Automatic Weather Forecast for Radio

[NoiseandPatterns]

Hi everyone,

I wanted to share something we’ve been working on at Noise&Patterns.

A while ago, one of the radio stations asked us to build a system that automatically generates weather forecasts as ready-to-air audio files — complete with a professional voice and a jingle.

The project turned out to work really well, and Radio Kwidzyn has been using it daily. That success encouraged us to make the tool available to other stations too.

This post isn’t meant as an advertisement - we’re genuinely looking for feedback from radio professionals:

• Could a tool like this actually help ease the newsroom workload?
• How do you feel about integrating AI into on-air segments?
• How much time do you typically spend each day preparing or airing weather updates?
• And what should we pay attention to as we continue developing it?

If anyone’s curious to try generating their own forecasts (no cost, one-month test), just drop us a line at [email protected].

You can also read more about the project here: Noise&Patterns - AI Solutions for Real Business Needs

Happy to answer any questions or hear your thoughts.

Best,
Wojtek from Noise&Patterns

 

[K.M. Richards]

Dave Scott -- famous for Century 21 Programming, back in the day -- was doing something like this decades ago with Unattended Weather. It pulled data from the National Weather Service and then assembled pre-recorded audio segments, including the temperature and sponsor credits ... and as I recall, for less than you are charging (in rated markets, he even had a barter arrangement for stations to get the service free by running a couple of minutes per day of national commercials).

 

[lanceventa]

@NoiseandPatterns Everything on your site is targeted towards European businesses. As this is a US based forum I think it first should matter whether you're even targeting stations here.

 

[Theater of My Mind]

This post isn’t meant as an advertisement - we’re genuinely looking for feedback from radio professionals

Cute. Do you think people in the radio business don't know an advertisement poorly disguised as "looking for feedback" when they see it?

The sole reason radio companies are interested in AI is to replace humans so they won't have to pay them anymore. You'll find some members on this site supportive of that mindset, unfortunately. But anyone who works hard on the creative side isn't falling for your "ease the workload" pitch at all.

Besides, if your radio station is still breaking for weather -- why?? It's not 1980 anymore. Everyone has the latest forecast on their phone and they know what the day will be like before they walk out the door in the morning. In 2025, stopping down for weather on the radio is pure clutter, at least in music formats where listeners will always tell you they want to hear less talk.

 

[K.M. Richards]

The sole reason radio companies are interested in AI is to replace humans so they won't have to pay them anymore. You'll find some members on this site supportive of that mindset, unfortunately. But anyone who works hard on the creative side isn't falling for your "ease the workload" pitch at all.

I'm somewhere in the middle of all that. Several of my consulted stations (including one I directly program) still has weather once an hour, but it's from a service with actual meterologists, not audio files and AI.

Besides, if your radio station is still breaking for weather -- why?? It's not 1980 anymore. Everyone has the latest forecast on their phone and they know what the day will be like before they walk out the door in the morning. In 2025, stopping down for weather on the radio is pure clutter, at least in music formats where listeners will always tell you they want to hear less talk.

Take 30 seconds to give the forecast at the end of a stopset where you have already told the audience going in that a long music sweep follows, and I believe they interpret it as reinforcement of what they already expected, without having to open an app to hear it.

That way, it doesn't get perceived as "clutter".

 

[b-turner]

The station I work for does weather twice an hour. It comes from what may be Dave Scott's system. We have a reputation as the go to weather station. It's a good reputation in tornado alley. Anyway our weathercasts are sponsored 24/7 and represent about 5% of monthly billing.

We use such a system because we would not be doing weather without it.

 

[fybush]

Something to consider...

As I start my newest career marketing broadcast software, I'm realizing how much even I didn't know about the ways in which things are changing under the hood for modern broadcast facilities.

