Thanks so much for another helpful and beautifully crafted tutorial, Curtis! I really appreciate your videos and they've helped me quite a bit as I make my way into the DSLR film making world. I'm struggling with a series of videos I shot right now, and this is probably the number one challenge I need to tackle - getting volume to a good point across a dozen interviews shot both outdoors and indoors. Surprisingly, it really is tricky, even compared to compression, noise removal, or sound sweetening. I'm very much obliged for your superb tutorials. Rock on!
Many thanks Curtis. I really appreciate the way you take your viewers through everything, including all the various items in a menu, terminology and explanations etc. You are providing a great service to all of us who are not professionals, but want to improve our amateur/enthusiast video productions (also subscribed to your online school, which I found very useful).
Curtis, there are plenty of "sound normalization" videos on UA-cam, so I picked yours at random, And am glad I did. I've subscribed to your channel and will be checking out the comments and your other videos. Thanks, Rowby
Great video! It's amazing that I was just struggling with normalization for some audio tracks this morning and worked very hard to get them sounding better. See a notification for UA-cam and there you are with this tutorial. Excellent! Always enjoy your videos and appreciate that you take the time to share your knowledge and discoveries with us! Thank you!
Thanks so much! I'm doing voice over work and the information in your video applies to that as well and made it very easy to understand. Personally, I find a lot of the other abbreviated video tutorials on UA-cam so abbreviated that I don't understand what steps to take. I understand that everyone is busy and modern attention spans have shrunk, but I found this long enough to be able to follow your steps and short enough that I didn't have to invest half a day! So thanks! I've subscribed and will look forward to other videos! And btw, I watched part one first and am really intrigued with Auphonic! So I'll try that and see if it doesn't do the trick for my auditions. Then the actual gigs will get the manual treatment.
Hi Curtis! Very thankful tips indeed. I find 17 min to be a but too long. Look forward to new tips! I'd love to get some advice on how to add a "radio sound" to the voice when recording a VO. Thanks again!
Irina Almgren Thanks Irina! Great feedback. For a radio sound, you might check out Mike Russell's Music Radio Creative channel: ua-cam.com/users/musicradiocreative
Curtis, thanks for all the great content. I notice in the online vids I watch with people who know what they are doing, such as yourself, the audio is in fact very consistent in terms of loudness across different videos. But, if you take me for example, I watch UA-cam a lot on my iPad, it has a lifeproof case on it (which muffles the audio a bit), my wife may be doing dishes in the nearby kitchen, or the ac is on, or somebody is watching in a cafe, or whatever. In short, when I have it cranked all the way up, your audio (and again others who know what they're doing) is not impossible to hear, but I'd say on the low side (possible to miss things because it's too low). I wonder if you consider this, because I would think my situation is rather typical. I realize these volume levels are "technically correct", but what would be the drawbacks of just cranking your levels up more, if they were consistently that high across all your videos?
Hi Andrew, yes, good question. I intentionally loudness normalize to -16 LUFS so that people can hopefully hear in less than ideal listening environments, situations, and with less than ideal playback speakers or headphones. It doesn't always work, as you have experienced, if the conditions combine to defeat the listening experience. However, there are costs to pushing the loudness even harder. When you push the loudness normalization too hard, you sacrifice dynamic range which at some point not only raises the noise floor, but also reduces audio quality. In addition, spoken word audio starts to sound really quite unnatural and unpleasant in terms of the listening experience. I'm not positive of this, but I also worry that pushing the loudness too hard increases the risk of damaging listers's hearing with prolonged exposure.
I use Audacity. I found that using the hard limiter with a residual level of 0.2 to 0.5 (depending of situation) followed by normalization seems to work better for my dialog tracks than using the compressor. By setting the residual level, it doesn't doesn't completely chop off the peak helping preserve the quality and some of the waveform..
I haven't verified it with a UA-cam video, but I tested it out with dialog I recorded along with the Orban Loudness Meter software and it seems to work as long as I normalize to -3db or -1.5db. Lately in my videos I have been improperly normalizing to -0.1db which is leading to some peaking. I'll fix that in future videos.
