StereoTool with Hans van Zutphen

What would you think of an audio processor designed by someone with little prior knowledge of audio processing? The very popular software audio processor, StereoTool, is designed and updated by Hans van Zutphen. Hans tells us how StereoTool got started, what portions of it are now in other major audio processors, and how radio enthusiasts are building their own audio processors for AM, FM, and Internet streaming.

Watch the Video!

Read the Transcript!

Kirk Harnack: This Week in Radio Tech, Episode 248, is brought to you by the new Axia Fusion IP mixing console, packed with features refined from a decade of IP audio experience; by the Telos Z/IP ONE IP audio codec - it's the best way to hear to from there; and by Lawo and the new crystalCLEAR virtual radio console - crystalCLEAR is the radio console with a multi-touch touch-screen interface.

What would you think of an audio processor designed by someone with little prior knowledge of audio processing? The very popular software audio processor Stereo Tool is designed and updated by Hans van Zutphen. Hans tells us how Stereo Tool got started, what portions of it are now in other major audio processors, and how radio enthusiasts are building their own audio processors for AM, FM, and internet streaming.

Kirk: Hey. Welcome in to This Week in Radio Tech. I'm Kirk Harnack, your host. I'm delighted that you're here. We've got a great show for you. This is going to be fun. I like it when we get to do a fun show. This is a show where we talk about radio broadcast technology, audio, I mean, the stuff that really counts. We don't talk about monster cables. We make fun of monster cables.

We talk about the light bulb at the top of the tower and everything in between. Streaming, and so if you're interested in that kind of technology, getting audio out to the masses, making it sound good, creating compelling content, this is where we do it.

I'm Kirk Harnack, founder of the show. Normally I have a co-host along with me, but Chris Tobin was unable to join us today, although he's going to be hosting the show next week. We've got a guest next week, Tom Morris, but this week it's a treat. Our guest on the show today is Hans van Zutphen. Hans van Zutphen. He is from the Netherlands. Hans, welcome in to This Week in Radio Tech.

Hans van Zutphen: Thanks.

Kirk: Glad you're here.

Hans: Great to be here.

Kirk: Yeah. I'll tell you what. We've got a commercial to do, but I want to give our viewers and listeners a little taste of what you're going to be talking about. People who watch the show, people who listen to the show, love audio processing. So many engineers think audio processing, that's a cool thing. They're interested in it.

People want to be expert at it. They want to get it right. They want their audio stream or their radio station to sound as good as possible. People are often looking for, what's the cool way to do this really well?

If you would, Hans, give me the taxi speech, the elevator speech about your product, what you do for audio processing. What's the short version?

Hans: It will sound better than sending 96 kilobit MP3 through a monster cable. Basically, it's just an FM processor, but in software instead of in hardware. It will do everything that you're expecting from a normal audio processor. One of the advantages is, on the PC you have much more calculation power and you can do a lot with it and make it sound really great. That's basically it.

Kirk: What we're talking about here, viewers and listeners, is actually a longstanding product called Stereo Tool. I've been hearing about Stereo Tool for some years now. I didn't know who wrote, who created Stereo Tool. It is an audio processor that is a software application. It runs on Windows. They also have versions for Linux and a beta version for OS10. Both popular versions run in Windows.

It uses an ordinary sound card that needs to be a modern sound card, probably a nice, good-quality sound card to actually put out a composite signal for your FM transmitter, but it also does processing for streaming. We're going to talk about that, Stereo Tool. Again, it's a software audio processor that runs on your PC. It's really cool. We're going to talk to Hans, the developer of this software and try to understand some of the tricks inside, including this very cool de-clipping algorithm that Hans came up with.

Our show is brought to you by three sponsors, as usual. It's brought to you by the folks at Axia and the Axias Fusion. They're our first sponsor. Let's throw a picture up on the screen, if we can. The Axia Fusion is a console, an Audio over IP console, that has been in the works now for about three years. Now it's shipping. I just got some pictures. In fact, I may try to put a picture up on the show notes.

