"I don't even know if I could make something-" *Immediately reconsiders* "I mea- I could make something that loud." Hilarious and informative Seamless, thanks!
mp3's are still limited to 0db, and going over will clip (or something to do with PCM vs lossy formats not having functional bit depth, but for now lets just go with 0 db). If you have a 32 bit wav with peaks over 0db and you convert to mp3, those peaks will clip. I'm fairly certain. you can, of course, test this as I did in this video by rendering a file that's over 0db as 32bit then converting into mp3. If, when you turn down the mp3, it still distorts below 0db, you have your answer :]
Ableton's engine does work at 32bit FP (just found a forum post from 2009). And one more thing... even if we had 32-bit FP physical outputs, we'd need Royal Airforce jet-engine based generators to power the amplifires that would feed the speakers that would deliver the 1500dB range (and this setup would probably have to be installed on the Moon for safety), so it's not about truncating to 24 or 16, it's about the physics of the world we're currently born in. But your bitwise explaination does the work perfectly for the most.
At the end of the day, everything's going to come down below 0db one way or another. The end result using 32bit is only different right up until you mix it in the master. 32bit internal playback means you only need to care slightly less before that point. But most of the usual tropes about mixing and mastering still apply.
So let me get this str8, I should really only use 32 bit float for when im rendering out my master tracks to be sent to another studio for mixing or mastering and just in case one of my wav files is distorted they can bring it back from the dead!, so this would really be good for vocals
You can't. FL Studio's internal rendering engine (as far as live playback is considered) will only ever perform at 32bit float. You CAN export in 16-bit or 24-bit.
Probably. I seriously doubt Maximus would work too well with a 900db signal. On the other hand, the resulting behavior might be interesting or unique :]
So later we will see audio interfaces and monitors which are also compatible with 32 bit float audio. Does this mean that this development of non clipping audio files and systems will reinforce the loudness wars?
I was just asking if I could render it in 32bit float, and then save it for remastering later on just in case I learn better ways to limit and compress and master(like I'd listen to a song without a limiter lol)?
Thank you! I've never understood what 32bit float was. Now I have a good idea what it means (although I probably have to re-watch this video a few times :))
in terms of bit size, like 16/24/32 bit.... for every additional bit, the number of possibilities doubles. 8 bits has 128 posibilities 9 bits has 256. go ahead. 2*2^32 vs 2*2^24 It's a HUGE difference. 25 bits would have twice the headroom as 24 bits. 26 would have 4 times..... 27 would have 8 times... 28 would have 16 times.... 29 has 32 times, 30 bits has 64 times, 31 has 128 times.... 32 bits has 256 times the headroom of 24 bit. just look how binary numbers work.
I understood what is 32bit FP but I dont understand what is the benefit of using it. Especially why do I want to go red with any channel to turn it down afterwards anyways.
now i know stuff :) Vey good explanation! i usualy try not to go over 0db in master monitor from fl, but it sounds to quiet, although my ears/speakers "said" that its good
Strangely enough, Garageband does this. You can't choose to turn it on and off, but if you push a track past 0db, there is no effect on how the track sounds. The master track doesn't clip either, so the only time you hear the clipping is when you export it with normalisation turned off. I find this really useful, cos I'm always pushing tracks way too far. I have used a newer version of Garageband where individual tracks do clip, but my 2009 version doesn't, very useful. :D
In the future can you make a mixing and mastering tutorial? Even though you said you don't do it the traditional way could you still explain it or at least explain your way? Keep up the good work btw.
1st off, thanks! I think you are probably the BEST fl studio user there is and I think you also know that. My question is, what does this mean for my render to mp3? I would love to always render to wave, but when someone wants a beat and i can only send them mp3's because of the size of the wave file, I am forced to send them mp3 with highest bitrate possible. So how is a mp3 render affected by 32 bit floating point? Thanks again!
