How to you calibrate the sound?

Alok-Sharma wrote on 6/23/2021, 11:49 PM

Hi,

In this comment https://www.vegascreativesoftware.info/us/forum/how-to-control-the-volume-when-playing-the-video-from-the-timeline--129991/#ca808398, @3POINT mentioned a very good point which I was actually curious about.

Adjusting volume for monitoring should be done at the end of the chain and not at the beginning or somewhere between. I control the volume directly at my monitor speakers, all other volume controls before stay at 0 dB.

Just like we calibrate our computer monitor for colour precision, I am curious to know how sound is calibrated. I will try to explain my query as good as I can.

Let's say a customer has sent me a video that needs to be edited.

Now on my computer, I would be having different sound settings like volume levels apart from the fact that I am using Logitech Z506 Home Audio Speaker which produces a magnificent & superb sound especially the bass. Even though it is 5.1, but I am using it as 2.1 as my laptop supports only stereo.

Now, when I do the editing and deliver the final video to the customer, the audio & other sound levels might not be the same. For, e.g., the audio is playing loud enough on my computer but too low on the customer's end. Besides, the customer would be having a different setup like ordinary speakers, a budget sound system or even a home theatre. So it's quite obvious that there would be a significant difference in sound at both the ends. In short, what I hear, the customer might not hear the same.

So there ought to be some way to standardize the audio levels so that it follows the global standard, i.e. not too low and not too high, just the right precise level.

I am curious to know what settings I need to maintain at my end so that when I play the video, I hear exactly what was originally recorded, i.e. bypass all the extra & additional sound processing or effects, just the naturally recorded sound. Post that, I can do whatever needed with the sound.

Besides, what external budget hardware you would recommend for my old Dell Inspiron 3542 laptop.

However, once I build my own desktop, I would have a plethora of options and combinations to choose from.

Comments

Dexcon wrote on 6/24/2021, 1:44 AM

So there ought to be some way to standardize the audio levels so that it follows the global standard, i.e. not too low and not too high, just the right precise level.

I am not so sure that there is a 'standard' per se. Sound on Sound magazine in association with Steinberg produced a series of 6 videos on Mastering techniques presented by mastering engineer Ian Shepherd. The first one is

They are all on YouTube.

I've only watched the first two, and my basic take-away from those 2 eps is that mastering is very much a talent built on self-development and experience - there's not a rule book to go by.

With music, one of Ian Shepherd's suggestions was to take one of your favorite recordings and use that as a basis to master your own music project.

Another suggestion - with wider application in that it can apply to any audio recording when being mastered - was to have a range of audio appliances on hand when mastering so that you can not only master by the recoding studio's high-end speakers, but also preview the 'sound' in domestic headphones and ear buds, TV or computer speakers and so on and so on.

With this in mind, I would suggest that if you are producing audio for a specific client, it would be a great help to find out what type of audio system would mostly be used by the client to play the project on, and make sure that you have something similar on which to preview the audio mix when mastering.

Based on mixing for the destination medium, pop music and radio advertisement tracks used to sometimes be specially mastered favouring the treble frequencies if the end source for that version of the recording was to be AM radio.

Volume is yet another thing. Have a look at the waveform for many modern pop songs on a CD. The amplitude of the waveform can be right up to the the vertical limits of the waveform which probably explains why they often sound a bit distorted due to clipping. Though the settings for 'normalising' an audio track are often adjustable, many mention that the peaks of normalisation shouldn't be less than minus 1dB. But generally I keep the peaks for original dialogue recording at around minus 2.5 dB though the level after mixing would likely end up to be different.

Another variable is the environment in which the audio is to be played. If its going to be in a large box-shaped hall without sound management like thick curtains or sound baffling, then the standing waves are likely to be a significant problem, likely exaggerating the bass frequencies making for an unpleasant boomy sound.

If your client requires a broadcast standard audio track (e.g. BS.1770-1 2, 3 or 4, EBU R128, OP-59, etc) then you'd need to establish which standard is required and do a bit of research online to find out the specifications for that standard. Unfortunately, the standards document is likely to be highly detailed and technical. For example, BS.1770-4 https://www.itu.int/dms_pubrec/itu-r/rec/bs/R-REC-BS.1770-4-201510-I!!PDF-E.pdf

Hopefully, @rraud would be able to provide some advice here as he has great experience in the audio world.

3POINT wrote on 6/24/2021, 2:14 AM

Audio should never exceed 0 dB, otherwise you well hear distortion (clipping). When you normalize audioclips, the loudest part in the clip will be set to 0 dB. This works fine when there is just one audiotrack in your project. But as soon when there are two or even more audiotracks added (music+voice over) and all of them are normalized to 0 dB, the total output will be +9 dB (+3 dB for each audiotrack) = clipping. To avoid you can set the Master Volume to -9 dB or set the trackvolumes of each track at -9 dB. Another possibility is to normalize each track to -9 dB instead of 0 dB by setting in the audio preferences:

To be on the safe side, I never put the total output at 0 dB but at -3 dB. Working this way, my audio has always the same volume everywhere. The end volume is controlled by the volume control of the player/amplifier/TVset.