So many of the things that we used to turn to external hardware or software to do, are increasingly becoming tightly integrated into an overarching suite of broadcast software. Time/temperature/forecast is definitely one of those things. Whether or not you're a fan of AI in general in radio (and for the most part, I'm not), there are definitely places where it works and works efficiently. If you're in a daypart that wouldn't have had a live or live-sounding host before, AI voices are widely available as part of modern automation systems like the Myriad software I sell. You don't need an outboard box or a separate piece of quarter-century-old software when you can have your automation system grab forecast and temperature data and read it out in your own talent's voice or in an AI voice that sounds a lot more lifelike than the old Tom Churchill boxes ever did.

Sure, it might be working for you now - but is there any support for it when it breaks? Or when Microsoft tinkers with Windows drivers?

 

[K.M. Richards]

Sure, it might be working for you now - but is there any support for it when it breaks? Or when Microsoft tinkers with Windows drivers?

In my experience, Microsoft's insistence that everything be based on whatever they think the standard should be at the moment has caused more problems for users -- both individuals and businesses -- than they solve.

I fault the software developers for buying into that mindset, like lemmings going off the edge of the cliff. (See what I did there? LOL)

 

[Theater of My Mind]

If you're in a daypart that wouldn't have had a live or live-sounding host before, AI voices are widely available as part of modern automation systems like the Myriad software I sell.

Also, if you're in a daypart where there WAS a live host before, the same AI voices are replacing them. But so far, I haven't heard an AI voice say anything worth talking about.

Consider how we got here. Back in the good old days, radio stations needed a real person in the studio 24/7 to manually play the records or CDs along with the commercials. Since they had to be there anyway, they were paid to be entertainers too. Those DJs came up with very creative ways to amuse, inform and connect with the audience between songs, and it was a real art form. Sure, there were liner card readers and even tape-based automation system for less engaging formats, but the top radio personalities we think of today came from that creative, live DJ setting, where he or she had to be there to run things.

Now with today's digital playout systems, as you know, Scott, no one really needs to be in the studio, ever. So first we got voice tracking, and now AI, still using the old model of talking between songs, but to what end?

Well, I just took a listen to Joel Denver's AI "Phoebe.fm" station after reading Sean Ross's first listen.

I have a lot of respect for Joel and the top radio veterans on his team programming the stations, but while the demo I heard is impressive on a technology level, it's lacking where it matters. Those in the business who have weighed in focus on how human the AI voice sounds, and how seamlessly it flows. What no one seems to consider is the lack of real entertainment value from this AI host. Some very lame jokes. And a lot of filler. It's really banal. This is an Alternative format which should be especially conducive to the host talking about the music, some interesting background details about the artists, whether they're touring, coming to town, etc. To my ear, this AI voice is trained to tell dumb jokes and talk for the sake of talking. Garbage in, garbage out. And while the voice technology is improving, you can still tell it's AI. There's no comparison to real human radio hosts, whose art in years past paved the way for whatever this is.

Is it good enough for radio companies? Probably yes. And that's probably a big part of why radio finds itself where it is today. They just want to invest in tech, not people anymore.

 

[K.M. Richards]

This is an Alternative format which should be especially conducive to the host talking about the music, some interesting background details about the artists, whether they're touring, coming to town, etc. To my ear, this AI voice is trained to tell dumb jokes and talk for the sake of talking. Garbage in, garbage out. And while the voice technology is improving, you can still tell it's AI. There's no comparison to real human radio hosts, whose art in years past paved the way for whatever this is.

And that is why I still pay Freddy Snakeskin for voicetracks on Flashback Weekend.

 

[TheBigA]

Is it good enough for radio companies? Probably yes. And that's probably a big part of why radio finds itself where it is today. They just want to invest in tech, not people anymore.

If investing in live & local talent would cause listeners to throw away their phones and laptops, they'd do it in a minute. But it wouldn't. And they'd just be in the hole spending money on services that could be done just as well by a centralized host in a single studio like K-Love or SiriusXM.

Because as you put it so well:

Besides, if your radio station is still breaking for weather -- why?? It's not 1980 anymore. Everyone has the latest forecast on their phone

Why spend money on live & local talent? It's not 1980 anymore. You could be Bob Pittman.

