ACX Audiobooks Specs: Peaks, Valleys & Misconceptions

[Music] wouldn't be talking about ACX and audiobook specs and the peak values because I get this all the time where people are trying to hit minus three and I have about three PowerPoint slides here and then we're just gonna go into demonstration if you have questions go ahead and ask a Kelly thanks for joining in julio glad you're there all right good good good and Barbara yeah good to see you there alright so we're talking audios book specs and here's what happens here okay so most of you already know this minus 3db is the official maximum amount so your specs have to come in you have to the RM x rms measurements and that's between minus 18 and minus 23 and you know about that but then there's this peak of -3 that's the official maximum but it's not a target I see people way too many times attempting to hit minus 3 with their peaks and trying to balance out the two and that is not the way it works anything less than -3 goes and I'll show you some of the reasons and then the way this whole conversation got started was this that we have we have a situation where people are going ahead and they're encoding to mp3 and they make sure they do a second opinion and they check their specs they do a CX check and they check and they say all right I'm at minus three and then they encode to mp3 and then if they check later it ends up being different and the reason is encoding where you turn it for it's not just translating or it's really encoding where you're taking audio that's WAV format and you're moving it into mp3 and during that process you have this thing called masking and we're going to talk about psycho acoustics okay but before we do any of that I'm just gonna tell you there's a whole bunch of you there that could care less about the details so here's the executive summary and you just shut this off right now set your limiters to minus 3.4 so and you'll always be good okay so that should be just the sample rate minus 3.4 so it'll work out time of course minus three is not a maximum it's a target you could be at minus four you could be at minus five nobody will hear the difference between my audio at minus three four the peaks or minus three point five for the peaks especially if you have a long chapter you've got 20 minutes audio one thing hits that phil never hear it and just a side note I'm talking about audiobooks today when I'm doing a voiceover project or an e-learning project or a video sync something that's going like that that's not an audio book lots of people do use the same specs they use for audiobooks and that's fine that'll work it's not what I would do I think there are better options for voiceover then and and the reason we want all this Headroom we want to be at minus-3 for audiobooks is audible and ACX are going to process our audio after we've submitted it to them and some other publishers as well and if they're gonna do some processing what they have done is they've allowed themselves some Headroom so that they can do their thing and they don't have to adjust our audio up and down because of the work that they're gonna do next the final product is not at minus-3 you go analyze any audiobook and you'll see it says that's that's our submission that's not really what it's going to end up like okay so let me scroll over here and see if anybody said anything hello Annette hey Dan hey Hugh glad you guys are there and Wade all right then everyone so by the way this Scrolls by and I don't see them all so if I don't sorry about that I don't mean to be rude and miss somebody all right I'm sorry but that's the end of my slides so I'm gonna get out of that so here's what we're gonna do we're going to talk about mp3 encoding and what happens is we have to talk a little bit about Wave mathematics because encoding starts with this whole concept of psycho acoustics and what it means is this I put on my glass as I look a little different if I put a mask on I would make a lot of people happy that they don't have to see me but you would just see my eyes or you would just see my mouth and some parts are masked so I don't need to have a nose if I have the right mask on I don't need to have ears I don't need to have any hair if I put a mask or a hat on I put a hat on and you don't really know if you don't know me already saw me on the street if I had a hat that covered all my hair you don't know whether what's behind the mask well in audio when we encode from wav file to our final mp3 for delivery to our client we get this interesting situation mp3 is taking out stuff and when we take out stuff strangely enough the signals can go higher or the peaks can go higher and it confuses people because you kind of think if I take something away it shouldn't go up but Peaks are not the same as total volume for example the rms level will never go up when we do this and let me give you a couple examples if you d click you'll sometimes see the peaks go up if you go ahead and encode mp3 or you D noise you D noise you're taking out stuff you shouldn't see these go higher but you do and here's why so I have an example here of a waveform it's a twenty Hertz signal it's all the way at the bottom I'm gonna switch it over so we only see the waveform here and then you're looking at this and okay where's it going there's a little meter over here if you're not familiar with our X I'll assume most of you are in the RX group and so right there I have that sitting right at minus three you can see this meter right here going right to minus three top and bottom okay so here's something really interesting let's say that's the simplest thing we can have that's me throwing one little pebble into the pond and I get some waves and this how audio works this would be a nice low signal matter of fact most people aren't going to hear this this is that this is at twenty twenty twenty Hertz okay but it's a nice a big one to look at this so what I'm gonna do here is I'm gonna go over right to here and I'm gonna grab a little bit of audio and I'm gonna just highlight it I need to make sure I'm there okay so I'm gonna take that and I'm gonna copy it so this is the same signal but I'm you know what I take that back I'm just