Inter-Sample Peaks Explained: Why is Your Master Clipping? True Peak vs DAW Meter

[@Btvstudio]

No lie people, David's courses are great. He's a great teacher.

[@everythingNotHere]

Indeed I’ll second that !

[@SWIFTYRNORTH2009]

This explains why my tracks sound good at home on the monitors but like absolute garbage in my car. I've been learning so much from your channel already man

[@gurubuzzzz]

If your mixes sound garbage in your car then you have a garbage room. Nothing to do with peaks.

[@cortical1]

Explained crystal clear without dumbing it down. This is why David is the choice among scientist students.

[@gurubuzzzz]

Sorry to say but there is zero scientific explanation here. This is just rhetoric. "Why does it happen" has not been addressed, only the alleged cause and effect are talked about...

[@cortical1]

@@gurubuzzzz Only the alleged cause and effect are talked about? Maybe you have a better explanation of the cause and effect you'd like to share. Let's hear it. 🤔

[@precisionsoundworksstudio]

I may have comment on this video in the past, but just came back to watch this again. Absolutely fantastic description.

[@cybus96]

yoooo thank you for the knowledge ! got a true peak meter and realised my track was reaching them peaks but i couldn't find anybody who explained it so well and got strait to the point! and i also just have to say i feel like im being taught by an argonium sorcerer from oblivion , love the haircut and much love from canada!!!

[@disciplethepoet]

Youre a man of the people

[@jon0830]

still the best channel for mixing and mastering!!! no BS, good topics. thank you

[@themagicianofsound]

I always master to -1dBTP now, not only because of this analog conversion but also because of the various (sometimes not so clean) conversions made by the hundreds of different online streaming platforms and online stores. Here is my way to test this: if I export a song from my DAW into an m4a compressed file, I clearly see that the m4a version has higher peaks than my original project in my DAW. This may be because of the conversion algorithm. The peaks of an exported m4a are higher than if I export to WAV (which is identical to what I see in my DAW) or mp3 (sometimes 0.1 or 0.2 higher peaks than in my DAW). I found peaks 0.8 dB higher in some cases in m4a exports. It's especially the case with piano peaks with very strong notes. I got my lesson only once and will never repeat that mistake again, under-estimating the consequences of distortion that can happen when online platforms or even artists manipulate the music (ex: to add it into a video while making conversions). When I made that mistake a while ago (a while ago for me means a few months ago hahaha and not years like you), I listened to an amazing piano song that the artist uploaded to her channel, I did the mastering and it was absolutely perfection when I sent it, but there was distortion for all peaks in the uploaded video on KZbin. I downloaded the version on KZbin to my laptop to analyze it and I was clearly seeing that the peak were all going above 0db! I suspect a conversion that the artist did in her process when bringing my WAV version into her video clip. It is the first time it ever happened in hundreds of songs that I mastered or remastered but the piano peaks were very strong. At that time, I was mastering at -3dBTP. Although the artist didn't matter, I did haha... the video got more than 5k views so I was having sleeping problems for a few nights! 😂😂😂 It's the hard way to learn haha! Best Regards from your friend François in Montreal, Canada! 🎼🃏🎶🎵🙏👌😁🌠

[@futurebeats898]

For MP3 export you'll need atleast 1DbFs of headroom. Otherwise The audio downsizing will make it clip. You can check it for yourself in the ozone mastering comp.

[@hotzemusic]

i believe David has a video covering this. or at least touches on it in a video about commercial tracks, when converted to MP3, clip when you put them into a DAW for reference for example.

[@Baphometrix]