I'm curious as to where you'd put these steps inside of a normal workflow? Currently this is something I'm looking to incorporate since there's a lot of highs and lows in my gaming commentary and I want to create a good sound. Currently my process is: Noise Reduction - Vocal EQ - Normalize at 98.8% and ship. I'm curious if this could completely eliminate the need for normalization but would the new process look like: Noise Reduction - Vocal EQ - Match Volume - Hard Limiter. ? Would love to hear your thoughts on this and thanks for the great tutorial!
Hi Curtis I enjoy all your videos thank you for doing all of this... do you know how well this can be done with audacity or sound processors like reaper... more budget programs?
Maury O'Con Reaper has a plugin available called SWS that evidently measures integrated loudness and true peak. I haven't used it yet but it is worth a look. As far as I know, Audacity does not have any included plugins or features to measure integrated loudness or true peaks.
Could you do a video comparing Audacity to Adobe Audition? Obviously Audition is vastly superior, but for people like me who are considering stepping up, knowing the gains to stepping up to Audition will help me decide to make the step.
Hi, I discuss this some in sound post-processing course: school.learnlightandsound.com/p/sound-post-processing-1 I will also add to the list to cover in the future on UA-cam. Thanks!
Love your videos--thanks so much! Just tried this and wanted to mention that with the current CC version seems to use "true peak"--appears to default to -1db (at least it mentions the peak in terms of true peak)
There's also the new Essential Sound Panel in the latest Premiere. Have you reviewed the process of loudness normalization using that and also parameters of the true peak limiter? Would you still apply compression for dialogue audio?
Hmm, your audio kind of looks like it has a DC offset (uneven distribution between waves under the baseline and over the baseline), which will limit your ability to normalize to some extent. Weird thing is, that it looks centered everywhere there is a pause. Maybe something you should troubleshot by replacing equipment (mic, preamp, soundcard) one at a time? Also, I'm not sure I heard any difference at the point of the "true peak" that wasn't there after you applied the limiter. I heard something which could have been either analog distortion from loudness, or some temporary acoustic from back of your throat, either way, i percieved it as being there both before and after you applied the limiter. I used to work as a studio engineer in a studio that primaraly recorded rock/metal music, and as a last step (while mastering) we always compressed, limited and normalized to 0 db. The "true peak" would have been lit constantly, which has been the db standard in rock/metal music CD:s since the 1990's at least. So this importance placed on "true peak" seems odd in respect with my experiences. Maybe true peak and the loudness level at -24 db is important when it comes to broadcasting and receiving via radiowaves so that the signal never is too hot for a radio receiver? (which would imply it is of less importance in purely digital situations such as youtube-videos). Anyway, graphical representations is a great tool in many cases, but you should also trust your ears (and also take breaks to let your perception of sounds re-calibrate to non processed sounds).
Valpskott Thanks for your insights! I thought that I had a DC offset issue as well. But it turns out that in periods without signal/sound, the waveform returns to infinity and a DC offset repair tool didn't do anything. I learned that this is actually waveform asymmetry which is evidently a little more common with male vocals and some microphones. I now use a phase rotation plugin which corrects that issue. Here's a more recent episode that includes that: ua-cam.com/video/VcMOBrdbHgk/v-deo.html I used to normalize to -0.1 dB as I learned from music production but it turns out that with digital, you can still get inter-sample peaks that exceed 0dB if you do that which is what dBTP is supposed to help measure and address.
This is great Curtis. Is this something that could be applied at the track level? Meaning, could I apply this technique to all of the dialogue tracks in a multi-track session?
Thanks Joe. I haven't done that to be honest, but as I'm think it through, the types of issues you're trying to solve will require different workflows. If the loudness of each of the dialogue tracks are where you want them, you could just take the master mix into waveform view and do the "Match Volume" from there. But you'll want to address any issues you have with the individual tracks first, perhaps even loudness normalize those that are significantly quieter than the others. This could be a good idea for a future episode.
Great series, very helpful! Quick question - to compensate for the true peak in the last step, why use a hard limiter as opposed to another pass of compression?
Thanks and good question. I is a judgement call. If there are just a few transients with a single peak each, I find that limiting doesn't leave audible artifacting. But you could certainly do another round of compression as well.
Can you go through your entire process? Because you say that you should clean up your audio before you do this part but Sometimes I just can't hear the background noise enough to edit it out before increasing the loudness. I would just personally love to see your entire process. (If you already have a video up like that maybe i'll run across it watching through the playlist.)