 

[K.M. Richards]

As I have said -- most recently in the past 24 hours -- the listeners who abandoned us for their phones and laptops and satellite are the reason we have had to downsize.

But, having abandoned us, I feel they have forfeited the right to criticize what we have become because of their actions.

 

[b-turner]

So because radio is not running DJs we are running people away from radio to online streaming that is void of DJs and any talking except for commercials unless you are actually streaming a radio station. That seems to be what is being said: No DJs on radio drives people online to No DJs streams.

 

[K.M. Richards]

No, that is not what I said.

The people "ran away" before we found it necessary to eliminate live DJs, etc. because their running away cost us ad revenue. The elimination was the effect they caused.

Now, we are blamed for the degradation of programming that was the result of their actions. That's what I find unfair: They left because they didn't like the fact that radio is mass appeal, and now they think they have the right to criticize what we have become because they left.

 

[b-turner]

That seems to be what critics are saying about radio. I'm in the business and have seen the revenue vanish. Radio just can't afford to do anything but. Yes the audience, at least a portion, went elsewhere. And I tell people broadcast radio has never been better researched than it is today which is why consensus programming is what it is.

For what it's worth, the voice-tracked small market station I'm selling for these days is doing very well, setting sales records and doing better than ever with listeners. We have 1 in 6 listening with 50 signals received. Sure we are in the right community for this to happen but it also helps we have a great owner with decades of programming in major markets under his belt.

 

[fybush]

"That's what I find unfair: They left because they didn't like the fact that radio is mass appeal, and now they think they have the right to criticize what we have become because they left."

They do have the right to criticize it, at least here on these boards where you are not a moderator or administrator. If these boards had been created purely as professional forums only for those actively working in radio, you could argue that the rules should be different. But they weren't, and they aren't. These are forums that (at least in their current incarnation) are very actively open to anyone with an interest in broadcasting, no matter how annoying you might find that.

That said, I think there's a kernel of truth to what you're saying, which is that we have to deal with the radio audience as it exists today, not as we wish it might have been. That's how I look at all the different elements of what's turned into an unexpected series of career changes for me - when I started in the business 35 years ago, it was plausible to think I could get through an entire career just as a radio news writer, or at least as a radio newsman in general. Those jobs are mostly gone now, and the ones that are left are tenuous at best and don't pay enough to put two kids through college, as I'm presently doing.

Hence the last couple of decades of adaptation - learning how to do trade publishing, contract engineering, FCC application work, station brokerage, and now this last very surprising twist in which I find myself not only doing software sales but really enjoying it.

I think I'm valuable to my new employer, which is new to the US/Canadian marketplace, because I am able to be both knowledgable and honest with them about what the state of the radio industry is really like right now. I get paid to do a lot of what I was already doing, which is talking to lots of people in the business about what the business is like right now and how my company can help them solve some of the technology challenges that keep them from doing the most they can with the level of staffing they already have.

That's where I thought I was coming into this conversation before it started doing the usual R-D spiral outward. It's absolutely not my job as a salesman (or as an engineer) to tell a station owner that they should have a live body in the studio if they can't afford one, no matter whether or not I think it might be a good idea. It is my job to show them how the things they couldn't have easily done with software (or hardware) even a decade ago are much easier to do now, especially when it's something my company's software can do a lot better than some of the more established players in this market.

So if the question is, "how do we provide better weather reports to our audience," my answer can't be, "you don't need weather reports because people have phones." It can't be, "allow me to complain about how some listeners have abandoned radio and I'm angry about it." It can't be, "AI sucks and you need to have a live human in there." None of those things answer the question that's being posed here.

