THIS is why engineers tell you NOT to EQ like THIS when mixing
Вставка
- Опубліковано 5 лип 2024
- THIS is why engineers tell you NOT to EQ like THIS when mixing
Many engineers tell beginners not to make big wide eq moves when mixing .. but why?
My name is Paul Third and I am a Scottish youtuber / audio engineer / mixing engineer / audio geek. I mostly cover audio engineering related content ranging from audio plugin shootouts / plugin comparisons (acustica audio plugins, universal audio etc etc) to actual analog vs digital / gear vs plugins plugin tests via access analog and mix analog. I even include ddmf plugindoctor tutorials in my plugin reviews so you can be your very own plugin tester and experiment and understand whats actually going on under the hood. I also discuss digital music distribution from time to time and like to give my viewpoint on online music distributors such as onerpm and distrokid.
All of my audio blind tests involving music production software are conducted in avid pro tools 2021 which is my main daw and I also use HOFA blind test 4U as my blind test software. In terms of my audio interface I record and monitor through my audient id44 and use an audio technica AT2050 for all of my voice overs.
Join and also subscribe to the channel to get access to perks:
/ @paulthird
🌠 HELP ME TO KEEP DOING WHAT IM DOING
/ mixingwedesdays
⬇️ MERCHANDISE
paul-third.teemill.com
🏷️ AFFILIATE LINKS
pluginboutique.com/?a_aid=60a...
www.thomann.de/gb/audient_id1...
PRODUCE LIKE A PRO PODCAST
• Working Audio Tools Po...
🎶 FOLLOW MY DEBUT ALBUM JOURNEY
• MY DEBUT ALBUM JOURNEY
🎤 RELEASING A SONG?
Music distribution playlist: • Playlist
Songwriting playlist: • MIXING SESSION WITH PA...
📖 CHAPTERS
0:00 why have i decided to make this?
0:38 do big eq boosts and cuts really cause that much of a phase shift?
2:51 phase nulls with different eq boosts
4:45 do bigger eq boosts affect saturation and aliasing more?
6:45 plugin alliance hg2 black box harmonic sweeps
8:03 so does this mean you shouldn't make big wide eq moves?
🖱️CONTACT ME
Email: paulthird.taysound@gmail.com
Instagram: / paulthird_uk
Website: www.paulthird.com
And for anybody wondering why people feel the need to class this all as a myth I can take a guess its more centered around stock digital eq's. The 'myth' is that stock digital eq's will cause phase issues the more they are boosted and cut. This is actually partly true because nearly every stock DAW eq won't come with linear phase options (with the exception of logic pro??) HOWEVER this also applies to ANY eq that doesn't have linear phase as an option and that also includes Analog eq's (and that goes for non linear emulations as well)
And for the "boosting big with stock eq's can cause distortion and aliasing".. Its been classed as a myth because others have not understood the correct context behind it. It's not just stock eq's, its every eq that makes big wide boosts going into saturation or distortion. it's just that the sole inclusion of stock eq's fit others narrative better.
All of these things are about context, if you don't take time to understand the context behind the "myth" then it will still sound like a myth to you
Seeeeeemples 🤓
Thank you. Maybe stupid question. Dos all this apply to using low and hi cut filters as well? Like cutting all hi information form kick. Cutting all low info form guitar (for bass) etc. Does all this create a lot of phase issues? If yes - what we can do instead or how to do it better?
FL studio’s does 👀
Which I find funny considering how much people like to crap on FL studio 😂
@@AJOrpheo I use Studio One but FL was next of my choice. Best GUI ever. Sadly not my workflow.
All eq moves that aren't linear phase will result in a phase shift but I wouldn't worry about it. There is no linear phase in analog. It's just a byproduct of eq. Certain phase shifts can actually give you an audible benefit
You're a fantastic teacher, Paul!
🤜🤛
You make these types of critical analysis SO EASY to understand. Thank you for not towing the line and just revealing the truth.
🤜🤛
I feel like when you link up with Dan Worrall the world will start to shift on its axis and turn into what it should always have been
I don't know if that day will ever come tbh haha
the fans demand it. Plus Whytse!
Woooo Paul, I never thought that going into a Saturator with a higher treble eq would produce more aliasing!
I checked it in Plugin Doctor putting the Black Box after and then putting it before the eq no longer generated that excessive Aliasing!
Then in a mix that we are doing we put the saturators first before the eq that boost the highs and it is impressive how the clarity improved!
Thank you very much Paul, it is one of the best videos you have made without a doubt
🤜🤛
@lunairys Very interesting point to put the saturation before the eq and probably also before the compression. After Pauls video that makes sense. I have to test it.
Nice video, Paul. It’s important to remember that linear phase eq is a fairly recent technology and how many superb sounding albums were made in the analog/non-linear world. In other words, don’t get hung up on it - if it sounds good, it is good.
Totally. Definitely not something to over think but nice to know all the same. I've heard many say thats why they use tools like spectre and stuff to eq with saturation rather then eq but in the end I have way more important problems to think about in a mix haha
"Fucked is fucked" That's my philosophy as well. Cheers!
🤓
Good work. This type of content is not only useful but not so ready available. Keep it up.
🤜🤛
It should be a known fact that minimum phase EQs work exactly because of the phase shift.
Technically speaking EQ is a function of phase and not the other way around: you delay a frequency group by a certain amount and this gives you the boost or cut.