+HollowProductions Here's one where I ran through the process. I only use noise reduction if I must so don't feel like you have to do that every time: ua-cam.com/video/VcMOBrdbHgk/v-deo.html And here's one on noise reduction: ua-cam.com/video/1lIjFBG6w58/v-deo.html And if you prefer Audacity, here's noise reduction in Audacity: ua-cam.com/video/vlJSTavIY3k/v-deo.html I'm working on a course that covers dialogue audio post processing in depth so look for that some time in January. Best wishes!
Very nice video, thank you :-) Is there a way to create "actions" for that, so that you might just mark the audio and start the recorded action, which automatically sets all these parameters? Just like in photoshop?
As far as I know Audition doesn't have actions like photoshop yet, but you can save presets for each of the plugins which helps a little--you'll see a little icon with an arrow pointing down at the top of each plugin/effect which allows you to save your setting for use in the future.
Hmmm, it looks like you can actually save normalization (with true peak limiting) in the encoder presets now so it is applied automatically when you render the video (with Adobe, anyway).
Curtis, so what's difference between normalize to -0.1db and Match volume? Also, for the noise reducetion, do we do it before match volume or after OR even both before and after?
Jason, the biggest difference is where the "Perceived Loudness" parts of your audio end up in the final audio. You could technically end up in the same place by normalizing to -0.1db and then compress and peak normalize again, etc. The problem I've found with that approach is that it is too focused on peaks which do not, by themselves, define the perceived loudness of an audio clip. So using Match Volume/Loudness Normalization, you can get more consistent results. Secondly, peak limiting at -0.1db can still result in clipping so using "True Peak" limiting can solve that problem. As a workaround, I have been peak limiting to -2.5 db and have so far been able to avoid clipping.
Thanks a lot for your videos, they're very helpful! I have a question though, do you know what's the loudness standard for film and cinemas (does this apply)? Also, I can't say I've searched enough in your channel but, have you done any videos on audio recording techniques for film (such as handling boom poles, managing cables, etc.)?, or, if you know, where can I get info on this? I'm studying at a film school but we haven't delved deep into sound. We use a tascam dr40, a 2 channel sound devices mixer, a rode ntg1 mic and sennheiser hd201 headphones, so I have a hard time with cables sometimes.
Hi Jorge, thanks! I don't believe that there is a published standard for loudness for cinema. I would probably go with -23 or -24 which are the loudness standards for TV broadcast in Europe and the US respectively. Cinemas usually have nice playback systems so using the TV loudness standards should work nicely and the nice thing about them is that they leave a bit more dynamic range than loudness normalizing for mobile and web. But if there's someone here who knows more, please feel free to comment!
As far as I understand, normalizing to -24 LUFS is quieter than to -16 LUFS, so there's more headroom for dynamics in cinema? I remember my experience of making DVDs. The Dolby Digital sound was much quieter than original WAV, because it used -27dB normalization at authoring, and then +27dB at playback. I didn't like that Dolby Digital codec is making my sound so quiet, so I combated that by setting that parameter to 0 dB and it sounded like original. Now thanks to your explanation of loudness standards I understand why DD normalized to -27dB. Do you think it's good that this is not a required parameter for other audio codecs like aac for video?
Great video. I am trying to follow your method but Audition won't let me drag the files to the match loudness window. I imagine it has something to do with the "supported files"part of it. I am using .wav files 16 bit 44.1 mono so I don't know what it can be. Perhaps the naming of the files that includes underscores and numbers? Have you encountered this problem?
I just want to be clear. When I make a video I most often have multiple clips of audio possibly in different environments. would you suggest I bring the levels as close as I can by my own ear. Then, export one .wav file for processing "loudness and normalization". Or, execute this process for each individual clip. Why or Why not. Thanks
Rhett Brown at RetroWreck I process each file separately. There may be times, particularly when doing narrative work when you may want to go quieter than -24 LKFS which would be an argument for doing each separate. If, on the other hand, you're doing something where all the clips are supposed to take place in the same environment, you may want to do them all together. In either case, having the ability to standardize your perceived loudness lets you make these types of choices with more confidence.
I had no problem using that Single Band Compressor with the 2015 (?) version of Audition, but just tried it with the new version for 2017, and can't get it to work. I mean, there is no change in the peaks of the audio file after Apply.
