Full On Cues?

[Audio]

Hi,

We are using SPL with BBP and are having an issue with the volume level of our cues. Indeed new tracks following a jingle sound like they are not starting at a 100% volume (although they are set to). It sounds like the start of the songs is ducked for some reasons… I was wondering if you’d have any idea as to what this could be linked to and how we could fix this.

One avenue is, that looking at the files’ waveforms, our jingles levels seem higher than the songs’. All files are .wav files. I am assuming there is too little headroom left for the song following the jingle to be heard fully at the start, although we’ve pushed the segue point on the jingles as far as we could (fading out).

One option would be lower the jingles level, but how do I go about doing this without messing up the quality? Neither our music or our jingles are compressed so we wouldn’t go mess up with that aspect..

Your thoughts?

Thanks!

Audio

[JesseG]

[quote author=”Audio”]One option would be lower the jingles level, but how do I go about doing this without messing up the quality?[/quote]

That’s my thought. To do it with high quality… would be to convert it to 24 or 32 bits if it’s 16 bits right now… lower the levels… then dither back to 24 or 16 bit if you have to. Minimum quality lost, and most likely an inaudible change. Other than the music being at the same loudness again. 🙂

Cheers,
-J

[Guillou]

Lower jingles would be a great option because we use SPL + BBP and we have no problem.

BBP is excellent !!! 🙂

[Audio]

Thanks Jesse!

I had a further look at various music tracks and jingles earlier and noticed that peak amplitudes are all over the place, anywhere between 0 dB and -10 dB… Since we are going to be normalizing everything to the same level, what would you say would be the ideal level to use ahead of BBP?

Regarding quality, you mentioned converting everything to 32 bits ahead of the normalization, which software would you recommend for us to do so?

Thanks again!

Audio

[RodeoJack]

Just an addition to what Jesse’s already contributed here:

In the broadcast environments I’ve been in, the common procedure would be to start your jingles at, or slightly above the level of the preceding cut. The succeeding song should be above the level of the jingle’s end, assuming you want that song to dominate. The idea is to make each cut force your processing into increased compression, which puts the incoming cut on top and forces the outgoing cut down at the ratio of the increased compression.

This is a nifty trick, which works well in a live studio environment. It can also be done with automation, though you have to plan your system carefully to pull it off well.

On the other hand, if you notice that BBP has to increase levels between the end of your jingle and the start of your next song, you’ll get the effect you mention. The jingle has pushed the processor into higher compression than is necessary for the song, so there’s a small period where the processor has to catch up. It’s the reverse of what you want to see the processor doing.

[Audio]

[quote author=”RodeoJack”]Just an addition to what Jesse’s already contributed here:
This is a nifty trick, which works well in a live studio environment. It can also be done with automation, though you have to plan your system carefully to pull it off well.[/quote]

Thanks for the input! Very interesting indeed!

We are running mostly in an automated mode. How would you set this for automation? Would it be sufficient to set SPL to operate a light fade before the mix point or would we have to prep our wav files in some ways?

Thanks

Audio

[timmywa]

Someone correct me, but what he’s describing sounds like the processor’s AGC is lowering the output during the jingles (ducking) and once the next element comes on that is at a much lower volume, the AGC has to try and bring back the audio to the desired levels. If the processor’s ultimate goal is to provide a consistent sound and your track to track volume is so different, you’ll get that response on all but some of the fastest and loudest presets, like New York or Eruption, etc.

When I do live shows, I don’t use a mixer so I just amp my mic up quite a bit so that it ducks the music in the background. Sloppy? Yes. But with my available equipment, it does ok. I’ve had to experiment where to set the mic’s volume so that it doesn’t totally obliterate the music so that when I "hit the post" the processor doesn’t have to recover and the opening vocals are still drowned out.

Hope this helps.

[jameskuzman]

[quote author=”timmywa”]Someone correct me, but what he’s describing sounds like the processor’s AGC is lowering the output during the jingles (ducking) and once the next element comes on that is at a much lower volume, the AGC has to try and bring back the audio to the desired levels.[/quote]

That’s how I’m interpreting the issue as well.

It sounds like the jingles are at a higher level than the music, so the processor is being driven further into compression to lower the levels; then, when the music starts, the AGC has to release a significant amount of gain to bring up the level of the music.

Other than increasing the "Speed" to accelerate the release time – which would, I assume, also change attack times, ratios, and other parameters, and in turn change the texture of the processing – I think lowering the levels of jingle audio files as Jesse described is the way to go. A bit labor intensive perhaps (but how many jingle files could there be?) but well worth the time investment.

Side note/wish list/semi-whiny/on the soapbox comment: If there was a version of Breakaway that un-bundled the parameters adjusted in a single slider, the end user would have significantly more control over the sound and in cases like this might be able to tweak the release of the AGC to help the processor handle this situation better. I know, I know, it might not make good business sense for the price, but given the quality of the Breakaway lineup I’ll bet there are people who would pay more in exchange for more control.

Off my soapbox now 🙂

Jim

[Audio]

We will amend the jingles levels, but the dilemma we have now is to which level (?) since we have uncompressed musical tracks at all kinds at level as well.

Upon checking, our jingles are sitting at around 0 dB and so are some of our tracks (say Oasis) but we also have some tracks sitting closer to -10 dB (some Coldplay, Evanescence, etc.).

So which dB level would you recommend we normalize our complete library to? The reason I am asking is that I am unsure whether bringing up the levels on the -10dB tracks will hurt the audio quality of these tracks? To which level would you recommend we normalize our library to?

Thanks,

Audio

[jameskuzman]

Hi Audio,

May I ask you to clarify something?