Alok-Sharma wrote on 6/26/2021, 8:41 AM

Thanks @Dexcon for the detailed information. I will watch those 6 videos one at a time and at a slow pace. Since everything is fairly new for me, I do get curious and sometimes overexcited.

Thanks @3POINT for the awesome tips. Until this date, I hardly knew anything about the sound level, especially the fact that if there are more than one audio tracks, the output will keep adding. I will definitely keep this in mind.

Dexcon wrote on 6/26/2021, 8:49 AM

@Alok-Sharma ... the 3rd Sound on Sound YT tutorial is about 'how loud should you master?'.

I do get curious and sometimes overexcited.

The learning curve never goes away no matter how long you've been doing it.

Musicvid wrote on 6/26/2021, 8:56 AM

There is a big difference between maximizing audio without distortion at the desktop, and loudness leveling for delivery.

Peak levels for hard media delivery are usually ≤ -1dBFS

Audio for actual broadcast must conform to US and European standards, the former being slightly more conservative. Vegas Loudness Meters are the right tool for this.

Audio for internet delivery "should" be no more than 6 - 8dB louder than broadcast levels, and most social media and Youtube penalize anything louder than than about -15 LKFS. https://www.loudnesspenalty.com/#

Search terms:

EBU R128

ATSC A/85

rraud wrote on 6/26/2021, 10:49 AM

Most music streaming hosts (Spotify, TIDAL, iTunes, Amazon Music, Pandora, ect) want the loudness at -14 LUFS ('Integrated' is the key LU factor), or thereabouts. Many streaming hosts re-encode and change it if it is not within spec. YouTube re-encodes submissions regardless and will lower the level if it louder than -14 LUFS.. but will not raise it if it is low.
LUFS is not to be confused with dBFS

Alok-Sharma wrote on 7/5/2021, 2:54 AM

Thanks @Musicvid & @rraud

I agree all this is going above my head but eventually, I will start understanding them once I get more deeper into all this.

Howard-Vigorita wrote on 7/9/2021, 12:02 AM

Personally, I do the opposite of what is mentioned in the opening post... I do my audio mix first. I like doing that in Vegas with my A-cam loaded so I can see who's doing what. But use it's soundtrack mostly as a timing reference. I usually put a volume envelope on it and knock it down to about -15 db and only bring it up for applause or audience reactions. The bulk of my audio content usually comes from an external flash disk recorder. Usually either a Sound Design 744T or a 788T depending on the number of sources. Theatrical productions usually call for a full scale mixer with capture to either an internal usb or a computer. Once I have the mix semi-finalized I create a mixdown track aligned with my a-cam and later bring in any other cams for alignment, muting all the camera audio tracks.

Calibrating the sound means different things to different people. Shaping a soundtrack for delivery is often the job of a mixing engineer and sometimes a mastering engineer. All the calibration before they start their work is to the physical space and the playback system so that they can rely on what they hear. Calibration of the content, as has been mentioned earlier, usually involves adjustments like eq, volume, panning, dynamic range, and reverb. Taking care not to exceed 0 db. More esoteric adjustments are usually reserved to a mastering engineer... things like manipulating phase relationships and pre-delays. These things allow depth positioning in a soundstage and require specialized equipment like a Neve Masterpiece. An advanced reverb unit like a Bricasti can do some of these things though allot of folks are using impulse modeling plugins to roughly simulate similar spacial manipulation.

Vegas is pretty good at doing the things that a mixing engineer does. You just really need to know the customs of the genre and your delivery target. Avoid mixing with headphones, particularly with vocal music. If delivering for mastering, they want no compression on the mains at all. Classical music targets usually prefer wider dynamic ranges... rms values between -20 and -30db, usually accomplished with a-b spaced omnis at a distance. In Rock, rms values around -15 or lower are common. Jazz often prefers close and intimate perspectives with medium compression... usually starting with closer mic placement. If your deliverable is for theatrical application, they want their dialog more forward for intelligibility... Vegas volume envelopes are an easy way mark dialog lines and lift them. Delivery for tv broadcast is more concerned with transmitter over-modulation... they require peaks to be -6 db max... the Vegas WaveHammer audio fx provides an easy way to conform to that.

Alok-Sharma wrote on 7/9/2021, 8:47 AM

Thank you @Howard-Vigorita for the detailed explanation.

Every information shared by the community members really means a lot to me, even though sometimes I find it a little difficult to understand. So I look up the terms on the internet to know their meaning.

