100%

I just found this looking for a way to not be whispered to then shouted at while listening to my flash briefing. Has there been any movement on this? I have a Gen 1 echo.

I frequently need to increase or decrease volume across the different news sources in a Flash News briefing. It would be great if Echo devices had a volume leveling similar to Amazon Music that can level the volume between commmercials and regular programming.

Amazon Music when played on an Echo device also desperately needs volume leveling. Volume leveling is available in the Amazon Music app, but it is really needed on Echo devices.

I do not want to switch to another product but constantly saying Alexa Volume up, Alexa Volume down is getting ridiculous. LEVEL all audio to output the same or I see many leaving the echo brand. Dot is not enjoyable when I have to say up and down every song...

BBC news is the worst for this. It was fixed at one point but it’s gone back to what it was. Alexa pretty much shouts at me then BBC opens up with the news and I can just about hear it.

Right now if someone says "Alexa turn up the volume on my TV/Speaker/Ect" when we get the directive https://developer.amazon.com/docs/device-apis/alexa-stepspeaker.html#adjustvolume

it just includes the payload with the property volumeSteps and defaults to 10 if not specified.

The problem with this is that you have chosen the default step size for us but not let us know that you have done this. My speaker has a volume from 0-25 so adjusting the volume by 10 causes a huge increase people are not expecting. But since you have not told me that they didn't specify a volume i don't know whether to increase it by 10 like you say or if a user has just said increase the volume, in which case i would like to increase it by 4 or 5.

My solution is to either allow us a way to specify what we want to be the default volume step size when we discover a device, or add a boolean property to adjustvolume telling us whether or not a volume was specified at all.

The Alexa hardware team would need to implement an Automatic-Gain-Control, then Dynamics processor pair of dsp functions to accomplish volume 'leveling.'

That function would have to then be un-activated for text based updates, reactivated for an audio brief, un-activated for the next text based update but still keep the previous audio levels, then carry those settings forward for the next audio brief's amplitude...etc.

If you need a band-aid solution in the meantime, you can use a compressor on your content to reduce the peaks gain by 3-6 decibels, then normalize the audio with make-up gain so the newly reduced peaks are around -4 decibels on a digital scale.

Those numbers should help reduce any deficits with a skill-dev's levels, but keep in mind that professional audio engineers do things that cannot be easily explained or taught on forums, but many studios and online audio freelancers should be able to process your files for you.

For a technical response only; 'levels' aren't always relevant more than 'density' (dynamic range of content), for example, a McDonald's radio commercial may only have 1 decibel of difference between its loudest and quietest audio peaks, but where an amateur commercial may have 20 to 30 decibels of difference.

If a skill-dev wants to truly give themselves a reference they can work toward, I'd recommend the following test:

1) activate a number of broad-scale flash briefing skills (major networks etc)
2) plug an audio cable from your device into the line input of an audio card
3) set the Amazon device's output moderately high, but not clipping or 'in the red' on your
computer.
3) record a full volley of audio from those skills
4) normalize the recorded audio to -4 decibels or so
5) save the file as a reference for later
6) compare and contrast

Hopefully this is helpful to some devs.