I just got a picture of some of the first Fusion consoles that are now on the air. There's one on the air at a cluster of stations in San Francisco, a brand new Fusion console. This console is just absolutely gorgeous. I've been doing some webinars about the Fusion console from Axia.

A couple of the highlights are that, first of all, it's gorgeous. If you're going to spend hours and hours of your life, days, weeks, months, years of your life operating this console. Hey, let's make it look good. Let's make it easy to operate.

The Fusion console, even more so than Axia's previous consoles like the Element, the Smart Surface, remember that one? The Radius, and the IQ. Even more than those, there's power under the hood of the Axia Fusion console. As far as the construction, it's just gorgeous. I wish I could bring one right here and show it to you.

One of the features of the construction is, instead of the markings on the console being painted on, or even being printed on Lexan, they're actually laser etched into the metal and then double anodized over it so that they will never rub off. An Axia Fusion console will look, keep it clean, as pretty ten years from now as it does the day it's brand new.

There's nothing really to wear out on the console, especially on the metal parts here. The whole top surface is machined aluminum. It's not sheet metal. It's machined aluminum. It looks really gorgeous.

One of the features of the new Axia Fusion console is that it has this very powerful back feed and record mode setup. There are some presets that you can pick for back feed and record mode, and they work great. They work just like the Element console, or the IQ, or the Radius console. But in the Fusion, you can set up very customized back feeds.

How is this useful? If you're doing a live, remote broadcast, or as our friends outside the US say, "an OB" - an outside broadcast, you can set up a back feed that feeds automatically, let's say the program channel down the left side, and a chat channel down the right side, or you can flip them if you want.

The point is, you can do a very complicated remote broadcast that, prior to the Fusion, would require some external mixing, a little rack mount mixer, or a mixer box or something. It would require outside stuff.

The Fusion does all of that internally now and it does it all automatically. The operators don't have to worry about a thing. They load their profile. They put their remote on the air. They have a conversation. They're feeding program out there. It's really awesome.

Every channel on the Fusion has its own OLED display, and that means that it shows you the name of the source, but it also shows you confidence meters for the audio coming in and back-feed meters, audio going back out to the source, like headphones for your talent, or a mix-minus feed going to a phone caller. That's all metered in real time on the console.

I'd love for you to check it out on the website. Go to TelosAlliance.com, or go to the AxiaAudio.com site. It'll take you to the same place. Go check out the Fusion Console from Axia. They are shipping now. People have been waiting for these for a couple of years now. They're finally shipping and working great out there. Good reviews of those things.

All right, thank you very much, Axia, for sponsoring This Week in Radio Tech. Check out that Fusion console.

Our guest on the show is Hans van Zutphen. Hans is the creator of Stereo Tool. Hans, let's start at the beginning. We're going to show some videos coming up in the show of Stereo Tool in action. Let's start a little bit at the beginning. What in the world inspired you to create an audio processor in software as an application? What made you think you could do this?

Hans: Actually, that is really strange because I didn't think at all that I could do this. I've always liked being busy with audio and radio. I remember getting up when I was about six or seven years old and trying to find stations on the air. A few years later when I was about 11 or 12 I think, I made the first program that I ever made to digitize audio. I just kept doing stuff with audio and with programming. At some point in 2001, I started an internet radio station. For that, I wrote a few small algorithms.

Kirk: Wait, 2001? That was a long time ago. People were barely streaming in 2001, right?

Hans: I know. I actually heard a pirate here that had an internet stream. I thought, when a pirate can have it, why can't I? So I looked it up and two days later, I had a station running. I had to make it sound good. Obviously, if you don't do anything it doesn't sound good. Also, it had to broadcast to a very small bandwidth, 56 kilobit MP3. That was dimension [sounds like 00:09:43] that was available at that provider.

Actually, after a while I managed to get a good sound out of it by broadcasting in mono. To do that, I had to do a lot of other tricks to get a good sound out of it. I made some filters and never thought about them again. Then I worked at Ace Metal [sounds like 00:10:01] where they make those big machines that are used for the chip manufacturing industry for Intel and AMD.