If I'm being asked that question by a station owner, then, my real-world answer goes like this: "I'm not a big fan of AI when it's trying to fully replace a live or good voicetracked human voice. But I also think it's better to have some current local content in off hours than to have none at all. And if you're going to do that, I think the state of the art in 2025 is to do it as part of your automation suite instead of from another external hardware box or additional software package that you'll have to pay separately to support and update. As with a lot of things that used to require separate boxes, this can all be part of your main automation. Change something in your playout log or in your scheduler and your AI voice for weather will change right along with it." That's my answer whether I have my engineer hat on or my sales guy hat on. If it's my sales guy hat on, I'll add, "we'll work with you to make it all sound as seamless on the air as anything does, and probably at a combination of lower price and better support than you'll get from anyone else."

Which is basically what I said in my initial answer.

 

[K.M. Richards]

... but it also helps we have a great owner

That always makes a big difference.

 

[K.M. Richards]

They do have the right to criticize it, at least here on these boards where you are not a moderator or administrator. If these boards had been created purely as professional forums only for those actively working in radio, you could argue that the rules should be different. But they weren't, and they aren't. These are forums that (at least in their current incarnation) are very actively open to anyone with an interest in broadcasting, no matter how annoying you might find that.

I'll agree to disagree with you, Scott.

I simply do not think that people should make a conscious decision to turn their backs on us and then come here to tell us -- after the fact -- how "terrible" we are.

I refuse to have people who do not accept that their decision is what led to the present-day situation, and suggest that we should all pack it in, because they say so.

I have always had a low tolerance for armchair quarterbacking, and this is the worst form of it.

And I especially have a distaste for those who, upon having the realities explained to them, come right back as if the answer is wrong and insist that they are right.

That said, I think there's a kernel of truth to what you're saying, which is that we have to deal with the radio audience as it exists today, not as we wish it might have been. That's how I look at all the different elements of what's turned into an unexpected series of career changes for me - when I started in the business 35 years ago, it was plausible to think I could get through an entire career just as a radio news writer, or at least as a radio newsman in general. Those jobs are mostly gone now, and the ones that are left are tenuous at best and don't pay enough to put two kids through college, as I'm presently doing.

There, I agree with you much more. Those of us who are still here, despite the naysayers, have managed to adapt to a changing business model. And at the same time, I am amazed that, after a career that has spanned more than a half century, I'm still able to do this and make money doing it.

Thankfully, I have no kids to put through college, and that means I probably have more in my investment accounts than most people in this business. So many of the people I worked with during my career (if they aren't already dead) got out a long time ago; I just never found anything else as satisfying as radio programming. Thankfully (again) this is a God-given instinctive talent and I still get to use it.

So if the question is, "how do we provide better weather reports to our audience," my answer can't be, "you don't need weather reports because people have phones." It can't be, "allow me to complain about how some listeners have abandoned radio and I'm angry about it." It can't be, "AI sucks and you need to have a live human in there." None of those things answer the question that's being posed here.

But, to circle back to my original dissatisfaction, those are the kind of arguments that are made here which, quite honestly, make me nucking futs.

I respect you, Scott, and you already know that. But I cannot shake the feeling that the naysayers are being unfair and greedy by their actions and words. They don't want us anymore? Fine. That's reality. But please don't come back here and tell us that we are worthless for adapting and surviving.

 

[crainbebo]

I actually noted an AI weather forecast earlier this summer while FM DXing in Idaho. It was an E-skip catch, WTBX 93.9 in Hibbing, MN. VERY obvious AI-generated forecast, robotic delivery, no nuance...not one of those Weatherology soundbite things ("Fair skies. Tuesday. High of. 63.") Which is why I am so thankful for the syndication that Dan Holliday, Jennifer Narramore, and Steve Hamilton provide to dozens of stations across the country. Real voices with years of experience.

Radio will die a very quick death if AI DJs, AI morning zoos, and AI live remotes become the norm. I'd rather hear a prerecorded voicetrack from a real human.

 

[kevtronics]

I simply do not think that people should make a conscious decision to turn their backs on us and then come here to tell us -- after the fact -- how "terrible" we are.

It depends if they "turned their backs" on you when they found better alternatives, or you drove them away and caused them to seek better alternatives... or a combination of both.

To me, replacing human voices with AI is one more step in what people are calling "en$hittification" these days.