In your initial post, you mention the problem when transitioning from a jingle to a song, but you didn’t say you were having a problem between two songs.

I ask only because in your last post, you mention that some of your song levels are significantly different from one another as well, and question whether you should start working on your entire library in some way.

If you’re not having song-to-song issues, I wouldn’t touch the music library regardless of what the levels are. Yes, levels are concrete, tangible points of reference, but part of the job of the processor is to allow you to send audio of different levels in and compensate for that. I would trust your ears rather than get too caught up in what a .wav form looks like or where it peaks. If you’re not hearing song-to-song discrepancies similar to the jingle-to-song differences, I’d trust your ears, focus on "fixing" the jingles, and leave the music alone.

If you’re hearing the same issue between songs, that’s a different matter altogether.

The other thing to consider – and I hope I’m not opening up some huge can of worms here – is where the "power" is in the jingles.

In other words, if the jingles are "dense" and posess high RMS power, they could – depending upon how the preset (and the processing core at large, for that matter) interprets that quality – handle the audio differently than they would a more dynamic audio source with a higher peak level.

If you want to experiment with how significant the "texture" of the recording is to a processor, open the .wav editor of your choice and load up a song mastered within the last couple of years, and then open something from, say, Steely Dan. Study the visual .wav files. Look at the peak levels. Then run them through Breakaway and observe the behavior of the AGC, the multi-band compressors, and in particular, the peak limiters. Even if the peaks are hitting the same levels, you’ll observe drastically different behavior on the processor’s meters…

Jim

[timmywa]

[quote author=”Audio”]We will amend the jingles levels, but the dilemma we have now is to which level (?) since we have uncompressed musical tracks at all kinds at level as well.

Upon checking, our jingles are sitting at around 0 dB and so are some of our tracks (say Oasis) but we also have some tracks sitting closer to -10 dB (some Coldplay, Evanescence, etc.).

So which dB level would you recommend we normalize our complete library to? The reason I am asking is that I am unsure whether bringing up the levels on the -10dB tracks will hurt the audio quality of these tracks? To which level would you recommend we normalize our library to?

Thanks,

Audio[/quote]

Assuming your uncompressed tracks are .WAV, you can find a tool called wavegain ( http://www.rarewares.org/others.php ) and use it to bring all your audio to the same level. I’ve not used wavegain, but it’s partner, mp3gain has a default of -89db. Jesse recommended I lower that to -85db for what would be industry standard. If wavegain is the same, then I would recommend levels like I mentioned. You don’t need to be anywhere near 0db, as you wouldn’t have any headroom.

Hope that helps.

[Audio]

[quote author=”jameskuzman”]In your initial post, you mention the problem when transitioning from a jingle to a song, but you didn’t say you were having a problem between two songs. [/quote]

Jim,
I actually only noticed the different songs level looking into this matter earlier today, but the problem is indeed with "loud" jingles preceding "lower level" songs. And since we play a jingle between each song, that’s the problem! What the different songs’ level does explain though is why this "ducking" is not taking place in each instances.

You are making a good point regarding the jingles’ RMS power values, I will have to dig into this a bit more and test different settings against some low level songs. You are correct BBP is able to cope with the different songs level quite well, usually.

I can not help thinking however that bringing all audio to similar levels would smooth things out.

Audio

[Audio]

[quote author=”timmywa”]Assuming your uncompressed tracks are .WAV, you can find a tool called wavegain ( http://www.rarewares.org/others.php ) and use it to bring all your audio to the same level. I’ve not used wavegain, but it’s partner, mp3gain has a default of -89db. Jesse recommended I lower that to -85db for what would be industry standard. If wavegain is the same, then I would recommend levels like I mentioned. You don’t need to be anywhere near 0db, as you wouldn’t have any headroom.

Hope that helps.[/quote]

Thanks Timmy, that should help a great deal, I’ll look into it 🙂

Audio

[jameskuzman]

Thanks for the clarification, Audio!

I would suggest that you begin with adjusting only the levels of the jingles downward as Jesse and Timmy have described here.

See if that takes care of the problem first.

If my "theory" that the jingles have substantially higher average levels (not just peak levels) this might go far enough to address the issue without having to do anything with the music library.

Just my two cents 🙂 Keep us posted though?

Jim

[JesseG]

[quote author=”Audio”]Since we are going to be normalizing everything to the same level, what would you say would be the ideal level to use ahead of BBP?[/quote]
The peak level tells you nothing about the actual loudness though, and it will encourage a "loudness skirmish" within your own plant and production studio.

How about getting the Breakaway RTA (check latest software topic) and using dB LKFS to set where your actual loudness should be at. -23 dB LKFS is pretty close (within 1dB) to where the EBU P/LOUD group’s R128 draft is having the new international standard for where reference loudness at. The idea for TV broadcasters at least is that programs can be mixed at that loudness level, and put straight onto the air without any audio processing at all. 🙂

But even if someone will be processing for loudness at the end, the idea itself still benefits everything before then greatly. For instance, one could reduce the range of the AGC/Multiband so that stuff that’s more quiet stays more quiet.

Anyways yeah… Breakaway RTA, check it out. 🙂 Mainly for the level meter. You can resize it skinny so that’s all that shows, if you want.

[quote author=”Audio”]you mentioned converting everything to 32 bits ahead of the normalization, which software would you recommend for us to do so?[/quote]
Whatever software you’re using to normalize aught to be able to do this (including a decent dither back to 24 or 16 bits depending on what your playout software is compatible with), or it’s probably not very trust worthy software to begin with.

[Author]

Posts