Kirk: Yeah.

Hans: I had been there for about five years. Then I went to Phillips Healthcare, where I was going to work on image processing for x-ray systems. There were about two months in between. I thought, "Well, I need some sort of an exercise project to learn C++, a programming language. Then I remembered that I had all of these filters lying around. I thought, "Let's just convert those to C++ and make a useful interface around it."

I did that and again completely forgot about it. About a year later, I found a website where I could upload it. I did that. Then suddenly there were 5,000 downloads in the first week and 90,000 in about two or three months. Of course, then I wanted to make a new version. That just kept going and going. More and more people started asking for new features. Still, it was just a hobby project in the evening. It was beside my normal job.

Kirk: I've got a question for you, in talking with you in the past couple of days, and you just mentioned it a moment ago, you are using the word "filters" to describe some of the things that you're doing.

Hans: Yes.

Kirk: This is kind of a different use of the word filters for me. When you say filters, are you describing processing algorithms - compressors and so forth, or do you actually mean high and low pass filter?

Hans: No, just any processing algorithm. Basically I call all of those things filters. I don't know why, but yeah, I do.

Kirk: That's okay. Call it whatever you like.

Hans: It was just a hobby project. It stayed that way for years. Many times, I wouldn't touch it for months in a row and then work on it for a few weeks and bring out a new version. Then that was it. Around 2011, things started to catch up. I needed more and more of it. I got more and more feedback from people.

In the beginning of 2011 there was this pirate radio station audio processing event in the Netherlands. I actually decided to go there the last day. They just arranged a table so I could put my stuff up there. There, I heard all the other processors. The good thing about this "processing free day", as it was called, was that they prepared a bunch of difficult audio. Or a whole set of difficult audio with all kinds of volume changes, bright and not bright audio, and just very difficult to process stuff.

They basically asked everyone who was there to process out that audio and to upload it to a website.

When I got home, I figured out when comparing the audio of all the other processors against mine that I really had a lot of work to do to catch up with what everybody else was doing. Basically, I had two options, keep doing this as a hobby and never really get somewhere, or quit my job and start working on it full time.

At the same time at this event, I also spoke to Leif Claesson, who was actually there to present, for the first time, his Omnia.9. It was actually shown for the first time there on that pirate station event, so that was also a bit crazy.

He had a de-clipper in it. I thought, "Oh man, I really need to have one of those things." I actually figured out a way to de-clip audio years before, I think. When I was working on my master's thesis, I had some side projects that I was doing at the time with the algorithms that I made for that. I basically made something that was intended to visualize audio in glasses, something like Google Glass, for deaf people.

I had a lot of analysis steps and algorithms that I made for that master's thesis. Some of those were actually usable to predict what kind of audio would be coming next. I knew that I had something that would be usable to de-clip audio. The reason why I wanted it was that, by that time I had a clipper that didn't distort.

People kept sending me files and telling me, "Ah, I hear distortion in that file." Then I'd play that file through my processor and yeah, it did distort. Then I'd listen to the original. Ah, it's in the original.

That kept happening more and more. So then a few days later, I suddenly got the idea of how to make a de-clipper. I actually made the first version of it in three evenings, just after work. At that point, I had agreed with Leif that we would send each other licenses of each other's software so I could use his MPX tool and he could use my Stereo Tool.

I had not thought about it anymore, so I sent him the license and he basically sent me an email back asking as a joke, because I had told him that I was thinking about a de-clipper algorithm, "How's your de-clipper going?" I sent him some before and after audio. He heard it. Within an hour, I got an email back, "What's your phone number?"

Two hours later, we had an agreement that it would end up licensed into his Omnia.9 and that I would be coming to the NAB to visit the show and visit the Omnia booth.

Kirk: Let me make sure I've got this right. Your de-clipper algorithm that you had thought about some years before and then you had recently implemented, this is the same algorithm that is in an Omnia product. It's in the Omnia.9 because you and Leif have this little cross agreement to share some ideas?