Does this work if I record something onto FL? As in, say for example I record a guitar that has an extremely loud clipping signal using FL's audio clips recording method, will I lose info if I then bring the clipped out audio clip down or would this apply to that as well? I'm basically asking if it records audio clips in 32 bits too or if it's just a rendering option. Thanks! I really love your videos, they've helped me a lot! Greetings from a fan from Argentina!
I only just saw your comment 2 years later lol you probably figured out the answer but I'll answer for others. I'm not 100% sure on this but taking the information Seamless just explained in the video I don't think you would be able to do this since the interface itself is creating the signal but like Seamless said 32bit interfaces don't really exist at the moment but if you were to ever find one in this ever technical-advancing world, as long as it stays there at those other important settings 44,100 khz ect. then you would be in luck. Boy I'm looking forward to 32bit audio interfaces aren't you?
So let's pretend there are 2 inserts, one is at +5db other is at +15db and the master is compressed and limited to 0db, are we going to hear the +15 insert louder or not ? why ?
What about gain staging your mixing plugins - for example distortion, saturation, also "vintage shit" in case you use that kinda stuff? (I mean sure, with waveshaper and maximus it's simple as shit) :D I'm asking because I've never noticed a huge problem, but then again I've been too ignorant and lazy to even think about gain staging so far
Thanks nice tutorial ! but I got 1 question, I was visiting a proffessional and he uses cubase. In his settings he could change the quality of cubase, he could set it on 16, 24 and 32 bit float. NOT rendering but in his program he did that. Is it possible to see in which condition I'm producing? is it 16, 24 of 32 bit float? Hope someone could understand it :) Bye !
FL Studio isn't using the MP3 format to its full potential. MP3 has an internal dynamic range of at least 200 dB, with good portion of it over full scale. The only program known to me that writes MP3 without clipping is Reaper. Other ways exist to bring quiet MP3 encodes over full scale, such as processing with mp3DirectCut. Many programs can also bring the level down to recover clipped peaks. Other lossy formats do not clip, but again some software doesn't support it. Internally the MP3 format breaks the audio down into subbands, which have a certain signal level. Their average power is recorded as the global gain, which is usually somewhere in the middle of the allowed range even if the sum of all the partial components would peak near full scale. There are two schools of thought concerning floating point in audio. One says that float wav files above full scale are 'illegal' (the values 'must' be within range of -1 and 1), and that is sometimes used as an excuse for not implementing headroom in software. Technically, it is possible to make clipping MP3s that satisfy both the users who want loud and squashed music, and also ones that want good dynamics, by utilizing the headroom. While most encoders do not support making these files, all good players know how to turn them down.
+TheRemonie the device (say an amp) can take say the maximum amplitude of plus-minus 5V. But you feed 5.5V into it. What happens? In those moments when the signal reaches beyond 5.0, the device registers only 5.0 (it can't receive more than that). Then imagine as if the top of the signal is clipped with the scissors - cut flat. This introduces some growling into the sound, that's distortion.
The 32 bit internal system of FL Studio is completely different to 32 bit depth of audio files. The 32 bit engine refers to how it manages and handles data. The same way you have a 32bit OS as well as 64 OS. You can't run 64 bit plugs on FL Studio because it is only 32 bit. Cubase is 64 bit. Logic X is 64 bit. FL Studio isn't.
but now FL Studio fully support 64bit plugin and We can use full range of memory above 3GB. but 64bit version of FL uses VST Bridge when you try to load 32bit plugin.
***** Yes but he is confusing 32 bit programs with 32 bit audio depth. That's why he was saying "it runs 32 bit internally". I haven't misunderstood whatsoever. He refereed to the recording bit depth as "the internal system of FL Studio". That's completely nonsensical. All he did was show the difference between a 16 bit audio file with a 32 bit audio file. This has nothing to do with FL Studio. But he is pretending that FL Studio has some advantage for this when you can do the same on almost any DAW. When you mix in FL Studio, it doesn't have an internal system that runs at "16 bit or 32 bit" etc. this only applies to recording - and how well your recording is depends on your sound card and sound drivers.
yooha yang Nope. I understood exactly what he is talking about. It is seamless that I believe is confused. I have an audio science degree. I wrote essays on this shit and had to do binary calculations and work out the theoretical dynamic range of 16 bit and 24 bit myself. I understand what float is too - as well as dither. I can give you detailed explanations of this without referring to Wikipedia (which isn't a primary source of information).