I love your videos, David, but I'm surprised you put this one out without mentioning that clipping is happening all over pretty much any master that's even close to "loudness war" targets. Especially since you are a proponent of using clippers/saturators for loudness (such as Saturate by Eventide). I mean, any pre-master that is hitting louder than roughly 13 LUFS integrated probably got to that point by having some degree of either hard clipping or saturation happening on various important tracks and bus points. And most modern brick-wall limiters have a final (hard or semi-hard) clipping stage at the very end, which is often *hidden* (except in some limiters like Elevate), to catch the transient peaks that slip through the earlier time-domain attack phase of the limiter. (Or are enhanced by the "transient enhancement/preserving" part of the algorithm.) And then there's the subject of oversampling modes, which can easily push the internal peaks in the limiter above the ceiling, until that final (hidden) clipper shaves them off at the ceiling again. And there's the fact that so many artists, engineers, and labels will ask for the loudest masters possible and therefore the final masters are still put put out at 0 dBFS sample peak and still sound fine even after lossy transcoding, because the material that's getting clipped is mostly short, high frequency drum transients and everything's leveled and shaped so that the clipping doesn't soften those drum transients too much that it starts sounding "bad". I feel like telling newer bedroom producers who struggle with getting louder masters (because their genre requires it) how the *theory* works is a *good* thing, but it has to be tempered with some "but in the real world, here's what the pros really do" comments (and theory and practical advice) too. It's good to understand where and how and why clipping in a lossy transcoding can be theoretically and potentially "bad", but it's also good to understand how you can minimize the badness and still hit your genre's loudness targets as transparently as possible. Hard clipping in strategic places all over your project--AND at the master bus before you head into your final limiter--can be a vital technique, but it requires a mix that is set up to survive this process as transparently as possible.

[@eranliber5993]

Dear sir - you are a great GREAT teacher. Very good and informative stuff. Thanks for the quality content

[@pascalmerschaudio]

great explained thanks mate....love your videos

[@AA-lq5pu]

You are soo cool! Now I feel I understand this. Thank you! I just subbed.

[@Syklonus]

I always set the limiter ceiling to -1dB. I gave up worrying about volume a long time ago, becasue until volume dials on stereos disappear then it doesn't matter. Sounding good is more important, and 1dB of volume difference means shit.

[@jmons33]

🔥🔥🔥🔥🔥🔥 no more confusion on the subject! Thank you 🙏

[@josephvalo]

Awesome Video! Simple and strait to the point. I cannot wait until you have a new course in the Pro Mix Academy. I really enjoyed and learned alot from the course. I was kind of confused about the different peaks and so forth before this video. Thank you so much🤘🍻

[@sevrra]

Thank you David for covering this !

[@claudius3124]

Another great video, very informative, and clear as usual! Thanks Davidf! Have a great day!!!!

[@MrGanax]

Now it does really make sense to me, thanks David. When you talk about DA conversion that can imply peak distortion, does it still occur with DA converters with High end specifications (or instance SNR above 110 db) ?

[@NinaShreds]

That was awesome, thank you so much!

[@The_Absurdistt]

Always a pleasure David. Thanks.

[@Shotty4HF]

Really well explained. Thank you.

[@MerajTypeBeat]

You explained it so well! Thank you! Had to subscribe

[@mixbustv]

Welcome!

[@J3unG]

Excellent tutorial on the True Peak system. Thanks!

[@Barncore]

A little gem of a vid. You always think of good topics to talk about. Most bedroom producers would never have known about these kinds of little concepts if it wasn't for vids like this. Quick question, when you master, what do you set as the output ceiling? -0.3? -0.5? -1.0? Something else?

[@arieloxford6790]

This is one of the rare cases where analogue can actually be far better than digital. I have just completed a master where the music was very heavy sub bass with a lot of pure analogue synth tones throughout. No matter what I tried, the level was just having to be reduced too far in order to prevent inter sample peaks. It was a good mix that was supplied. It was purely about the spectrum and makeup of the content. The way around the problem was to print through my Focusrite limiter. The D/A conversion on the way out reproduces the True peaks so the signal being fed into the limiter represents the true reconstructed audio. The hardware limiter can then deal with peaks appropriately. Digital is brilliant, but sometimes only analogue hardware can provide a quick easy fix.

[@jurymax99]

Thank you so much, this explained a lot of things and saved a lot of time!!

[@Bonzvy]

If the mix will be on streaming services, which is 99% of the case, you not really need to push the level boundaries (not loudness) as hard as you can. The streaming service will take the level down without asking us 🤷🏽‍♂️ thats why I always set my ceiling between -0.5 and -1db

[@J3unG]

I've never gone past -1.0 dB in any of my mixes over the last 30 years. Yes...even in in the analog tape world, nothing beyond -1.0dB. For final mixes and for limiting. Never more than -1.0dB.