Hans: Not so much to share ideas, but to share technology. I basically licensed it to him as a library. That's basically it.

Kirk: Okay. Sorry to interrupt. If people can get your de-clipper, and we're going to explain, by the way, what this is. Remember, we have some viewers and listeners who may not really think, "What's a de-clipper? Isn't the process supposed to do clipping? It's not supposed to do de-clipping." We have a video to help explain that in a few minutes. Go ahead with your story about after this happened now.

Hans: The reason why I hadn't made it yet until that point was because the algorithm that I had thought of during my master's thesis would take about a year of calculation for one second of audio, so that would be completely unusable.

The new idea that I got after being at this free day was something to make it very efficient and not necessarily produce 100% optimum results, but, I have to carefully phrase it, one of the optimal solutions to the data that's available and that's not clipped.

What it basically does, for people who don't know what it is, if you take audio that's recorded too loud and it sounds really bad, you can basically fix the parts that are clipped and that are distorting. There are a lot of filters that claim to do this, but many of those just put some sort of shape on top of the audio which looks better but doesn't necessarily sound better.

Kirk: Oh, okay.

Hans: This algorithm that I came up with basically guarantees that it will always be better than what you had, and also a whole lot better, and usually almost indistinguishable from the original before clipping. It's actually even in use now by some forensic police labs in Germany.

Kirk: Okay, so they maybe have a telephone recording or otherwise bad recording of some evidence that they need, and they run your de-clipper and it sounds better?

Hans: I don't really know what they use it for, actually, but I know that I sold a number of them to forensics labs. They are using it for something.

Kirk: Interesting. I realize there are trade secrets in your algorithm, but can you... Maybe we should watch a video and then maybe you can explain a bit what you can about how this reconstructs audio that's been lost through the clipping process. Would that be a good idea to show the video?

Hans: Sure.

Kirk: Okay. Andrew, let's show that first video, start with 16 seconds in.

[Music 00:18:32 - 00:19:30]

Cool. We can hear with our ears the clipped version sounds distorted, pretty badly. Then we can see with our eyes that the de-clipped version has peaks and substantial audio restored. Tell us about what we saw and heard right there, Hans.

Hans: I can tell you what I'm not doing. I cannot tell you what I'm doing.

Kirk: Okay.

Hans: What I'm not doing, which would actually kind of work a bit, is you have all of these short sections of audio that are good. Maybe you have ten samples that are not clipped. Then you have a few that are. Then you have 20 other samples that are not clipped. What you could do is take those 20 samples that are good and then just try to go to both sides and calculate what would have been there based on those ten samples.

Kirk: Yeah.

Hans: That would be better than the original. But if you have very extreme amounts of clipping, it would actually probably work very well for very small amounts of clipping. But if you have a lot of clipping, it wouldn't work anymore. That's actually all I can tell you because anything more would give away too much. Sorry.

Kirk: When I first heard of de-clipping, of course my brain, my engineer brain is thinking, "How would you do that?" Of course, I was thinking maybe you just draw a line. You have the waveform going up, then clipped, and somewhere it comes back out of clipping. You just draw a line between these two. That's just drawing. That's not a complex waveform. That's a simple waveform. You're going to lose all the harmonics. You're just going to be drawing the lowest frequency thing you have in there.

What you alluded to a minute ago was you are looking at samples prior to the clipping. You could assume that during the clip period, the harmonics and the fundamental have the same relationship. You just take a sample of that.

You build it up and bring it over the crest and bring it back down the other side and replace it that way. I think that's what you alluded to that is what most people would do who are designing a de-clipper. Yours is not that. It's different than that. That's what you can't talk about, right?

Hans: Yeah, exactly.

Kirk: Okay.

Hans: Basically, what I can say is that I look at a lot more samples. If you only look at 20 samples, you just don't have enough information. You also have to look at samples that are maybe 100 samples away and also not clipped, but separated by many pieces that are clipped in between. You basically have to try to combine everything you can find about the audio that's still there.

Kirk: Okay.