+George Taylor hi George. What do you have the ratio set to? And you have confirmed that you have set the threshold lower than the peaks? Also, attack is set to less than 10ms?
+Curtis Judd Okay. I think I just found my mistake. It looks like it was a threshold problem, after all. I do appreciate your response, as well as all your instructional videos. They've been a big help to a lot of us.
Can we just do thatin Premiere Pro ? Also, what if the audio file is already inside a the video file? Also what if we want to mix both external audio file + internal audio file in the video? Thanks
Hi Alex, yes, of course: ua-cam.com/video/RrRn8stkEBs/v-deo.html Mix all your audio then at export, set the loudness target. What this does make more difficult is that if you need to compress your audio to hit your loudness target, it is a fair bit more difficult to do in Premiere. But give it a try shown with the linked video and see how it works out. Good luck!
I have a question....Im putting up clips of music i did for a film, and im wondering if I should normalize the audio or leave it "as is." It is a bit softer than other music vids Ive seen in the same genre....Here is one of the vids so you can see what Im doing: ua-cam.com/video/ZiWA6CRMJKM/v-deo.html Appreciate your help =)
Hi Feona, I think it depends on where it will be "broadcast" and it is probably not unusual to have a few versions. For anything that will be listened to via mobile devices and web, you'll want to consider -16 LUFS for stereo tracks (-19 for mono). For TV broadcast, -23 LUFS in Europe, -24 in the US. I'm not so sure about films that will be projected in a proper theater, however. Perhaps Paul Figgiani would have some insight on that.
Feona, I would say your responsibility as a composer would be to produce without shooting for a specific loudness target. Ultimately this will be up to the people in post. They will mix accordingly and produce a final master that will be (in terms of loudness) adjusted based on the delivery platform. Of course most composers/mastering engineers have personal preferences when building a piece. Allot comes into play - things like dynamic range, impact, etc. It’s pretty much subjective. I analyzed your track. It looks like you may have Peak Normalized to 0dBFS? I wouldn’t recommended that. You have some ISP’s (Intersample Peaks) occurring above full scale. Take a look at this high-res snapshot: f-video.s3.amazonaws.com/fionna.png … the top is your original source lifted from UA-cam. In the lower version I applied slight compression just to tame the wayward transients. This allowed me to limit peaks to -1.0 dBTP without any sound quality degradation. In the future I would suggest you keep things well below full scale. Also note that peak amplitude has nothing to do with perceived (average) loudness. It represents the proportional voltage of an audio signal. If you are looking to create mixes that vary perceptually, you will need to look into other descriptors (Program Loudness, Loudness Range, etc.).
Hi Curtis, You should look at Tone Boosters TB Barricade and TB EBULoudness. TB EBULoudness is much usable than Loudness Radar. www.toneboosters.com/tb-barricade/ www.toneboosters.com/tb-ebuloudness/
Thanks! Interesting article. I think the intent is to compress the peaks by that amount as every recording will need tweaks to the threshold setting to get that much compression.
STILL an excellent video, 2.5 years later! Thanks Curtis, huge help.
Thanks Trent!
Thanks so much for another helpful and beautifully crafted tutorial, Curtis! I really appreciate your videos and they've helped me quite a bit as I make my way into the DSLR film making world. I'm struggling with a series of videos I shot right now, and this is probably the number one challenge I need to tackle - getting volume to a good point across a dozen interviews shot both outdoors and indoors. Surprisingly, it really is tricky, even compared to compression, noise removal, or sound sweetening. I'm very much obliged for your superb tutorials. Rock on!
Glad to see you on board Curtis! Sounds like you understood all the tech.
Thanks Ray! Wouldn't have learned about this now unless I listened to your podcast!
Many thanks Curtis. I really appreciate the way you take your viewers through everything, including all the various items in a menu, terminology and explanations etc. You are providing a great service to all of us who are not professionals, but want to improve our amateur/enthusiast video productions (also subscribed to your online school, which I found very useful).
+microbroadcast Thanks!
Curtis, there are plenty of "sound normalization" videos on UA-cam, so I picked yours at random, And am glad I did. I've subscribed to your channel and will be checking out the comments and your other videos. Thanks, Rowby
Thanks Rowby, glad it helped!