The same happens in the analog domain as LCR filters do exactly that, they apply delay to achieve the desired result.
Phase equals delay.
As for the linear phase EQ variety, digital trickery is then applied to recover any phase shift occourring.
Tbh I shouldn't have had to make this video because your right it should be a known fact.
But when you have audio channels with 17k+ subs needlessly confusing matters with stuff like "people believe that digital eq plugins cause problems due to phase anomolies and aliasing. I added a bunch of eq on a track, countered that eq in series with with reversed settings and the result nulled 100% with the unequalised sound.. Proving that stock digital eq boosts don't mess with phase like people are lead to believe"
.. And then using that example in a video talking about why it's OK to use stock digital eqs to make big wide boosts and cuts.. It just blurs the lines of the facts.
Now after looking at his comments section it looks like what he's trying to say is that digital stock eq plugins are 100% accurate unlike analog which may have phase variables and these are what he calls "phase anomolies". But then at the same time states that due to the basic roots of digital audio science ie cancelling a boost out with its exact opposite cut and then flipping the phase with the unprocessed to 100% null.. Proves that "all digital eq's do is boost or cut" and completely disregards the fact that the bigger you boost or cut results in a more audible phase shift and thus why making these big eq moves in digital eq's without a linear phase option can to many, result in negative results or "ruin the mix" in their own words (not mine)
love every video appreciate the in depth look in quick-form content
🤓🤓
This is a good video. It’s important to distinguish the roles of the artist/producer vs an engineer. Either look at processing like EQ differently. A bedroom producer or musician looks at an EQ and thinks “tone shaper” and who’s to say they can’t boost, cut, futz up whatever they want? It’s part of their creative vision. We engineers on the other hand think of EQ, Compression, etc as tools for adjustment or repair. We tell people when mixing, recording, or mastering to be sparse with their processing because our goals as engineers is to create as little distortion (changing the waveform in anyway than the original) as possible. Whether it’s linear or standard phase shift EQ, we’re going to shoot to get away with as little as we can. There are exceptions such as artistic engineers like CLA or Manny M, who get paid to change an artists sound into something mainstream.
Learn your tools so you can understand the consequences and benefits of using them.
Edit: added in a thought and a missed word.
🤜🤛
I was going to mention CLA. There’s videos everywhere of him using huge EQ boosts without any caveats other than if it sounds good, it is, etc. He neglects to remind the viewer he’s been engineering near the top of the industry for decades, using the same console with insane numbers of premium outboard units to correct & tweak his spur of the moment maneuvers. That’s why I refuse to pay to watch guys like him when I get more and more useful info from guys like Paul Third. 👍
Remember that when he does do these big eq moves it's usually into compression as he likes the sound of going just too far with EQ and then taming that back to the sweet spot with compression.
Has its own sound so it's normally not just a random huge splurge with eq, it's pretty well thought out as he gets an extra bit of grit and colour.
@@russellszabadosaka5-pindin849 it’s also important to remember that he’s using analog gear which is way more forgiving than the digital realm. So someone on their laptop boosting 12 db in an EQ is going to get a much harsher result.
Nah man that's an old myth. No reason why analog gear would sound less harsh unless it involved a bit of saturation. Most analog eq's are super clean and would almost null out with a digital eq if the curves & phase match even at extreme settings
Passing signals through an ssl console is still a very clean process
First time i actually followed your analytical lessons on a large screen- way to go!
🤓🤓
One of my favorite stories which I cite often is from an interview with Malcolm Toft talking about the Trident A-Range, and how when an engineer told him about the phase-shift the console's EQ introduced he responded: "Okay, but do you like how it sounds?" A big reason why that console sounded so good (its emulations are my go-to EQs) is because its designers were musicians/recording engineers instead of electrical engineers, so they didn't worry as much about the technical side, and did what sounded good musically, because that's the goal. I concur with you that the science is handy to know, especially if you need to fix a problem, but it is only a means to the end of getting a good sound.
Totally, lots of acustica stuff I like is a lot to do with just phase. I mix more with pre's these days instead of the eq's just for the passthrough
@@PaulThird Indeed, not only do they give you more character, but they're more efficient to work with; why would I spend 5 minutes sweeping through to find that just right frequency and q factor on a digital parametric when I can pick from frequencies set for me by the guy who tracked "Hey Jude," pull it up and in 5 seconds go 'yeah, that's pretty good,' and save my energy for more creative decisions?
Thanks a LOT!. We need more information like that! It really helps to understand how really things work in digital realm and what to do and what to avoid!. Please, make more short informative videos like this. Especially for all analog emulation plugins and how to be careful with them. Because I thought they act like all other plugins but they are not.
👍👍
Thanks Paul, this is a big help in understanding what I've been seeing on my VU meters. 6:15 The buildup of the harmonics creating audible noise. Science ! 🤯 Cheers
🤜🤛
Thanks for the video. Great explanation!
🤜🤛
Simple workaround a very successful producer gave me... don’t touch the eq, just change patch, or sound you use until it works
🤓
What i tend to do is use linear phase on sounds that don't have clear transients and non linear on transient sounds. The reason i grab a linear phase eq is because the sound retains it's energy or presence a little more. Thats at least how i experience it.
🤓🤓
Great vid as always Mr WAT.