You are the best. By far. Please don't under estimate how appreciated you spending your time going over basic things like this is.
"I don't even know if I could make something-"
*Immediately reconsiders*
"I mea- I could make something that loud."
Hilarious and informative Seamless, thanks!
mp3's are still limited to 0db, and going over will clip (or something to do with PCM vs lossy formats not having functional bit depth, but for now lets just go with 0 db). If you have a 32 bit wav with peaks over 0db and you convert to mp3, those peaks will clip. I'm fairly certain. you can, of course, test this as I did in this video by rendering a file that's over 0db as 32bit then converting into mp3. If, when you turn down the mp3, it still distorts below 0db, you have your answer :]
Depends on your definition of "need". And you wouldn't need to render it out to get the benefits, the internal playback works in the same way.
Ableton's engine does work at 32bit FP (just found a forum post from 2009). And one more thing... even if we had 32-bit FP physical outputs, we'd need Royal Airforce jet-engine based generators to power the amplifires that would feed the speakers that would deliver the 1500dB range (and this setup would probably have to be installed on the Moon for safety), so it's not about truncating to 24 or 16, it's about the physics of the world we're currently born in. But your bitwise explaination does the work perfectly for the most.
Imagine the width of the cables...
installed in the moon would not probably solve things, because sound does not travel through space i guess? sorry bad english
With this dynamic range I think we'd hear it even through the space ;-)
@Hi Mod well now the discussion can take two paths depending on whether it's a philosophical or flatearthian argument :-D
At the end of the day, everything's going to come down below 0db one way or another. The end result using 32bit is only different right up until you mix it in the master. 32bit internal playback means you only need to care slightly less before that point. But most of the usual tropes about mixing and mastering still apply.
This makes things for mastering engineers really interesting, and actually quite a bit easier.
So let me get this str8, I should really only use 32 bit float for when im rendering out my master tracks to be sent to another studio for mixing or mastering and just in case one of my wav files is distorted they can bring it back from the dead!, so this would really be good for vocals
really really good video man
Ok, how do you change it from 32 bit to 16 bit float? Like i looked trough the fl manual and i cant find how to change the playback
You can't. FL Studio's internal rendering engine (as far as live playback is considered) will only ever perform at 32bit float.
You CAN export in 16-bit or 24-bit.
Any producer in 2021?
Thanks for this proper explanation Seamless, you rock!
Probably. I seriously doubt Maximus would work too well with a 900db signal. On the other hand, the resulting behavior might be interesting or unique :]
So later we will see audio interfaces and monitors which are also compatible with 32 bit float audio. Does this mean that this development of non clipping audio files and systems will reinforce the loudness wars?
I didn't understand this at all. Are you able to go over this in your next livestream? I'd really like to know what 32bit Float is.
You can infact get soundcards and interfaces that run at 32bit.
I was just asking if I could render it in 32bit float, and then save it for remastering later on just in case I learn better ways to limit and compress and master(like I'd listen to a song without a limiter lol)?
Oh god.. thank you so much! i searched for this information since two days..! THANK YOU
Thank you! I've never understood what 32bit float was. Now I have a good idea what it means (although I probably have to re-watch this video a few times :))
in terms of bit size, like 16/24/32 bit.... for every additional bit, the number of possibilities doubles.
