In the end all sample rates can sound good. Most important is a good song and incredible performances.
24bit / 96k for me. Because I can (have CPU, storage for it). Customers can opt later for “mastered for iTunes” if they choose.
take a look at this--Fabfilter--Samplerates: the higher the better, right?
the best explaining