😎😎
It is very interessting and realy helpful. Thank you! :-) Is it a solution to do a decrease of frequencies instead of increasing them - and after this to increease the whole level of the signal? Have you done a video about a good way to use EQs or is it not possible to avoid this problematic?
My view is that you give the track exactly what it needs and don't overthink. If it's need cut, cut it. If it needs boosted, boost it and move on.
When I use Parametric EQ 2.. I will drop the main level of the EQ to it's lowest level -18db.. Each band can only boost to +18db..At -18db " The overall level" I can still hear the input source clearly. I will then use my bands & boost and shape the signal anyway I choose to bringing the bands back to -0db.. This never goes over 0db.. I can really make a thin sounding VST sound huge and more rounding like this.
👍👍
Im going to try this , Thanks !
Good video and perfect explanation - I'm curious how this apply to a plugin like Shoothe 2 in terms of alliasing?
Nowhere enough distortion made in the plugin to cause aliasing. Basically like putting a signal through proq3 working as a dynamic eq. Extremely clean. Soothe is meant for smoothing out harshness so it's never anything to worry about
@@PaulThird - thanks so much for your reply 🙏
Good video Paul, love ya pal. That 11db gain from phase shift is huge and would be unexpected by those who don't know. You'd get away with the distortion in analogue if it still sounds ok, but in digital you are going to end up aliasing all over the place lol.
11db increase in difference is just 15db compared to 5db boost haha 10db & 15db boost compared to 1db boost both have over 20db more difference due to phase shifting haha 😅
Makes me wonder about parallel saturation (such as spectre) and (if oversampled), is it a better alternative for wide boosts or top-end boosts? Baphometrix certainly claims it to be so in her Clip to Zero series and the audio examples used were a convincing argument
Pretty sure that's what jaycen joshua does a lot. Uses spectre instead of eq a lot. He definitely does it on vocals
Ho mate totally unrelated but I want to get into acusticaplugins and I saw a little shootout you did between p42 and gold. I want a colourful eq and some saturation options along with some compression. Would you say gold is a good way to go or have you any other recommendations? Cheers
Depends on the exact vibe you want. Describe what you are after and I'll tell you what I think
Real life sound has phase shift with frequency changes. Linear phase is an option for very specific purposes, but using it as the default is less natural sounding.
True, no real life sound is linear phase so technically linear phase is less natural sounding in theory but doesn't mean in a creative context it doesn't have a place like any other process.
Natural is a pretty subjective term. You could use tons of natural phase processing and make this sound unnatural. That's why we just use our ears and go with what sounds best
@@PaulThird Very fair! :)
Excellent Paul as always. What about Natural Phase on Pro Q?
Not looked into it tbh. Is that not what many call analog phase?
Thanks, Paul. There seems to be regular moaning by well-known UA-cam producers that amateurs and hobbyists, like me, never do enough automation when creating tracks or in sound design. I think the situation gets worse during mix-downs. With self-mastering, well, it's just slamming on some plugins, and the job is done. This missing attention to detail is often a lack of critical listening skills. Concentrating on the side effects of EQ moves is even more difficult. This all describes my shortcomings so please keep hammering away at this stuff as one day it might click. 😄
🤜🤛
A fun experiment I thought would be to take a full complete old project mix with 40 to 50 tracks, replace the original audio, vocals and instrument's etc., but keep all the plugins, and volumes but replace the audio with either a traingle wave or even 3 sine waves to form a chord on each track. Dont actually listen to it, but just render two versions, one oversampled, and one with the oversampling off and compare with the Rx. Obviously it will have the same affect as you see comparing a single track but it will show the Cumulative affect on a typical mix. How much garbage will exist?
Even though lin phase and min phase don't null, the "difference" between them is actually just a slight delay - not a change in tone. a static phase shift is almost never going to be audible unless it is blended with the dry signal.
There is a difference in sound. I understand what you are saying but in practice if you match curves but if the phase doesn't match then are audible differences in an A/B test. For example boost 15db at 10k on a drum bus and compare that with the linear phase. I can hear tonal differences in the cymbals where it sounds harsher with minimum phase.
Great video Paul, really opened my eyes.
I wouldnt know what phase shift sounded like though even though I have spent countless hours hammering away with EQ alone. But I'm looking forward to doing some heavy lifting with linear phase EQ and the gentle tweaks with standard EQ to see if anything good comes of it. cheers!
Always worth experimenting with. But just be careful with linear phase and transients especially if doing really narrow boosts. I really only see it as a problem in the low end due to pre ringing so I use mixed phase from time to time but it depends on the source and tbh its not even an issue in a lot of cases but no harm in experimenting yourself 🤓
@@PaulThird Ahh cheers for the further info.
A lot of the time I'll be cutting quite a lot with low shelf
Great video! Well explained and great examples.
🤜🤛
Great video Paul...How ever...
Dinnae be hitting me back with my howevers! 🤣
Been off all screens for 4 days, firmly recommend it. Wouldn't have guessed such a video gets >8k views. Am surprised you didn't mention those kinds of boosts/cuts before hitting a comp in order to make it react harder to let's say pickups or consonants. Same is true for saturators. Rule of thumb: If your comp, saturator or whatever lacks control, you need to compensate via EQing the signal that goes in and either blend in a duplicated track, or "reverse" your EQ move after the additional processing (note: don't reverse it 1 to 1, do it by ear).