[@michaelmurdoch]

This really helped! Fantastic explanation :)

[@MusaDays808]

Nice One, Big @Dave, You're Both A Great Teacher & Inspiration To Guys Like Me Out There & On Behalf Of Those Guys, I'd Like To Say Thanks 🙏 Keep Going Big Man 👊

[@MusaDays808]

@@mixbustv Yeah, No Doubt Big Man 👊

[@simonepiva6969]

Ohhhh finalmente ho capito. Thank you for this video. Grettings from Italy

[@ZoldarMusic]

Finally I undertand this! Thank you!

[@alexinfanti4252]

Great!!!! Awesome Video! Awesome lessons!!!! Thanks a lot!!!!

[@PaulHunt3]

Great simple explanation - thank you

[@ZakOnTheBeatProductions]

Thanks bro. Now I understand what is true peak 😁

[@your_chel]

Thx David, ur videos its diamonds from YT mines :))

[@VictorMan289]

True peak is more of a safety net than a mastering technique. A 96kHz sample rate /w 24bit depth wav master is more than enough for a daw meter to catch whatever peak your wall of sound has. It is scientifically proven that even a 48kHz sample rate digital recording is sonicly indiscernible from the exact analog copy. Also, no audible transient will have such a dynamic range that it will go +3 or 4 over 0 lufs without being caught by even a crap limiter plugin. Where I think true peak is handy is on jazz, classical music or other music genres that require no distorsion and high dynamic range.

[@deltarizkia3492]

Hey dave do you have video review of plugin Waves Linear Phase Multiband/Waves LinMb ?

[@mixbustv]

No but I show how to use it in my mixing courses

[@jmizzbeatsj-jigga7888]

I order that motu 828 es audio interface from watching your review on pre8 es. Can't wait for it to come. them eve monitors I needs lol

[@jmizzbeatsj-jigga7888]

@@mixbustv ahw maan congrats all I know when its coming from you I know it's the truth that's why I went and bought it.

[@PaulBell88]

Sometimes instead of high pass filtering I will use a notch filter for say, a bass so I can remove the problem areas without cutting everything. I am noticing, however that sometimes after the cut, the peak of the signal is higher. Is this related to what we are talking about here? BTW, I have been racking my brain trying to figure out who you remind me of. I finally remember. You look a bit like Alex Van Halen!

[@williamwilkinson2959]

txs

[@3rett115]

You say you won't hear this 'when listening in your DAW'; but even then, are my monitors not doing a DA conversion? I don't understand. I thought everything we hear, whether it comes from playback on the DAW or a finished 2-track, is going through a DAC. If true, wouldn't you still be able to hear the distortion, even listening to playback from your DAW?

[@mixbustv]

Yes of course it's going through a DAC, but I want to believe that first your monitoring DAC is better than a cellphone dac or a Bluetooth speaker DAC. Besides that, you're listening to a wav file, you are not listening to a compressed file, compressed (as in MP3 or other lossy format) done by who knows what streaming platform. That's why if you're a certified mastering studio for Apple they want you to audition the files thru their codec and with their tools.

[@rocthemasses]

Hi Hi ! Do you have any video explaining how "Bella Kelly´s "Heartbreak Hotel" was mastered beceause the signal shoots well over into the red WITHOUT DISTORTING!

[@mixbustv]

Aside from very little snips on effects, I'm not disclosing anything about her songs for now. At some point in the future maybe

[@rocthemasses]

@@mixbustv Will be waiting to hear and I´ll keep my wild guess on how you did it to myself.

[@BelAnuRa]

nice Video :-) but if you Master i higher loudness over -8 Lufs, Levels from Mastering in the Mix is not honest anymore in showing True Peak, same as different other True Peak Meters ... i tried already a bunch of it and until now the TC Electronic one was the honest of all ... but i am on search for others :-) if you have some tips let me know ... cheers Robbie

[@MNmusic2024]

How to determine the value in the true peak limiter?? Some true peak limiters doesn't have a value like fabfilter l2 and it sets the value to 0 , is that ok?