8 bits has 128 posibilities
9 bits has 256.
go ahead. 2*2^32 vs 2*2^24
It's a HUGE difference. 25 bits would have twice the headroom as 24 bits. 26 would have 4 times..... 27 would have 8 times... 28 would have 16 times.... 29 has 32 times, 30 bits has 64 times, 31 has 128 times.... 32 bits has 256 times the headroom of 24 bit. just look how binary numbers work.
had no idea it was playing back in 32 bit float informative thank you
So you're saying that I don't need a limiter or compressor until I go to master? if I render it in 32bit float first o.O
It depends on the project
I understood what is 32bit FP but I dont understand what is the benefit of using it. Especially why do I want to go red with any channel to turn it down afterwards anyways.
+Marcin P just freedom of movement, like whatever, you don't care, it can't clip, ever.
Thanks Seamless. Very lucid explanation.
aren't some plugins designed to work best within a certain range? so if u feed a 5db signal in, the plugins algorithm wont perform as intended.
tq for everything seamless
now i know stuff :) Vey good explanation! i usualy try not to go over 0db in master monitor from fl, but it sounds to quiet, although my ears/speakers "said" that its good
Sure, but that would be functionally identical to limiting the track before hand and rendering straight to an mp3.
Strangely enough, Garageband does this. You can't choose to turn it on and off, but if you push a track past 0db, there is no effect on how the track sounds. The master track doesn't clip either, so the only time you hear the clipping is when you export it with normalisation turned off. I find this really useful, cos I'm always pushing tracks way too far.
I have used a newer version of Garageband where individual tracks do clip, but my 2009 version doesn't, very useful. :D
What software do I us to record my vocals
if you reset your settings, does anything delete?
In the future can you make a mixing and mastering tutorial? Even though you said you don't do it the traditional way could you still explain it or at least explain your way?
Keep up the good work btw.
Coolest video from you ever.
1st off, thanks! I think you are probably the BEST fl studio user there is and I think you also know that. My question is, what does this mean for my render to mp3? I would love to always render to wave, but when someone wants a beat and i can only send them mp3's because of the size of the wave file, I am forced to send them mp3 with highest bitrate possible. So how is a mp3 render affected by 32 bit floating point? Thanks again!
Many people think you can't clip, you can. It's not analogue. As you said, as long as you render in 32bit float.
5:40 Bricks have been shat. This was useful, thanks
Good to know. Thanks, Steven!
Does this work if I record something onto FL? As in, say for example I record a guitar that has an extremely loud clipping signal using FL's audio clips recording method, will I lose info if I then bring the clipped out audio clip down or would this apply to that as well? I'm basically asking if it records audio clips in 32 bits too or if it's just a rendering option.
Thanks! I really love your videos, they've helped me a lot!
Greetings from a fan from Argentina!
I only just saw your comment 2 years later lol you probably figured out the answer but I'll answer for others. I'm not 100% sure on this but taking the information Seamless just explained in the video I don't think you would be able to do this since the interface itself is creating the signal but like Seamless said 32bit interfaces don't really exist at the moment but if you were to ever find one in this ever technical-advancing world, as long as it stays there at those other important settings 44,100 khz ect. then you would be in luck. Boy I'm looking forward to 32bit audio interfaces aren't you?
So let's pretend there are 2 inserts, one is at +5db other is at +15db and the master is compressed and limited to 0db, are we going to hear the +15 insert louder or not ? why ?
damn... every producer should know this
render in 32 bit wav or not
Good Video!
Please make a video about the whole thing with mixing with Fx (Limiter etc) on the master channel!
Thanks! :)
What about gain staging your mixing plugins - for example distortion, saturation, also "vintage shit" in case you use that kinda stuff? (I mean sure, with waveshaper and maximus it's simple as shit) :D I'm asking because I've never noticed a huge problem, but then again I've been too ignorant and lazy to even think about gain staging so far
Seamless have Facebook and Twitter?
Thanks nice tutorial ! but I got 1 question, I was visiting a proffessional and he uses cubase. In his settings he could change the quality of cubase, he could set it on 16, 24 and 32 bit float. NOT rendering but in his program he did that. Is it possible to see in which condition I'm producing? is it 16, 24 of 32 bit float?