Great video! It's amazing that I was just struggling with normalization for some audio tracks this morning and worked very hard to get them sounding better. See a notification for UA-cam and there you are with this tutorial. Excellent! Always enjoy your videos and appreciate that you take the time to share your knowledge and discoveries with us! Thank you!
Hi, glad it hit at the right time! Thanks for the feedback!
Thanks so much! I'm doing voice over work and the information in your video applies to that as well and made it very easy to understand. Personally, I find a lot of the other abbreviated video tutorials on UA-cam so abbreviated that I don't understand what steps to take. I understand that everyone is busy and modern attention spans have shrunk, but I found this long enough to be able to follow your steps and short enough that I didn't have to invest half a day! So thanks! I've subscribed and will look forward to other videos! And btw, I watched part one first and am really intrigued with Auphonic! So I'll try that and see if it doesn't do the trick for my auditions. Then the actual gigs will get the manual treatment.
Thanks Robin and glad it helped!
thanks for the video, man. My videos have never been so much better!
Great to hear!
Fantastic video! So helpful to me as I start to venture into more professional shoots that require higher attention to detail.
Thanks, glad it helped! Good luck!
Great stuff. Thanks Curtis. I've been looking for this for quite some time.
Thanks James!
tonsfocus Thanks and best wishes on your videos!
Thanks so much for sharing! I just downloaded Audition and this is very helpful!
Glad it helped and I hope you enjoy working with Audition!
Hi Curtis! Very thankful tips indeed. I find 17 min to be a but too long. Look forward to new tips! I'd love to get some advice on how to add a "radio sound" to the voice when recording a VO. Thanks again!
Irina Almgren Thanks Irina! Great feedback. For a radio sound, you might check out Mike Russell's Music Radio Creative channel: ua-cam.com/users/musicradiocreative
Curtis Judd Yeah, Mike's channel is too good.
Curtis, thanks for all the great content. I notice in the online vids I watch with people who know what they are doing, such as yourself, the audio is in fact very consistent in terms of loudness across different videos. But, if you take me for example, I watch UA-cam a lot on my iPad, it has a lifeproof case on it (which muffles the audio a bit), my wife may be doing dishes in the nearby kitchen, or the ac is on, or somebody is watching in a cafe, or whatever. In short, when I have it cranked all the way up, your audio (and again others who know what they're doing) is not impossible to hear, but I'd say on the low side (possible to miss things because it's too low). I wonder if you consider this, because I would think my situation is rather typical. I realize these volume levels are "technically correct", but what would be the drawbacks of just cranking your levels up more, if they were consistently that high across all your videos?
Hi Andrew, yes, good question. I intentionally loudness normalize to -16 LUFS so that people can hopefully hear in less than ideal listening environments, situations, and with less than ideal playback speakers or headphones. It doesn't always work, as you have experienced, if the conditions combine to defeat the listening experience. However, there are costs to pushing the loudness even harder. When you push the loudness normalization too hard, you sacrifice dynamic range which at some point not only raises the noise floor, but also reduces audio quality. In addition, spoken word audio starts to sound really quite unnatural and unpleasant in terms of the listening experience. I'm not positive of this, but I also worry that pushing the loudness too hard increases the risk of damaging listers's hearing with prolonged exposure.
Thanks Curtis. I will definitely look into this -16 Lufs normalization.
I use Audacity. I found that using the hard limiter with a residual level of 0.2 to 0.5 (depending of situation) followed by normalization seems to work better for my dialog tracks than using the compressor. By setting the residual level, it doesn't doesn't completely chop off the peak helping preserve the quality and some of the waveform..
Thanks Harry. Does this ever peak after uploaded to UA-cam?
I haven't verified it with a UA-cam video, but I tested it out with dialog I recorded along with the Orban Loudness Meter software and it seems to work as long as I normalize to -3db or -1.5db. Lately in my videos I have been improperly normalizing to -0.1db which is leading to some peaking. I'll fix that in future videos.
Harry of All Trades Hey, cool--thanks for the details!
I'm curious as to where you'd put these steps inside of a normal workflow? Currently this is something I'm looking to incorporate since there's a lot of highs and lows in my gaming commentary and I want to create a good sound. Currently my process is: Noise Reduction - Vocal EQ - Normalize at 98.8% and ship. I'm curious if this could completely eliminate the need for normalization but would the new process look like: Noise Reduction - Vocal EQ - Match Volume - Hard Limiter. ?