[@mixbustv]

Yes, but also I don't use true peak

[@gaudinni]

Thank David

[@shellb656]

when we're listening to the mix through headphones or speakers, doesn't that mean that the sound already went through that last DA converter that you're trying to compensate for? it's not like we are physically inside the DAW listening to the mix. so I see no benefit using a true peak meter at -1 or even -0.1. at 0 is even questionable why you would use it

[@mixbustv]

I'm not sure why you think we need to "compensate" for a DA. f course all the RECORDED music goes thru a DA when we listen to it, but the idea that ISP could be a problem has nothing to do with the DA in our studio, it has to do with low performance DA's in peoples phones, BT speakers, car stereos, Laptops, etc.. consumer media. Those could, and are, more sensitive and less tolerant so ISP could,, on paper, be an issue. In reality, they're not for the reasons explained in the video and in the follow up video (which is more recent) kzbin.info/www/bejne/iaTOf2uZh7Otoassi=jJ4FnaQwfIq61RlE

[@farshiddel7844]

thanks you so much i am understand

[@everythingNotHere]

Sage advice.

[@thimovijfschaft3271]

What's up with ten video ending, but not actually ending because a part of the video will repeat afterwards?

[@azenkwed]

You are just a great fucking guy. thank you.

[@P3tzos]

Very good !!!! And what is the best option? -0,30 db on master or -1 db for the true peak.

[@mixbustv]

kzbin.info/www/bejne/iaTOf2uZh7Otoas

[@izzychago]

Thank you so much

[@MusicMan-tm6ps]

I have analyzed some professional masters. Many of them have true peaks above 0 db (wave cd files Genre Heavy Rock and Metal) but i don't hear any distortion on the monitors or in the car or they are not audible. What do you think about that?

[@jayguwa5050]

Was looking in to the same thing before. Got something sent back that had like truepeak 1db and clipping in daw. I see it alot too

[@MusicMan-tm6ps]

@@jayguwa5050 Slipknot - sloway firth CD Master: True Peak: +0,7 db Mastering: Bob Ludwig. But again, i can't hear audible distortion. I'm curious what david thinks about that.

[@Ionknw]

@@MusicMan-tm6ps i dont know but if i had do guess i would say that there is clipper included in the chain

[@MusicMan-tm6ps]

@@mixbustv Saw the video and sure the Mp3 files are all clipping when converted from loud waves files and I know why that happens. However i'm talking about cd wave masters that go way above 0. I can follow you well because I've been mixing and mastering myself for 10 years now. If i use my go to itb limiters on 3x or 4x oversampling then it is no problem to avoid intersample peaks if I set it to -0.3 or -0,2 db peak and it is still as loud as hell like all other commercial masters. Yes it happens sometimes that some peaks go to 0 or + 0,1 and yeah i don't care as well but not to +1 in true peak.I mean +1 on a wave file? In my opinion, this is just unnecessary. The ff pro l 2 does the job well and also other itb limiters. The other question is what really happens when you rip cd data as a wave? is it really lossless or is there something happening as well? have you ever try this? I m ashamed now because i never tried 😂

[@nightjaronthegate]

@@MusicMan-tm6ps The wav file format is an exact bit-for-bit copy of the recording on the CD, so it is lossless.

[@themagicianofsound]

This is my second comment under this interesting video, but it's a different point: why when I download great hits, none seem to be respecting these rules?

[@mixbustv]

Rules? :D no man, there are no rules other than one in the industry: you gotta be more then whoever was before you and after you in the playlist, period. More, better, louder, stronger, etc etc.. That's the only rule

[@themagicianofsound]

@@mixbustv 😂😂😂😂 haha 😂😂😂 But with the normalization performed by the streaming platforms, does it still makes sense to think that way? I mean when I want a song to pop in a playlist, taking into account the normalization factor, I'll use all tricks EXCEPT make the whole song louder in my DAW, which would have no effect other than decreasing the quality and even will make it play at a lower loudness level in some cases... You may have seen that I do remasterings on my channel and it takes me forever to find uncompressed mixes to start my work from... most of what I download from common platforms is still crushed at 0db and is terribly distorted! For example they should understand that with normalization, lowering the verses by 1db would make choruses to play louder in playlists (assuming the normalization lowers the loudness level everywhere still after that adjustment), I just don't understand the industry don't get it to optimize loudness in the new modern way? I do not work with artists dealing with labels though, they are all independants. Fortunately I am able to make them understand to stop crushing their sound to 0db as they were doing before working with me and when I make them compare their old sound with their new one on a streaming platform like KZbin by putting 2 windows side by side, they now understand what I was talking about and they will never do this agressive extreme limiting again in the future. At least, at my level, I have always been able to convince all the artists I work with to change their way of thinking regarding this. All are independant artists, so I guess it's different with labels maybe?