Tbh I didn't plan on making it. I was sent a video made by another youtuber and the information was pretty confusing and misleading in my opinion so I decided to make a quick video simply stating the facts so the correct information is out there.
It was purely made to counter the claims that stock digital eq's don't affect phase the more you boost or cut and that big eq boosts in a digital eq don't have any impact on distortion or aliasing.
I think a few people have thought that I'm saying this is why you should use linear phase and not to eq into compressors and saturators.. When it's just simply showing what happens.
I'm gonna make a video showing CLA's big 8k boost into la3 on guitars to show why "rules" are broken and why you just use your ears.
And yes... 8k+ views proves to me that being more educational with my content is where the growth is. Using my tests and doctor to educate is a better use of my niche than simply doing shootouts all the time
@@PaulThird phase is a thing. Sometimes it's a big thing - like when mixing live-recordings, especially of drums. Sometimes it's a rather small thing, like with EQing. You can easily get away without linear phase EQ. But it's just one click away from testing wether it sounds better or worse.. so just click. That's my approach, anyway. Still very surprised that video made >8k. Sometimes I'm at a loss with what people click and usually I think I understand people pretty well. Tells me people are quite insecure about their EQ moves.. and I thought people didn't know how to compress.
Obviously no need for an answer, I'm just rambling about.
Amen to facts mate. Amen to the facts. 🙌💯❤️
🤜🤛
This is why you compress pre-eq. It can be done either way around, but the order changes the initial sound in a drastic way.
Very true. CLA is a great example somebody who purposely eq's big but into compression so he can use the gain reduction to control how much 8k he brings into guitars. He purposely goes in hard into the la3a instead of just boosting to taste as he prefers the colour of the 8k compressed. He goes too far so he can use compression to dial it back. It's a sound in itself and quite easy from a workflow aspect. Max out the 8k.. Have up your compressor set up the same way and compress till the bite in the guitars is just right and not overly harsh
What I did wrong back in the day was just add in lots of 8k without going into a compressor and wondered why it never sounded right.. All I heard was "I crank 8k.." OK.. F*ck compression.. I learned a valuable lesson after that eventually .. All about the context haha
Not that I know what I'm talking about, but back years ago I've read that linear phase introduces eq pre-ringing which is a much audible phenomenon with transients, and it will create a time-based difference when compared with regular eq. So I'm not sure if nulling is a useful test in this instance. Also been told that it's the phase shift itself that creates the desired effect of cut/boost - for whatever it's worth. I guess if it sounds good, it is good, which is what matters at the end of the day.
Phase is a delay of frequencies. That's what happens when we eq. We slightly delay each frequency in accordance with the eq curve. If there were time based issues using linear phase then it would affect the full signal. You can see that the areas that aren't being boosted fully null as any irregularities in time would cause an audible signal to come through..but there is no issue with time using linear phase. That's why the only signal we see coming through in the null is from the area that's been boosted.
As for pre ringing and linear phase, it's one of these boogie man things that people push all the time. Yes linear phase can cause pre ringing but the chances of it happening depend on certain factors and whether its even audible is another thing.
Even in Dan Worrals video he admits that he needed to set it up a certain way with a specific kick in order to make the pre ringing noticeable. And it's important Dan said that cause loads of people make youtube videos and show examples of manufacturered pre ringing but don't tell anyone that they had to spend time to figure out how to make it noticiable. Nearly all of them are pre set up.
In actual practice, its an out and out lie that using linear phase will cause pre ringing EVERY time on transients and that when it does, its noticeable.
Take acustica for example. There's truncation (echos) in loads of their plugins and its something that some people prefer without even knowing and most don't even hear it.
Pre ringing is the same. You can use linear phase on a kick or snare and there be some slight pre ringing on occasion and you don't even hear it in the mix and you prefer it compared to natural phase. Very similar to aliasing.
I'm not saying use linear phase instead of minimum phase. But I do sometimes use mixed phase in kirchhoff which is natural phase for the lows and linear for the rest. Not cause I'm worried about phase, just cause the options there, it has a tonal shift and its easy to compare.
Everything we do in a mix has the potential to f*ck something else up. Phase obviously has an effect on drums and if you eq 3 snares all differently then that's gonna create a tonal shift due to phase cancellation as they are on the same signal, and that's why many eq on a bus and process as a whole.
Its just a part of eq and tbh I just use my ears in the context of the mix. If I like something better then I go for it
@@PaulThird Our ear is used to phase shift created by EQs for more than 40 years. In Germany we have the saying "the better is the enemy of the good." Sometimes linear phase is the better choice.
Hey, couple questions. what do you think would be the best EQ to EQ rap vocals then? should we use a Linear phase EQ?
2. If we don’t track with compression going in, would you compress first or EQ in your chain
3. in what part of the chain would you use a Saturator?
Im talking specifically for mixing home rap vocals. Thank you 🙏
I treat all vocals the same way. Smart eq for spectral balance and normally mixroom from mastering the mix in a genre specific preset to give me an idea and I may tweak that further, add to it with a clean digital eq or simply put up kirchhoff and do my own thing entirely but my advice is more to have a reference handy of what you want your rap vocal to sound like and use whatever tools you have at hand to replicate that.
My vocal chains are extremely clean but I do use a hard hitting 1176 in parallel to give me some grit if I feel I need it.