different this is why we're live watch this I'm gonna take the same signal and I'm just gonna copy a little bit of it okay then we copy that much I'll do it with a keyboard shortcut so you can see what I'm doing I just copied it now what I'm gonna do is I'm gonna go and I'm gonna paste it right here now most of you can guess what's going to happen I'm gonna add these two together I'm gonna do a special thing edit paste special and I'm gonna Aksum combine two signals together and watch what happens to my audio I almost zero it out and frankly if I had gotten the exact correct point it would zero that audio right out because I had put a negative with a positive and so now my audio is that nothing but if I undo that so this is what happens when you encode to mp3 I'm gonna take something out now I'm gonna take out my second pace and watch what happens when I undo that it goes back up and this is exactly what happens to people when they start out with their audio they encode to mp3 and that masking effect says this when mp3 is doing its job it's using this concept of psycho acoustics and they have figured out over a couple decades that some sounds that are quieter we don't hear like a mask you don't see every section of a person if they have a mask on so they make sure you can see the eyes they make sure you can see something if it's a kind of mask where you show through and the things that are prominent that's cool that's what you see and that's what you hear are the prominent things and they know that they can throw out some things to make the file size smaller that's why mp3 is so much smaller than a WAV file they throw out a bunch of stuff and when they throw out stuff it's the same thing as what I just did there you have two signals that came together and they create a peak and if I took them and I put them together differently like that the first time I put them at a certain spot which negated them I went down to zero if I start them at another spot and edit paste special and mix them together now my Pete's just went obnoxiously high and clipped so I added two waves that were identical and now that went higher depending on where they all come together will depend on what the results are and so people sometimes I don't used to wave mathematics they were used to a normal world where if I have five dollars and I give you three dollars I have two dollars left but waves don't work the same way as that adding and subtracting does for regular math they work based on principles mathematical principles wave mathematics and how they combine which just says I throw two pebbles in the pond when they come together the peak will go higher if the valleys if the peak in the valley come together at the same exact point then they'll be evened out or zeroed out like the first example there and then if I go ahead and do some other things where let me just so here's a minus five over here I'm gonna grab a little section you can't see it I'm gonna copy it and I come in here and I paste something over which is what really happens in the real world I'm gonna edit paste special and I'll mix those together this is what you end up with sometimes they go higher sometimes they go lower if I did it again right here and it usually gets a lot more complicated than this I'm gonna mix these together and right now where they all came together everything's just going lower but that's not how it works in the real world these are very very simple waves I only have one two or three of them combined if you go through and combine them like our voice does with all our harmonics and all the other things going on you can take things away during encoding during D clicking and you go higher so what does all that mean it means don't be surprised when you export to mp3 and your Peaks are higher than they were before you encode it all you have to do it's really simple you're not trying to get to minus three I've said that four times already that's a maximum not a target so if you set your limiters at - I do - three point four for mine but you could do I started it off many many years ago five years ago I was doing maybe three point two I'd bumped it up a little bit go if you're at three point five that would be fine just set your limiters so that it's not going as high to begin with then when you encode if you want to check it again great check it out a half-dozen no let's see let me give you an idea here I this is real world audio okay so I have real world audio and you take this signal here and you go ahead and you copy it and you paste it someplace else so I just go in here and I choose edit and I go ahead and pay special and I mix that in you can see what happens there I get some pretty interesting stuff some things been higher some things went lower if we analyze that you would see the exact same thing times 400 that we saw with the simple waveforms and that this stuff all adds together and when you take something out you still can have a peak go too high make sense I hope that make sense of people if it doesn't make sense and you can ask some questions there but the bottom line is this keep your limiters a little lower than you think if you're trying to hit a CX if you're doing voiceover we'll talk about that another session it's a little bit different you don't need to meet you won't hurt anything if you do a CX type specs audiobook specs for voiceover although frankly you won't be as competitive as you can be so that's a whole nother discussion there but don't be surprised when you when you encode something and I hack out this show that wants this so I'm gonna take this little thing here and I'm gonna say file export oh let me show you something here first what's this here's this original what is this this is only a couple minutes of audio it's at - not clear something that can't be - infinity there I cleared something but here's where at - 19.3 - and our Peaks are - for 8 - you can read so I shouldn't I shouldn't need to read those and then what I'm gonna do is this export and I'll export to mp3 constant bitrate 192 I'm gonna reopen that and boom it's gonna go ahead and I'm just gonna say demo I guess I can just leave that because it will that's gonna reopen that I have