[@3ATTR1X]

Cheers David, good info, I have always completed a master file using waves L2 at -01. I wonder though what is happening with my DA/AD converter UA2192 as I use it as a clipper looped so output and return input is clipped via Buzz Audio Arc 1.1 which adds saturation when needed. What I notice is as soon as I add a plugin to the new clipped file the peak levels that were cut off return. So I leave clipping as a final stage sub group. Interesting why plugins are adding back in peaks that were cut off. L2 does not add, but the L3 does.

[@3ATTR1X]

@@mixbustv L2 is the final file, just wondering how analog deals to the Inter-Sample Peaks DA to analog gear then AD as most times I am using it to clip (soft/warm) clipping. What I find is with the new tracks that have been recorded back in through the loop for clipping if I put a EQ or compressor plugin insert the plugin automatically adds 4-6db peak back on (so what was at 0db there is now 4-6db in the red, without any increase audible level. The only plugin that does not add peak level is the L2 which I use just to bring the file master level to -01 and dither. The L3 also adds peak level. So at a guess the plugin is making up the missing peaks? UA2192 has that in-built warm clipping. I believe the designer moved on to Burl.

[@3ATTR1X]

@@mixbustv Cheers thanks, hard to explain aye end up taking half an hour lol, and gone off topic ;)

[@emilianolopez5631]

Yoha! First here, thx for the info!

[@jendrikschmidt]

Thanks

[@marcinklimek8065]

So, we master our song at 0db in DAW and then... It's good idea to put Waves WLM Plus with True Peak Limiter as last in the master chain... Right David? :)

[@mixbustv]

Yes if you want to make sure you don't have any ISP, but there are many limiters out there with integrated ISP reading (ozone, Limiter #6)

[@tai08cdt2]

Could somebody explain to me why many songs i download to my PC it has over 0dbTP (eg: Hello - Adele get 6dbTP on DAW meter) ? Any problem about the songs !?

[@380stroker]

Crappy mastering

[@5urg3x]

Good video thank you I now know not to master to 0

[@iosonoio2493]

Are u italian ?

[@gurubuzzzz]

I could play 99.9% of people a clipped track and they won't be able to hear it...I call bullshit on this whole "True Peak" religion.

[@mixbustv]

Don't confuse the true peak clipping vs actual audible clipping tho

[@gurubuzzzz]

@@mixbustv You actually said "true peak meter displays the absolute peak of an audio waveform and that if it's clipping it will be heard in the analog realm" I'm not confused at all. I don't think anyone understands what true peak means. Let's imagine what happens to the speaker then the air and then your eardrum during an "Intersample Over" and before that in the circuit that drives the speaker. Does the analog part of the circuit have a square wave? Does the speaker make a square wave? I think you confuse the term true peak as some kind of clipping.... If you look at the movement of a speaker cone you will notice it has inertia due to it's mass, Yes ?? so this "True Peak" we speak of exists between the samples that set voltage that goes through an amp then hits the speaker coil. Between the 2 samples the speaker cone has to decelerate which involves stopping and reversing it's direction....if this so called true peak occurs at all, it is in voltage in the analog realm but is it actually an analog voltage peak or the extension of the cone on the speaker while decelerating into negative velocity. And if this True peak exists, it exists purely due to the limitations of speaker technology or circuit design. Which brings me to the point I was making that an actual clip that is 2 samples in length will also exhibit the same relative motion in the circuit and the speaker and is inaudible to 99.9999% of people and only exist theoretically to 0.0001% of people with a highly active imagination. The only place that I can accept that this may cause audible clipping is in a poorly designed analog circuit after the DA converter.

[@caigner]

Thanks for the really good explanation.