I compress first then eq dependent on how good the vocal recording is. If it needs work then I'll clean up first then eq but if it's a good recording. Compress clean with kotelnikov GE and then Lindell de-esser, soothe, smart eq followed by mixroom or my own eq moves with kirchhoff then multiband compression at the very end
@@PaulThird Great, thank you🙏 One more thing, do you know which EQ’s are colorless and clean/ natural? ones that won’t add any artifacts etc to my chain?
Most non analog emulations. Mostly all stock digital eq's, fabfilter, waves, kirchhoff etc etc list goes on and on
@@PaulThird Thank you🙏
Great observations; it had me poking around in PluginDoctor on the oscilloscope and it peaked something of curiosity which when I brought it over to the frequency response plot made better sense;
You might find this interesting to further explore (For the science part; practicality part plays very little into making sense of what I found)
At higher frequencies; the differences were indistinguishable when graphed; on even the frequency response graph;
However here were my settings and you can poke around yourself with it (Because YT doesn't allow photo attachments)
IN FF Pro-Q3 a bell curve @53.8hz +3dB Qfactor of 4;
Plotted 3dB of gain in Natural phase mode and ≈2.5dB of gain in Linear phase - Medium mode;
I think the differences may be smoothed out at higher frequencies because the bandwidth covers more frequency cycles;
All that said; good job;
It's been a long day, could you simplify what your end conclusion is by doing your tests. My heads minced. Are you saying that phase diffrences are more audible in lower ranges compared to higher frequency ranges?
@@PaulThird Short answer; there happens to be a difference in GAIN on the EQ's centre frequency at those lower frequencies;
My conlusion as to why; the audibility and relevance; inconclusive because these were very quick tests I put together and glanced over this morning during my coffee;
@@panorama_mastering I have not uderstood your point either.
And fact is a PMC loud speaker model series! 😀
I have a new headphone amp coming... F*ck speakers... That's the truth!!! 🤣🤣
@@PaulThird I luv my PMCs.
Dear Paul, do you have your mkxed/mastered songs available on spotify/youtube?
Would like to hear them.
Just go on my website. Paulthird.com
Great fact based knowledge. Easy to pick up. Every audio engineer can draw his own conclusions. For me it's once more clear how important gain staging is in the digital domain. Not done right it can impact your audio negative and you don't know why it doesn't sound analog although you uses all these analog modeling plugins.
🤓🤓
On a side note, how did you get a white background on the Kirchoff?
It's in the settings 🤓
Cool, I like yours!
So if we cut before an analogue emulation we won't be pushing the level and won't create aliasing. Boost after a linear emulation to avoid aliasing
Maybe a way around aliasing by boosting into an analogue emulation plugin would be to reduce frequencies either side of the frequency you want to boost, bit like a passive EQ.
Depends on how much high order harmonic content the non linear process creates and how high it makes it but yes I'd advise boosting after non linear processing to avoid aliasing for the most part but whether its necissary depends on the size and frequency range of the boost, whether the source itself has enough high frequency information and ultimately the harmonic behaviour of the compressor or saturator
Theres quite a few factors that have to be in place for it to be an issue but it's really just something to bare in mind that if you eq heavily into saturation that area will be saturated more so its whether you want that
So in hardware phase shifts also happen during the drastic eq changes but Aliasing doesn't happen in hardware ?
Correct. Analog shifts the phase but no aliasing as its purely a digital artefact. Even in modern times AD converters simply lob off the high end information past niquist before it reaches the digital domain
@@PaulThird Thank you Paul for the prompt reply.🙏
My dude!! Facts & science is what cripples the BS,,, keep em comin 🤘✌️
🤜🤛
Hey Paul, am I delusional if I think that the Sonnox Eq sounds better than the Pro-Q 3? I noticed when filtering the hiss in electric guitars the Sonnox sounds less masked than the Pro-Q 3, I believe there is a difference...
Eq match the curves (if you are stuck you can use plugin doctor or bertom curve analyser dependent on your daw) ensure both are identical in gain, do a null, if it nulls you are delusional
@@PaulThird 🤣🤣🤣
@@PaulThird I've just tested them, they don't null
Did you properly match the curves in an analyser? Because proq3 uses a different q factor to most eq's. If you have just matched settings, they won't null
@@PaulThird no, but I kinda tried to null them and the settings were totally different but you could still hear some noise, they didn't null perfectly, I should try the plugin doctor, the thing is that the same simple Eq move, a High Cut at around 6.5 K generates different results, so It will affeft the way you mix...
Hey Paul, could you do a video explaining WHY equalizing leads to phasing and aliasing? I know you've talked about it before, and so have both UA-cam mixing channels and magazine articles but fuckin' ... _Why!?_ Why does it happen!? Please explain it to my friend as if my friend had a random attention deficit disorder and a _somewhat_ inconsistent ability to connect the dots 😬
Pretty sure Dan's done videos on phase and in terms of aliasing I've done a video on it already.
Aliasing is caused by higher frequencies exceeding nuquist which is the frequency limit of the sample rate. So at 48khz, the frequency range is limited to 24khz. However what many forget is that harmonics are octaves of the fundamental so even though a frequency may be below niquist, when saturated it makes orders of harmonics in octaves so 10k would have its even harmonic at 20k, then its odd harmonic at 30k and it goes on and on dependent on how much harmonic distortion is generated.