Hope someone could understand it :)
Bye !
The crickets in the background
1600db is hilarious. Thanks heaps man haha :)
omg thank u this helped me alot
Thanks for the tutorials :D
I've regularly had other inserts going over 0 and I thought that was a normal thing it didn't clip. :D
FL Studio isn't using the MP3 format to its full potential. MP3 has an internal dynamic range of at least 200 dB, with good portion of it over full scale. The only program known to me that writes MP3 without clipping is Reaper. Other ways exist to bring quiet MP3 encodes over full scale, such as processing with mp3DirectCut. Many programs can also bring the level down to recover clipped peaks. Other lossy formats do not clip, but again some software doesn't support it.
Internally the MP3 format breaks the audio down into subbands, which have a certain signal level. Their average power is recorded as the global gain, which is usually somewhere in the middle of the allowed range even if the sum of all the partial components would peak near full scale.
There are two schools of thought concerning floating point in audio. One says that float wav files above full scale are 'illegal' (the values 'must' be within range of -1 and 1), and that is sometimes used as an excuse for not implementing headroom in software.
Technically, it is possible to make clipping MP3s that satisfy both the users who want loud and squashed music, and also ones that want good dynamics, by utilizing the headroom. While most encoders do not support making these files, all good players know how to turn them down.
SLICEX next! please bro
this is not a joke..how come i hear crickets in the background. isnt it suppose to be a sealed tight studio room.
great explenation! :D
If I export in 32 bit float on a wav file will the quality be better?
That's what I m searching still didn't get it ? If got answer then pls reply
thanks
Vid is 13:37 Long = Leet Long
Can someone give me the defenition of distording?
Distorts the waveform to make it sound louder and more powerful.
+TheRemonie the device (say an amp) can take say the maximum amplitude of plus-minus 5V. But you feed 5.5V into it. What happens? In those moments when the signal reaches beyond 5.0, the device registers only 5.0 (it can't receive more than that). Then imagine as if the top of the signal is clipped with the scissors - cut flat. This introduces some growling into the sound, that's distortion.
Thx
So always use 32 bit float
Well sure.
oh my god 4:45 ...your desktop must be brutally unorganized!
fuck ya seamless!
wow this was so useful! :o
The 32 bit internal system of FL Studio is completely different to 32 bit depth of audio files. The 32 bit engine refers to how it manages and handles data. The same way you have a 32bit OS as well as 64 OS. You can't run 64 bit plugs on FL Studio because it is only 32 bit. Cubase is 64 bit. Logic X is 64 bit. FL Studio isn't.
but now FL Studio fully support 64bit plugin and We can use full range of memory above 3GB.
but 64bit version of FL uses VST Bridge when you try to load 32bit plugin.
***** exactly, I think he misunderstood that
***** Yes but he is confusing 32 bit programs with 32 bit audio depth. That's why he was saying "it runs 32 bit internally". I haven't misunderstood whatsoever. He refereed to the recording bit depth as "the internal system of FL Studio". That's completely nonsensical.
All he did was show the difference between a 16 bit audio file with a 32 bit audio file. This has nothing to do with FL Studio. But he is pretending that FL Studio has some advantage for this when you can do the same on almost any DAW.
When you mix in FL Studio, it doesn't have an internal system that runs at "16 bit or 32 bit" etc. this only applies to recording - and how well your recording is depends on your sound card and sound drivers.
yooha yang Nope. I understood exactly what he is talking about. It is seamless that I believe is confused.
I have an audio science degree. I wrote essays on this shit and had to do binary calculations and work out the theoretical dynamic range of 16 bit and 24 bit myself.
I understand what float is too - as well as dither. I can give you detailed explanations of this without referring to Wikipedia (which isn't a primary source of information).
***** There may be ways around it but that's irrelevant.
Love you in a non homo way. Thank you a lot!
Dude music is damn interesting if u actually start understanding shit
hmmm
ilovefree99
y u so smert