that reopen so there's our file now if you look at the numbers here this was the original let me move those side-by-side so you see these numbers here remember this 19 point 59 the total RMS over here is 19 point 32 which is louder well mostly this is stupid negative numbers this one's louder this one's quieter and then look at our Peaks let me go back this our Peaks are at minus 4.8 - our peaks on the mp3 in this case are at minus 5 so this one went the other way this got quieter okay and that'll happen it that that's the thing that's maddening about it is people see it wants and they is that we generalize that that that's the way it always does haha if I go ahead and go through I just picked this one at random I picked something else and you'll see it go the other way I seen it go as high as one point how about one and a quarter DB higher in the peaks after encoding mp3 because it just depends on what's there in the audio already what frequency ranges are there do you have a lot of low-end stuff that when you take that stuff out because it was so far in the background changes all sorts of numbers they're so low-end stuff has a tendency to make bigger bigger changes but the point is always the same they are different don't let it surprise you this is a this is just one thing if I go through and I found some other short ones what do I have here this one short and now pull this guy in so this is something that's four seconds let's go ahead and see what happens on this one just to make a see if we can get something that's close we go to mp3 on this one our numbers right now are 1741 566 i'm gonna go through and save that guy i'll have the word trash here so i like to do that that means i can throw it away later save it and look at our numbers there let's see what that one did so this one was at five six six for the peaks this one's at five nine one i mean even in four seconds three seconds worth of audio there you can see that the rms is at 17 we encode and we're at 17 72 this got quieter but the peaks changed this one got this once louder and the peaks change and if you test a whole bunch of these you'll see that you'll see the patterns emerge and the long are these long chapters and audiobooks those i mean the longer the audio is the more likely you are to see changes and the more likely are to see things go higher shorter it is the less likely it is but it's still different and you don't need to really measure this stuff if your limiters were small if you like you could submit this to a CX in terms of at the die or mass is too loud okay for this one this other one was the other one I don't remember now in that you could submit this to a CX and you'd be fine because your total Peaks are in fives would anybody hear it the difference no they would not hear that this is at minus five versus somebody that's at minus three what they really are going to hear mostly is what's happening with the automat so let me check what's going on here what do you mean by saying giving yourself Headroom by setting peaks to minus three five okay John so when when I'm exporting from Studio One I'm gonna make sure my limiter the limiter on my raw audio is going to be set down to minus three point four when I first started I mentioned I used to tell people and I would set those stuff up when I was setting it at minus 3.2 over time I changed it to minus 3.4 just because it's better well 3.2 work yeah and the dirty secret is ACX actually allows some wiggle room there okay ACX will allow you to have some wiggle room and they will not give you a hard time if it goes I mean I think they just know people measure it in wave form they measured with a CX check they measure with second opinion those are against the wave files and they're not double checking it against the mp3 files but you could drag it in and see it if you ever want to just make sure that your files at WAV format before you go out and encode them have Peaks that are no more than about three point four three point three you'll be fine encode it to mp3 don't worry about it you can check a few of them if you want but you'll see and if you get one at two point nine something don't stress over it they'll take that I thousands of files have gone in that way and they just don't they're not hyper close they're not that's one stat that they just don't get hyper close about but on the other hand you give yourself enough Headroom meaning your limiters are saying I don't want a peak above three point four by the time you've encoded you don't have to worry about it think about it it'll totally work so John I hope that re-explained it or explained it better and if not then you be sure to ask another question okay so I hope you guys hope this helps some people the big takeaway don't stress over hitting minus three you're not trying to hit it it's not a target I could submit this at minus five and nobody would know that it wasn't a little bit higher it's okay no stress fair enough and then RMS itself you know you you want a good compressor and also I should note I'm not a fan of normalization normally I don't do that and normally I don't recommend normalization ok it's a weak practice it comes from the days before we had good limiters if you've got a good piece of software today you do not need it you have a decent compressor and you have some and you have some gain that you can add you just don't need to to normalize and normalization makes here punching is harder to get right especially on a long chapter that's a whole other subject that I'll deal with another day so I hope this helps anybody ask questions be sure to ask in the group if I was unclear about something I'll do another session another day explaining something if it was unclear but you know let me know of course I help people set up their rx there are expected and then you can do a final listen through and check little things and studio on in anyway hope you guys have a fantastic weekend and of course as always I look forward to seeing you on the wires you have a great one bye bye you

Channel: Red Baarns Audio

Views: 1,526

Keywords: ACX, Audiobooks, RX 7, Audio, Audio Books

Id: rX8jwIFnLcY

Length: 19min 0sec (1140 seconds)

Published: Fri Dec 07 2018