So there are frequencies (harmonics) being made past niquist which in the digital domain can't be computed as the sample rate is set compute 1's & 0's up till the limit of the sample rates frequency range. And these are bounced back to the audible range so the frequencies can be computed.
So knowing that. If you boost tons of high frequency content into saturation then it's gonna make more higher order harmonics which dependent on the amount of saturation and the amount of boost and the sample rate.. Its gonna exceed niquist and bounce back as unnatural frequencies
@@PaulThird Dan who?
Worral
@@PaulThird Oh sweet Jesus, so that's the name of the guy who did that video on aliasing; I've been looking for him forever!
Good video
🤜🤛
I love your videos. But this time, I have to raise my hand. I agree with you on the phase shift issue, of course. Especially, if you have two signals, being at the same pan position, keep in mind that boosting a frequency range of one signal also effects the second signal. If you need to add a lot high-freq content a good, oversampled exciter might be the better option.
But I'm not completely with you regarding the linear EQ-thing going into a nonlinear plugin. Just oversample the nonlinear one and the aliasing issue is solved to a major degree. And the pre-ringing of linear EQs makes them rather un-preferable for me. If you want a particular character from your saturated plugin and it doesn't have a tone controll you actually have to hit it harder with specific frequency ranges (and un-do the boosts with a following EQ). 🤷♂️ By the way: Most nonlinear plugins (like preamps or tape) cause phase shifts (which is very audible if you make a 50% dry/wet mix) so reaching for a linear phase EQ just to go into a nonlinear (waveshaping) plugin that causes a phase shift is kinda...pointless?
I'm not saying to use linear phase eq though haha I'm just showing the facts as there's another video doing the rounds saying that stock digital eq's just cut and boost and have no impact on phase or distortion/aliasing when making big cuts and boosts.
The saturation/aliasing thing is purely just a reminder to think about how your eq move plays out further down the line. For example If you use a coloured compressor or saturator that doesn't oversample and stick a crap ton of high through it then the aliasing is there. You can't oversample after the fact. It's locked in.
Theres ways to fix it but all I'm doing is giving others the context behind it so they at least know it's happening. Just food for thought really. I'm not telling anyone what to do, just giving the facts as that's when you start to understand why something sounds a certain way..
For example.. I boosted 10db on my snare using a clean eq and now I have distortion around that area.. Oh.. I forgot I had a fast 1176 after it.. Just shit like that.
CLA does it but understands why. He boosts a shit ton of 8k on guitars into an la3a so he can use the compressors GR to control the amount of bite he adds to the guitars. He could just add a ton of 8k on the SSL but he wants the grit and the GR to be dealing mainly on that 8k
GR is gain reduction. Phase is just a part of eq. Every EQ that isn't linear phase results in a phase shift. Most analog emulation compressors have a phase shift and essentially anything that changes the frequency response alters the phase as well. It's just a byproduct of mixing. Flipping the phase won't remove a phase shift, only linear phase can do that.. Or unless you get clever but that's just not practical.
In reality, the more you boost or cut results in more of a phase shift. It's up to the user whether that's a negative or it isn't. Just trust your ears.
@@PaulThird Ah, okay. I haven't heard of that stock-EQ myth before.
OS:...Sometimes I forget that Reaper's internal plugin oversampling on demand isn't implemented in every DAW out there. But there's still DDMF's Metaplugin 😀. However, one has to be aware of the necessity to do that, of course. So it's good that you gave people a heads-up to keep that in mind.
I fully don't understand why people think that these phase shifts are so evil (and no I don't need to be told all the same tedious literal reasons). I am saying that the 'sound' of inevitable phase shifting is the sound of EQ.
If phase shifting were so brutally universe-destroying, we would take the body off our guitars, violins, pianos... take them outside even to get rid of phase adjustments from walls, mountains, the audience!!! Definitely not use speaker cabinets to enhance the yumfulness of our metallicators.
I use eq nudges to bring character changes which make the instrument more vibrant exactly as how Mr Stradivarius and Mr Wharfedale tune their sound boxes to make the sound fabbo.
:-)
It's not evil. Just a fact of eq. For whatever reason somebody thought it was a good Idea to make a video stating that digital eq's just cut and boost and have no impact on phase regardless how much you cut or boost.
All this video is doing is providing the correct facts so that misinformation isn't spread and then regurgitated by others on other platforms.
As I said at the end. Just use your ears and make your own decisions. Only time I think about phase in a mixing context is drums, multi miked instruments and poorly coded anti aliasing filters at 48k as these issues can be fixed.
Apart from that I never really think about phase shifting in eq's. Unless I use split eq, which I use a lot and I know fine well that its some mighty phase shifting splitting the tonal and transient. I can hear it when you really get the 2 elements doing different things so I judge by ear what's too much and if the difference it has made is worth the big phase shift. I only use to bring up specific transient frequencies for drums but sometimes it can have its more abusive uses for badly recorded sources. Just a case of judging by ear if it's worth it
@@PaulThird I wasn't saying you said it was evil - or judgy in any way. You were very "just the facts ma'am". Go P3 v2.0 LOL
Really I was talking of the general obsession with phase in EQ when there are better things to worry about - like if Beyonce actually wrote (or even did) anything on her new album (that seems to be nothing but samples off other people's famous records if one believes the chatter)!!!
I also don't think of phase shifting when working, it is by ear/feel.
I will use EQ out-of-phase to make things like cabinets (I have a video or two on that).
:-)
My aspergers gives me what I call social dyslexia so best to just cover my arse 🤣🤣
But yeah I get why people would complain about phase shifting with big ass cuts and boosts and I also get why people wouldn't give a shit so I'm learning in these occasions its better to just give the facts and sit on the fence.. MUCH easier life on youtube haha
@@PaulThird I understand. People (like the spank right now on my video about Morph EQ) commonly say I said the complete opposite of what I clearly said. I think that is more about their passive-aggression and conniving as a result of their fear than anything to do with you or I expressing our Obama-given right to be abused by the likes of them.
You have never owned me an explanation.
:-)
@@PaulThird I put my ideas/approach into a video of my own - deliberately a bit provocative after the minute morphing dev debacle
ua-cam.com/video/eVNA-VqU2HI/v-deo.html
Interestingly, in my inexperienced dive into mixing, I found this is more problematic in dirty electric guitars (real ones, recorded with two mics on the amp) and vocals, not so much on drums. I might be wrong of course, it’s just what my conclusion from my hears and from checking the phase.
This is because you are dealing with multi miked instruments. If you don't eq them as one then you are affecting the phase relationship between the 2 mics.
Eq'ing causes phase which is a delay of frequencies so if you delay frequencies seperately then you alter the phase relationship so essentially the work that went into minimising the phase in the recording process was kinda for nothing if you go hacking away at eq seperately.
That's why it affects multi miked stuff way more. That's why I bus my kicks, snares, guitars and eq them as one instead of seperately.
@@PaulThird Thank you so much Paul. Yes, I do EQ both mics of the guitars together in a bus, usually just a HPF and a LPF. I try hard to get a great sound when micing to avoid any EQ afterwards. In the drums, however, I’m EQing everything separately. I’ll take your advise and review my workflow on that. Thanks again for taking the time to reply with such great info. I truly appreciate it.
Paul! We do not eq a single sine wave signal (or sweep) but a signal with a complex frequency spectrum . Thus a single harmonic from that spectrum has much less energy than an sine sweep even when you boost 4kHz by 10db. I hope you will catch my little drift here...
I do but if you go boosting 15db at 10k into a saturation plugin .. Your mostly likely gonna get a ton of aliasing
@@PaulThird I think you will never boost 15db at 10k on a harmonic poor signal which is the only case I can imagine where you will do this.
Yeah it's purely just an example of worst case scenario
thanks, hm, so what better to use on master buss with +-3-6 db correction in pro q, zero latency or linear phase? better accept phase shift or pre ringing? honestly i cant hear both of it or do not know what exactly listen to. sure, best way would be to mix it properly...
Tough one. Use your ears on that one. I use bass room and mixroom on my master and they are set to linear phase by default and I love the results so my advice is do a side by side comparison and pick the one you think sounds best
The difference around -48.5 is really big and very audible I must say - I hear such differences immediately. Only differences below -80/90 are negligible from what I've tested with my ears (while doing null tests of different eqs). Differences around -65/-70 are still really audible.
It is audible but in a blind test its still subtle as its only audibly affecting a small part of the frequency spectrum. It's there and if you listen critically you will pick it out but still subtle in the context of a mix as its only a db boost
sicne you;re older remember you cant even hear over 15Khz in your 40ies etc...12Khz if you're much older... so it might be psychological what you claim ?
@@levondarratt787 no, I work with audio every day that why I have very sensitive ear. I can pick up the differences between A i B in a blind test precisely if there is around -65dB null. Below -80 I can't tell. But differences higher than -50dB null are really big and very easy to hear for most people
Tbh he is right. It is audible at that level. Its the "really big and very easy to hear" part that's subjective.
Only way to prove it is to triple blind test with original, dB boost and dB boost linear phase. If you can pick each out in a 3 blind tests on the trot one after the other then it is easy to hear.
That's how you prove it for each person
@@PaulThird most people can learn to hear these small differences. The key is to focus your hearing on ranges (low end, low mid, hi mid, highs, air - how solid, clean/dirty/smooth/hard/natural/artificial they are) and transient part of the audio especially. The samples around 3-5 seconds are best to notice the differences while being very focused on each aspect at a time. If you know what to listen for you'll hear the difference immediately. Once you notice the difference you describe it in your head and store this sound in the memory.
It’s incorrect to say that linear phase eq does not distort. Pre-ringing is certainly a distortion of the original signal and much more bothersome IMO.
Is it distortion though? This is from the crave website
"As a side note, all EQ suffers from post-ringing but our auditory senses don't notice it nearly as much. Linear phase EQ effectively halves the post-ringing and instead provides an equal and opposite amount of pre-ringing."
So all linear eq's ring so does that mean that all eq's distort? Depends on how technical we want to get on the terminoligy of distortion. I'm talking total harmonic distortion that can be measured.
In regards to pre ringing in the low end that's why kirchhoff offers a mixed phase mode where you can have your low end natural phase and the rest linear phase. You choose where the linear phase starts. So there are work around in modern day times.
Tbh they all have subjective positives and negatives. Depends on each person's taste and ear
I certainly consider transient-smearing pre-ring distortion moreso than inaudible post-ring. On sources with less transient information, it matters less, but EQ your snare drum or toms to your liking then A/B linear phase vs minimum phase. Unless running a CPU-crippling FIR window, the difference is noticeable enough to make a video about it.
Thank you for helping the audio flat earthers
Haha 🤜🤛
Less than obvious whether this is area under the curve or gain with your example.
How do you mean?
Hypothesis: ... too much aliasiasing will tends to produce 'ear fatigue' faster than the 'same mix' without aliasing ...
... a bit silly hypothesis, because, the mix without aliasing is not 'the same mix' ... but what is pretended to be called 'the same mix without aliasing' is taking all the procedures for preventing aliasing to occur and set all the 'mixing parameters equal' ...
Next, testing the hypothesis is not easy at all to realize, because, each human has a subjective and circumstantial 'threshold' for experiencing 'Ear Fatigue' ... so, setting an experimental 'psycho metric' setup is practically a nightmare that requires an obsene number of already profiled individuals ( as experimental subjects ) for getting some sort of statistical variables that can build the construct 'Ear Fatigue' in the experimental model ...
The point of all this babble is that it could be possible to test the 'undeseable listening effects from aliasing' and why it is objectively recomendable to get rid of it, inclusive, if a great percentage of the population will not be aware of its presence in an audio source ...
It's pretty audible in the tests I've done. Less transient detail and overall less clarity. When it builds up its very apparent. It's unmusical frequencies being brought back into the audible range so I've never seen the benefit of it unless you want something to purposely sit further back in the mix. I'll sometimes use lofi on backing to give them less detail so they sit better behind the vocal
@@PaulThird Who knows?? Maybe the AI generation will tell to their AI system to make some 'vintage' Digital era Aliasing effects for their mixes ... 🤭
Alexa gimme that 00's over limited smeared with aliasing sound 🤣
🤓🤓🤓
Remember to check out my autism channel if you want to learn more about my life 🤓🤓
Damnit I just do not understand it...Worralls video wasn't easy to understand this one also doesn't fit in to my brain cells its like rocket science 😭 what phase guys I see it's but how does it effect the sound 😵💫
ua-cam.com/video/JTVIXOxCafs/v-deo.html
Try that one
@@PaulThird i do understand that the phase shifts in time so in could cancel out the original one when compar. but how does it have a negativ effect on the material in general...and should i always HP with LinPh in Pro-Q3 ?
In other words new engineers, stop confusing phase with non-linearity.
🤓
Yeah, but when you need 20 Db, you need 20 Db, so you do it.
Good luck having an acoustic kick in metal without that much raping of the sound.
Yeah. You do what needs done. I don't think about phase unless it's multi miked instruments but at the same time I don't really go hacking away with massive boosts or cuts unless as you said its really needed. I Actually do more LP and HP filtering for tone shaping . Out of interest kind of eq moves do you make in metal compared to other genres?
@@PaulThird
The high peak on a metal kick is something that Nature didn't provide :)
A Bass Drum is not supposed to have 15 Db spike on 8 KHz, that's not what you hear when you are in front of it... yet that's what you need, followed but a multiband compression to make it sound even less natural.
Of course that won't be enough, so you star cuttin all the mids until it sounds like Typewriter :D
Id probably just sample the shit out of it with some metal kicks haha
@@PaulThird
That's what most people do, me included when the source is not good enough, but when I can, it's fun :)
its subjective. correct. use your lugs.
🤓🤓
"going to have nyquist then going to fold back". this does not have to be like that. could be a bad plug-in or something else. clearly a bad implementation.
Nope, it's just digital science. If you create harmonics past niquist they are reproduced within the audible spectrum. It's a computing thing. Only way to fix that is to oversample the plugin or work at a higher sample rate to increase niquist.
The most common sample rate is 44.1 or 48khz which has a frequency range of up to 22k/24k.
If you have a non linear plugin bring fed a lot of high frequency content then it will create more higher order harmonics. The more orders of harmonics you make at these frequencies the more aliasing that will occur.
Many people confuse it for others by trying to use 1khz to measure aliasing which is unrealistic as it would take 23/25 orders of harmonics to exceed niquist which is a lot.
However take 10khz for example. 2nd order harmonic is 20k, 3rd is 30k. Only takes the first odd harmonic to exceed niquist.
So any frequencies saturating past 12k will excced nuquist as at 13k the 2nd order harmonic will be at 26khz which is past the frequency limit (nuquist) of that 48k sample rate.
Most modern plugins that include high levels of saturation internally oversample so when using the plugin at its reccomended input level it won't alias.
But as the black box shows.. Internal oversampling has its limits when cranking high frequency content into the saturation.
@@PaulThird "Only way to fix that is to oversample the plugin or work at a higher sample rate to increase niquist. " - which confirms my comment. this is one way doing it with proper filtering. Don't use plug-ins that pretends to be analog, yet create alias in the nyquist range.
me stupid.. i just use my ears.. if it sounds good I'll use it.. huge wide boosts or whatever...sometime phase mess can sound better.. ears are strange.
DO THE THINGY AND PUT IT TO THe POWER CUPLING AT THE gubibly goob ..all BULLS HIT ...use your ears
Honestly very boring way to deliver information
How else are you meant to deliver it?
@@PaulThird
Maybe he wants you to perform a tiktok dance skit.
@@ChristianIce 🤡