Would love to hear your thoughts on this and thanks for the great tutorial!
Exactly, loudness normalize and then hard limit as the last two steps.
Hi Curtis I enjoy all your videos thank you for doing all of this... do you know how well this can be done with audacity or sound processors like reaper... more budget programs?
Maury O'Con Reaper has a plugin available called SWS that evidently measures integrated loudness and true peak. I haven't used it yet but it is worth a look. As far as I know, Audacity does not have any included plugins or features to measure integrated loudness or true peaks.
Awesome Stuff! Thanks for the help!
You're welcome and thanks for the feedback!
Could you do a video comparing Audacity to Adobe Audition? Obviously Audition is vastly superior, but for people like me who are considering stepping up, knowing the gains to stepping up to Audition will help me decide to make the step.
Hi, I discuss this some in sound post-processing course: school.learnlightandsound.com/p/sound-post-processing-1 I will also add to the list to cover in the future on UA-cam. Thanks!
Love your videos--thanks so much! Just tried this and wanted to mention that with the current CC version seems to use "true peak"--appears to default to -1db (at least it mentions the peak in terms of true peak)
Thanks! Yes, they updated that which is great news!
There's also the new Essential Sound Panel in the latest Premiere. Have you reviewed the process of loudness normalization using that and also parameters of the true peak limiter? Would you still apply compression for dialogue audio?
Hmm, your audio kind of looks like it has a DC offset (uneven distribution between waves under the baseline and over the baseline), which will limit your ability to normalize to some extent. Weird thing is, that it looks centered everywhere there is a pause. Maybe something you should troubleshot by replacing equipment (mic, preamp, soundcard) one at a time?
Also, I'm not sure I heard any difference at the point of the "true peak" that wasn't there after you applied the limiter. I heard something which could have been either analog distortion from loudness, or some temporary acoustic from back of your throat, either way, i percieved it as being there both before and after you applied the limiter.
I used to work as a studio engineer in a studio that primaraly recorded rock/metal music, and as a last step (while mastering) we always compressed, limited and normalized to 0 db. The "true peak" would have been lit constantly, which has been the db standard in rock/metal music CD:s since the 1990's at least. So this importance placed on "true peak" seems odd in respect with my experiences. Maybe true peak and the loudness level at -24 db is important when it comes to broadcasting and receiving via radiowaves so that the signal never is too hot for a radio receiver? (which would imply it is of less importance in purely digital situations such as youtube-videos).
Anyway, graphical representations is a great tool in many cases, but you should also trust your ears (and also take breaks to let your perception of sounds re-calibrate to non processed sounds).
Valpskott Thanks for your insights! I thought that I had a DC offset issue as well. But it turns out that in periods without signal/sound, the waveform returns to infinity and a DC offset repair tool didn't do anything. I learned that this is actually waveform asymmetry which is evidently a little more common with male vocals and some microphones. I now use a phase rotation plugin which corrects that issue. Here's a more recent episode that includes that: ua-cam.com/video/VcMOBrdbHgk/v-deo.html
I used to normalize to -0.1 dB as I learned from music production but it turns out that with digital, you can still get inter-sample peaks that exceed 0dB if you do that which is what dBTP is supposed to help measure and address.
This is great Curtis. Is this something that could be applied at the track level? Meaning, could I apply this technique to all of the dialogue tracks in a multi-track session?
Thanks Joe. I haven't done that to be honest, but as I'm think it through, the types of issues you're trying to solve will require different workflows. If the loudness of each of the dialogue tracks are where you want them, you could just take the master mix into waveform view and do the "Match Volume" from there. But you'll want to address any issues you have with the individual tracks first, perhaps even loudness normalize those that are significantly quieter than the others. This could be a good idea for a future episode.
***** Hey thanks, will need to give that a try...
Great series, very helpful! Quick question - to compensate for the true peak in the last step, why use a hard limiter as opposed to another pass of compression?
Thanks and good question. I is a judgement call. If there are just a few transients with a single peak each, I find that limiting doesn't leave audible artifacting. But you could certainly do another round of compression as well.
hi curtis ... which shot gun mic best cheapest one... please..
Hi Prakesh, The cheapest one I've used is the Takstar SGC-598 and it is surprisingly good for the price: ua-cam.com/video/Ecs0GWaPtto/v-deo.html
Can you go through your entire process? Because you say that you should clean up your audio before you do this part but Sometimes I just can't hear the background noise enough to edit it out before increasing the loudness.
I would just personally love to see your entire process. (If you already have a video up like that maybe i'll run across it watching through the playlist.)
+HollowProductions Here's one where I ran through the process. I only use noise reduction if I must so don't feel like you have to do that every time: ua-cam.com/video/VcMOBrdbHgk/v-deo.html
And here's one on noise reduction: ua-cam.com/video/1lIjFBG6w58/v-deo.html
And if you prefer Audacity, here's noise reduction in Audacity: ua-cam.com/video/vlJSTavIY3k/v-deo.html
I'm working on a course that covers dialogue audio post processing in depth so look for that some time in January. Best wishes!
thanks for the quick reply and i'll definitely check those videos out!
Very nice video, thank you :-) Is there a way to create "actions" for that, so that you might just mark the audio and start the recorded action, which automatically sets all these parameters? Just like in photoshop?
As far as I know Audition doesn't have actions like photoshop yet, but you can save presets for each of the plugins which helps a little--you'll see a little icon with an arrow pointing down at the top of each plugin/effect which allows you to save your setting for use in the future.
Hmmm, it looks like you can actually save normalization (with true peak limiting) in the encoder presets now so it is applied automatically when you render the video (with Adobe, anyway).
+noodohs yes! Makes it easier. If you're normalizing for web/mobile to -16 LUFS, you'll usually still need to do some compression first.
I always do, probably more than I should. I sure do like compression...
LOL!
Curtis,
so what's difference between normalize to -0.1db and Match volume?
Also, for the noise reducetion, do we do it before match volume or after OR even both before and after?
Jason, the biggest difference is where the "Perceived Loudness" parts of your audio end up in the final audio. You could technically end up in the same place by normalizing to -0.1db and then compress and peak normalize again, etc. The problem I've found with that approach is that it is too focused on peaks which do not, by themselves, define the perceived loudness of an audio clip. So using Match Volume/Loudness Normalization, you can get more consistent results. Secondly, peak limiting at -0.1db can still result in clipping so using "True Peak" limiting can solve that problem. As a workaround, I have been peak limiting to -2.5 db and have so far been able to avoid clipping.
Thanks a lot for your videos, they're very helpful! I have a question though, do you know what's the loudness standard for film and cinemas (does this apply)?
Also, I can't say I've searched enough in your channel but, have you done any videos on audio recording techniques for film (such as handling boom poles, managing cables, etc.)?, or, if you know, where can I get info on this? I'm studying at a film school but we haven't delved deep into sound. We use a tascam dr40, a 2 channel sound devices mixer, a rode ntg1 mic and sennheiser hd201 headphones, so I have a hard time with cables sometimes.
Hi Jorge, thanks! I don't believe that there is a published standard for loudness for cinema. I would probably go with -23 or -24 which are the loudness standards for TV broadcast in Europe and the US respectively. Cinemas usually have nice playback systems so using the TV loudness standards should work nicely and the nice thing about them is that they leave a bit more dynamic range than loudness normalizing for mobile and web. But if there's someone here who knows more, please feel free to comment!
As far as I understand, normalizing to -24 LUFS is quieter than to -16 LUFS, so there's more headroom for dynamics in cinema? I remember my experience of making DVDs. The Dolby Digital sound was much quieter than original WAV, because it used -27dB normalization at authoring, and then +27dB at playback. I didn't like that Dolby Digital codec is making my sound so quiet, so I combated that by setting that parameter to 0 dB and it sounded like original. Now thanks to your explanation of loudness standards I understand why DD normalized to -27dB. Do you think it's good that this is not a required parameter for other audio codecs like aac for video?
Great video. I am trying to follow your method but Audition won't let me drag the files to the match loudness window. I imagine it has something to do with the "supported files"part of it. I am using .wav files 16 bit 44.1 mono so I don't know what it can be. Perhaps the naming of the files that includes underscores and numbers? Have you encountered this problem?
Hmm, not sure. Might be best to contact Adobe support. Best wishes!