Webster-Struthers wrote on 10/29/2024, 4:39 PM

I am trying to master a 10 song collection I have produced on Vegas 21, and trying to balance the loudness from track to track. Using the loudness meter with conjunction with the master volume seems to have no effect. How can I use a reference track, and adjust the loudness from track to track ?

john_dennis wrote on 10/29/2024, 5:10 PM

@Webster-Struthers said: "How can I use a reference track, and adjust the loudness from track to track?"

Tools / Generate Loudness Log.

At the end of the log are these data. Match Integrated while keeping the other two measures under control.

Results:

Mom. (max):      -7.06 (LUFS) at 00:04:33.237
Short (max):      -8.74 (LUFS) at 00:06:16.832
Integrated:     -14.63 (LUFS)

john_dennis wrote on 10/29/2024, 5:15 PM

I often do quick sound leveling first after a shoot. I sometimes do a batch render to watch all the clips before I cut them. When viewers hear a radical change in loudness around my house they let me know right away. Less so with differences in video.

Webster-Struthers wrote on 10/29/2024, 6:20 PM

I understand the concept of matching the integrated level. I just can't do it. Where do you make that adjustment ?

john_dennis wrote on 10/29/2024, 9:24 PM

@Webster-Struthers said: "Where do you make that adjustment?"

The simplest example. You want to upload to youtube and want to have Integrated less than or equal to -14 (LUFS). Assuming the peaks are well controlled, you could use the Volume fX on the audio track to make the adjustment.

It gets more and more complicated from there, but I'm going to eat dinner.

john_dennis wrote on 10/29/2024, 11:16 PM

There is a lot of information in this long thread:

https://www.vegascreativesoftware.info/us/forum/broadcast-delivery-with-vegas-pro--140074/?page=1

Webster-Struthers wrote on 10/30/2024, 9:23 AM

Thank you, John. So, after running the LUF log on a rendered audio file, you can literally change the loudness with the volume FX. So, if I was was shooting for -14.00, I could literally adjust the loudness of every single track to make them match ? That sounds so simple, yet so brilliant. Then I could use the CD layout tool, and render a complete file to burn cds from ? Am I missing something here?

rraud wrote on 10/30/2024, 9:32 AM

Sound Forge Pro 18 has loudness normalization,(integrated LU) unfortunately it does not have the most used -14 LUFS, but a 9dB makeup boost be added after normalizing to -23 LU (EBU)..
That said, normalization is not a substitution for overall compression, EQ, ect. (mastering).

ChrisD wrote on 10/30/2024, 9:47 AM

So, if I was was shooting for -14.00, I could literally adjust the loudness of every single track to make them match ? That sounds so simple, yet so brilliant.

It can be a bit of a black art, plus depending on how many tracks you have that may or may not have audio peaks and valleys, you could put the FX on the master output. Depending on your preferred workflow.

Perhaps this is a more concise thread.

Webster-Struthers wrote on 10/30/2024, 10:44 AM

Thank you, one and all. I am going back to the DAW to see what I can do, but I now understand how to adjust the loudness using the Integrated LFU's from the log in conjunction with the volume FX. The music is an acoustic trio with three part harmony, so there are some significant peaks and valleys. I can live with that as long as the loudness is perceived to be about the same on all the tracks. Thanks again.

Howard-Vigorita wrote on 10/30/2024, 4:20 PM

@Webster-Struthers Here's how I do concerts. Little harder with multiple musicians who tend to perform with their own touch and style.

I put the entire concert into a single project and mark off the songs as named regions. I put allot of detail in the region names so it passes through batch renders later when uploaded to YouTube. Then I render it from Vegas in Sony w64 format with the render option to save markers. Then pull the render up in Sound Forge and run Tools, Statistics on each region and note the reading in a spreadsheet. Current versions of SF report both rms and lufs. You can use either as long and you know the targets to aim for... I use rms but find it helpful to look at both. I also note the max volume level so I know how much boost a track can take without overloading. Some have mentioned getting similar info directly from a Vegas loudness log, but I'm accustomed to SF.

I start the spreadsheet by copying the region names from the Vegas Edit Details pane and pasting into Excel to get a track sheet. I try to maintain the dynamic relationship between multiple pieces from the same performer so that the pieces they intend to be quiet are not louder than their other pieces. With different performers appearing separately, I further adjust to try keeping the softer and stronger pieces in roughly the same ballparks. This concert was further complicated by brass which is especially strong. Going back to Vegas, I usually throw a volume envelope on the master bus to lift or cut the pieces that need it. If something needs a lift above 0db, that requires compression. The Track Compressor FX can be automated on the master bus so it only affects specific pieces. If a song is too low and needs a boost more than 3 or 4 db, expansion usually sounds better by not noticeably altering a song's dynamics or lifting the noise floor... Graphic Dynamics FX can be automated to do either compression or expansion but is a challenge to tweak. Here's a Vegas project I'm working on now: