Sound levels - speaking & music

[Rob Harlan]

How far up the dB scale are you peaking your sound for ceremony & speeches?

Do you try to ensure that the spoken words in ceremony & speeches peak at around the same line as music does in music-driven parts? Or are you keeping one lower than the other?

Obviously I stay well away from 0db in every case to avoid any distortion, but I was wondering - particularly from sound experts - how much lower than 0db spoken words and music are best limited to.

[Don Bloom]

I try to keep peaks to no more than -12db but if it gets a little over I can live with it and can generally take care of it in post with compression and EQ.
-12 in my mind is the magic number.
Others may say different YMMV

[Jeff Brewer]

I also like -12db as my midpoint. Of course people don't consistently speak at the same level, so it may jump to -6db or drop to -18db, but my average is at -12db and I love it there.

[Philip Howells]

Rob unless you have proper meters this is rather a futile exercise. Jeff thinks he can detect a variation of 12dB - that's means the quietest sounds are about one sixteenth the intensity of the loudest. I don't think so. On stage delivery by a Shakespearean actor might approach that range but not normal speech. What I think he's seeing is the lag and distortion in whatever on screen metering his set up has.

Whilst it is true that it's good practice to keep the average peaks at -12dB to -10dB on NLE readouts to cater for the inaccuracy of the readouts and transient peaks above that and of course never to permit any clipping, the balance between music and voice will depend on the type of music, the type of voice and the intensity of both.

Such balancing is best done with the ear, listening to loudspeakers not headphones at the levels the programme will eventually be played at - because relative balance varies according to the overall loudness. Apart from avoiding clipping, about which you obviously already know, slavish adherence to any metering, no matter how accurate, isn't the way to go.

Having spent many hours mixing voice and music, I also think that NLE systems are particularly ill-suited to such mixing - in my view riding the faders is the best way and most NLE systems only allow you to set the faders not ride them. One day I'd like to be making enough out of weddings to buy a decent audio control surface for my computer.

[Rob Harlan]

I record at a suitable level for the situation, be it ceremony, speeches etc., and don't usually break my back each time for an absolutely perfect recording level per-se. So long as the mics are picking up clearly, not too quiet (hiss to deal with) and not too loud (irrepairable clipping), I simply balance all together in post-production (utilising compression if necessary or advantageous).

Was really just wondering about the levels applied to the different sections of the wedding video during editing - i.e., say the opening montage track bobs around the -12db level when the song is at its loudest, would you then aim to have the vocals of the ceremony/speeches peaking at -12db also, or make a point of increasing or decreasing the loudest peaks of voices from that?

I agree with the ear (and speakers rather than headphones) being the best way to judge the overall 'loudness' impression.

[Philip Howells]

Rob, I thought you were referring to levels in the mix and on the final DVD. If you're talking about the levels for actual recording, my mentor and former sound engineer (former broadcast sound engineer) sent us tone from the SQN which we set to -18dB on the cameras (used to be -4dB for analogue cameras back in the Beta-SP days).

I'm surprised you're worried about hiss on a digital recording - the system noise level is virtually non-existent.

As to your second paragraph the answer is yes, provided it's genuine peaking and not you trying to do a manual compression by raising the quiet sections of music/speech so that they also peak at -12dB. If it's quiet then record it quiet.

[Allan Black]

I don't do weddings but after 30yrs with everything else I know that first time through everyone only listens to themselves off the DVD .. and that's how they judge it.

So regardless of what you listen to mix it on, a sample DVD played on a variety of sets is the test .. with specific attention being paid to whoever is paying the cheque.

Cheers.

[Rob Harlan]

To be clear my initial question was about the final sound mix of speaking and music on the DVD - i.e. do both speaking and music sections come out about the same loudness to your ears, or is there ever any need to reach for volume adjustment throughout the video.

About actual recording levels though, of course if we set those levels too low on the day then the audio must be raised in post-production = introduction and amplification of background hiss.

[Chris Harding]

Hi Rob

In the old days of analogue voice and music mixing the absolute maximum level that music was supposed to reach was said to be just 40% of the voice level. If I am doing something like a bridal arrival and the bride has asked for a special song during that segment AND still wants to have ambient audio I will on average pull the music track down at least -9db or -12db lower than the live track. However this CAN of course be way too loud where the music track is dominant (a soft solo piano track is going to be a lot less intrusive to the live track than a heavy metal track!!)

I think the real answer hear is just listen to it and decide yourself if the music track is supplementing the live track or sometimes over-riding it??? However in general with Vegas, I will drop in a volume envelope on my music track and initially pull it down to -10db as a starting point.

Chris

[Philip Howells]

Originally Posted by Rob Harlan

snipped

About actual recording levels though, of course if we set those levels too low on the day then the audio must be raised in post-production = introduction and amplification of background hiss.

If you have so much hiss on a digital recording then you either recorded it or there's something wrong with your system. The thing about digital recording is that there is very little "noise" so you can raise levels without seriously degrading the signal.

[Rob Harlan]

Not so much hiss, just more hiss than if the audio recording levels were correctly set high enough to not have to raise much in post-production. If there was not enough difference to be concerned about, then maybe everybody would just record audio with levels set low rather than aim for higher around the middle.

I have reasonable mics, though not ultra expensive. If a recording has been made on the low side then hiss, while not necessary serious (depends on each person's standards I suppose), becomes more present upon raising the audio to ideal levels. It's fairly easily dealt with with noise reduction filtering anyway.