I just want to be clear. When I make a video I most often have multiple clips of audio possibly in different environments. would you suggest I bring the levels as close as I can by my own ear. Then, export one .wav file for processing "loudness and normalization". Or, execute this process for each individual clip. Why or Why not. Thanks
Rhett Brown at RetroWreck I process each file separately. There may be times, particularly when doing narrative work when you may want to go quieter than -24 LKFS which would be an argument for doing each separate. If, on the other hand, you're doing something where all the clips are supposed to take place in the same environment, you may want to do them all together. In either case, having the ability to standardize your perceived loudness lets you make these types of choices with more confidence.
Thanks, I will try it both ways and see how it works.
I had no problem using that Single Band Compressor with the 2015 (?) version of Audition, but just tried it with the new version for 2017, and can't get it to work. I mean, there is no change in the peaks of the audio file after Apply.
+George Taylor hi George. What do you have the ratio set to? And you have confirmed that you have set the threshold lower than the peaks? Also, attack is set to less than 10ms?
+Curtis Judd Okay. I think I just found my mistake. It looks like it was a threshold problem, after all.
I do appreciate your response, as well as all your instructional videos. They've been a big help to a lot of us.
+George Taylor 👍
Can we just do thatin Premiere Pro ? Also, what if the audio file is already inside a the video file? Also what if we want to mix both external audio file + internal audio file in the video? Thanks
Hi Alex, yes, of course: ua-cam.com/video/RrRn8stkEBs/v-deo.html
Mix all your audio then at export, set the loudness target. What this does make more difficult is that if you need to compress your audio to hit your loudness target, it is a fair bit more difficult to do in Premiere. But give it a try shown with the linked video and see how it works out. Good luck!
Curtis Judd will try it today. Thanks a lot for the swift replay!
Limiting an already clipped file.. wouldn't that just hide a problem, without solving it?
Absolutely correct. If you have clipping, first need to repair that, then compress or limit.
I have a question....Im putting up clips of music i did for a film, and im wondering if I should normalize the audio or leave it "as is." It is a bit softer than other music vids Ive seen in the same genre....Here is one of the vids so you can see what Im doing: ua-cam.com/video/ZiWA6CRMJKM/v-deo.html
Appreciate your help =)
Hi Feona, I think it depends on where it will be "broadcast" and it is probably not unusual to have a few versions. For anything that will be listened to via mobile devices and web, you'll want to consider -16 LUFS for stereo tracks (-19 for mono). For TV broadcast, -23 LUFS in Europe, -24 in the US. I'm not so sure about films that will be projected in a proper theater, however. Perhaps Paul Figgiani would have some insight on that.
Feona, I would say your responsibility as a composer would be to produce without shooting for a specific loudness target. Ultimately this will be up to the people in post. They will mix accordingly and produce a final master that will be (in terms of loudness) adjusted based on the delivery platform.
Of course most composers/mastering engineers have personal preferences when building a piece. Allot comes into play - things like dynamic range, impact, etc. It’s pretty much subjective.
I analyzed your track. It looks like you may have Peak Normalized to 0dBFS? I wouldn’t recommended that. You have some ISP’s (Intersample Peaks) occurring above full scale.
Take a look at this high-res snapshot: f-video.s3.amazonaws.com/fionna.png … the top is your original source lifted from UA-cam. In the lower version I applied slight compression just to tame the wayward transients. This allowed me to limit peaks to -1.0 dBTP without any sound quality degradation.
In the future I would suggest you keep things well below full scale. Also note that peak amplitude has nothing to do with perceived (average) loudness. It represents the proportional voltage of an audio signal. If you are looking to create mixes that vary perceptually, you will need to look into other descriptors (Program Loudness, Loudness Range, etc.).
Paul Figgiani Thanks!
Hi Curtis, You should look at Tone Boosters TB Barricade and TB EBULoudness. TB EBULoudness is much usable than Loudness Radar.
www.toneboosters.com/tb-barricade/
www.toneboosters.com/tb-ebuloudness/
Thanks for the tip--those look great and reasonably priced, too!
docstore.mik.ua/orelly/web2/audio/ch04_03.htm Here they say the threshold should be at 3 to 9dB
Thanks so much for your video !
Thanks! Interesting article. I think the intent is to compress the peaks by that amount as every recording will need tweaks to the threshold setting to get that much compression.
your tout are bad cuz you are not explaining all things you doing on your audio
What is a tout?