Hans: The whole idea is that audio is redundant. That's why MP3 can work. The same thing can be used for de-clip.

Kirk: Interesting.

Hans: And that's it.

Kirk: Okay. De-clipping technology, one reason why we want to use this technology is because people who produce music, who produce CDs or MP3 tracks or whatever your downloading, for whatever reason, they want it to sound louder. It seems to me, and correct me if I'm wrong and maybe explain it further, but a lot of music producers, they have, maybe, some pretty simplistic tools to make their audio louder. At least a few years ago, this was true. They would just basically use a pretty simple clipper to make the audio louder.

If you go buy a CD and look at the waveform with an oscilloscope, you oftentimes on many songs, many CDs, you'll see actual square wave clipping going on. It does sound louder, but it sure doesn't work very well if you're going to process it further beyond that, like for a broadcast chain or a streaming chain.

You end up causing some real problems when your source material is clipped. Am I on the right track there? Do we have problems with source material?

Hans: Exactly, that's the reason. I really don't get it. I've actually tried de-clipping and then clipping again with my own clipper. The crazy thing is, if I do that, it actually sounds good. It's not less loud. There are ways to clip without having these horrible clipping distortion sounds, but for some reason, almost nobody in the recording industry seems to use it. I have no idea why, actually.

Kirk: I know that some people in broadcast, 10 or 12 years ago, let's say, Frank Foti and Bob Waterban [sounds like 00:24:13] sent a letter from the two of them to representatives of the recording industry saying, "Hey guys, stop it. Stop producing clipped audio. We'll help you if you want, but stop this." To my knowledge, nothing really happened. There are some song producers who do produce fine quality audio, but plenty of others who produce clipped audio on their final product. Wow. You said that you can actually de-clip and then re-clip and it sounds better than the original clipped audio?

Hans: Yes, because the harmonics are gone. The dynamics would still be missing if you do this. The dynamics are as much as will fit in there. At least the harmonics will be gone, so it will sound much cleaner and less crunchy.

Kirk: We've been talking about de-clipping. Clipping is a process that is typically thought of as being a very fast process. It's not audio compression. It's not automatic gain control. It's actually stopping an electronic voltage from going any higher. It used to be done with diodes in the electronic circuit.

With a de-clipping circuit, you're restoring what essentially a diode or an algorithm, cut off. What about restoring the dynamic range? Can you do a backwards function of audio compression?

Hans: Oh, that would be a lot more difficult because there's this very big difference between the two. If you have clipping, you know exactly what happens. You know these samples are good. These samples are bad. I just need to restore them from whatever is still there.

If you have compression and you know exactly what the compressor did, you might be able to restore it depending on if it's limiting or compressing. In the case of limiting, it will just be completely impossible. I should never say completely impossible, but it won't be easy. For compression, you have to know exactly what happened, otherwise you'll just be guessing something. You might be doing something wrong.

What you can do, that kind of works, is take out the transients and the kicks and the parts that should be loud and boost them. That's what I have done in my natural dynamics filter. What it basically does is it takes out the percussion of the song. Then you can separate them from the rest, boost the percussion, and mix it back in. You're basically going back to the mixing table indeed. I do have that.

Kirk: Tell me the name of this function again. Dynamics what?

Hans: Natural dynamics. It's actually the second video that should be standing ready.

Kirk: We'll take a look at that in just a few minutes.

Hans: Okay.

Kirk: You mentioned you could restore dynamics from compressed audio if you knew the exact characteristics of the compressor. It brings to mind that this actually is the technique that was used for years in noise reduction technology, where you have a compressor and an expander, like Dolby noise reduction and other brands.

Hans: Oh yeah.

Kirk: When we had audio tape and we had a limited dynamic range on the audio tape, maybe 50 or 60 decibels, but we wanted to have a 90 decibel dynamic range, you'd run it into a Dolby system that would take your dynamic range and compress it, put it on the tape, and because we knew the characteristics of the compressor, we could then expand it back out and come back with about exactly what we started with. We'd have a better dynamic range than the tape would handle.

If you don't know, if you have multi-band compression with variable release times and variable attack times, then all hope is lost. You can't recreate that at the receiving end, I suppose.

Hans: It will, at the very least, be extremely difficult. I can think of a way to do it, but again that would require a whole lot of processing power for one second of audio. I'm not going to say a year, but a lot. A lot more than we have.

Kirk: I take it that if we have some audio that is not ideal for broadcast or for processing, and if we can just de-clip it, that takes care of a large amount of the problem. Now we can process it the way we want. In other words, I was saying that, "Hey, what if you could also restore the original dynamic range from compression?"

I take it that, yeah, we might see some improvement, but it may not really be all that necessary to actually do that. It's the clipping, because that's real distortion. It's the clipping that we really want to get rid of. That's what your de-clipper does. It restores it.

Hans: Yeah.

Kirk: Okay.

Hans: Of course, increasing the dynamics, if you have a song that's really compressed to hell, it still helps. It will basically give the compressor something to work with or the whole processor. Instead of just having some sort of square wave, or maybe not a square wave, if it's compressed, but something that's very hard to do something with. If you have something that's already [inaudible 00:29:43], what can you do? You can make it less flat, but then you're probably just making things worse.

Kirk: Hans, talk to me about, in Stereo Tool, and this is your audio processing software, this is a full featured processor. You have the de-clipping. Take me through the different stages. What can you do to the audio in your Stereo Tool?

Hans: I'll just open it here so I won't forget anything.

Kirk: Okay.

Hans: There are a lot of things. Actually, the first step, which is something that I just... Okay, since I'm doing this and I'm trying to do things in a new way, which is different from what other people are doing, I'm trying to come up with new filters that other processors don't have just because that's more interesting work on, actually.

The first stage that I have just recently added after going crazy from listening to some Greek public radio station that uses my software which had a horrible 50 hertz hum in their shows. In everything basically, because they had an AM transmitter in the same building. It just sounded horrible.

The first thing I made was a filter that you can let it listen to silence for a few seconds and it will make a mask out of that. Just filter out any audio that is present during silence. If you have a 50 hertz hum which is coming into the processor, it will filter it out completely.

After going through that, we have the de-clipper, which you just heard. Then we have a noise-removal filter which just removes very soft sounds. In the latest version, I have a new filter but that's still in beta right now. That is a, this is going to sound weird, de-quantizer.

Kirk: What? A de-quantizer?

Hans: A de-quantizer. If you have, say, eight bit audio, which would be really bad but you can use it for anything, basically.

Kirk: I've got to say, this sounds like it's straight out of Star Trek. "Sulu, hand me the de-quantizer."

Hans: Yeah.

Kirk: I'm sorry to interrupt. What does it do? What does a de-quantizer do?

Hans: Basically, if you have audio that's cut off at 16 bits, it will make it sound like 18 or 19 or 20 bits. If you have 8 bit audio, you can make it sound like 10 or 11 bits.

Kirk: Really?

Hans: Yeah. That's completely useless for normal radio stations, but I came up with it and I thought, "Why not just put it in? Who knows if anyone can find some use for it?" You can just turn it off if you don't need it.

Then there is natural dynamics, which I just mentioned, which takes out the percussion and boosts it separately from the rest. After that, I have several phase rotation filters. One to protect the clipper in the end, which just makes asymmetrical sounds symmetrical.

There's also another phase-rotation filter which is actually phase-delay filter, which you can use to, if you have a bass kick, you can basically pull it apart into multiple frequencies. Instead of having a [makes noise], you will get a [makes noise], which many people seem to prefer. It does indeed sound warmer if you do that.

After those things, there's the AGC, which is just an AGC. There's not very much exciting to tell there, except one thing. It can handle very big volume jumps very well. If you have, for example, suddenly at 20 dB jump, it will not slowly drop the volume, but it will suddenly kick down. You will basically get rid of those sudden, very loud sounds.

Kirk: I take it, it's the AGC that typically will be in front of some multi-band processing.

Hans: Yes.

Kirk: Changing the AGC characteristics can really change the sound produced by the multi-band processing. If you can control very well the audio level being presented to the multi-bands, then you'll have a more consistent sound. You're saying this is what your AGC does, yeah?

Hans: Yeah. The whole idea is that after the AGC, the audio should actually already be more or less constant. Multi-band will just take care of the small details and make it sound fuller and everything else you expect from a multi-band, and make it consistent in the highs and lows.

You don't want a multi-band to handle very big volume changes. You could hear it in earlier processes. Actually, I haven't heard it in more recent boxes. Sometimes when there's a sudden loud kick that you hear different frequencies come up at different speeds when they come back. That sounds really horrible. That's what this AGC should protect against.

Kirk: Okay.

Hans: Then we get into the stereo section, which actually consists of four different parts. Yeah, four. The first step is the azimuth correction. That's something that actually almost nobody seems to have. If you have a very old recording, sometimes there's a shift in time between the two channels. If you then merge to mono, you get all kinds of cone-filter effects. Normally, you don't go back to mono.

If you have bad FM reception, then you will. I know that in the past, and I haven't heard it recently anymore, probably because nobody uses analog equipment for this, like tape. The whole sound gets weird and cone-filtered if your reception is bad. The first step in the stereo section is to align those two channels. It will really dynamically follow what's happening. Even if you have constant shifts and movements between channels, it will just follow that and compensate.

Then there is a five-band stereo widener, which is basically just a compressor on the left minus right signal, which does protect against too much stereo, like more than 100%. You don't want that to happen.

Kirk: Yeah, okay.

Hans: There is also a new filter that is called ACR Stereo, which basically controls in single band the stereo signal. It tries to do all the changes in the amount of stereo on peaks. That's when you don't hear it. If you are going to adjust things during time slowly, you will hear all kinds of weird effects. If you do it suddenly on the peak, you don't hear it. It can also mix in the left minus right signal with a delay. That's actually a very old mechanism that some Hartford boxes came up with in the 1980s.

Kirk: Yeah.

Hans: That also protects you against having a too strong signal at the L minus R, which would cause problems with multi-path distortion and issues like that.

Kirk: Okay, understood.

Hans: Finally, there is one filter for people who want to broadcast in mono. I started out saying that I once had this radio station at 56 kilobit MP3. I had to broadcast in mono to make any kind of decent sound with it. What that filter does, it can down mix to mono but without phase cancellation. You can basically even put two channels which are in complete anti-phase and it will just merge them and you will have a sound which sounds as full as the original stereo sound.

Kirk: What? Really?

Hans: But it's mono, yeah.

Kirk: Wow. That sounds a bit like the azimuth correction you had before. Is it a similar process or is it completely different?

Hans: No, it's completely different because azimuth will just align the two channels. You can still have instruments that have cancellation effects. Azimuth does help a bit, but it's not a full solution. This thing, actually if you combine them, you get the best result.

Kirk: Wow. Okay.

Hans: That's actually the reason why the program is called Stereo Tool. When I made this thing, when I made this GUI, I had actually three filters. I had the multi-band compressor, a single band compressor, and this stereo thing. And then "radio tool" was used too much already on the internet. The next thing that I have is an equalizer where you can just draw the line yourself by using your mouse. You don't have separate parameters, but just one line that you can just change.

Kirk: Okay. Nice.

Hans: Then there's a multi-band compressor with a configurable number of bands, anywhere between two and nine. Whatever you want to use, you can use.

Kirk: Why would you choose two bands versus nine, or nine bands versus two, or five bands, or seven? How do you decide what's going to be right for the format that you're playing or the sound you want?

Hans: For example, if you do speech, then you want a very low number of bands because you don't want all the different frequencies to be taken apart too much. You want it to sound natural. There are some voice processing presets in here, microphone processing presets, which actually sound very similar to some hardware microphone processors. For those, you need a very low number of bands.

If you go higher, you gain something in some sort of consistency between... In spectral balance, but it gets harder to control the peak level. There are different tradeoffs which you can choose whatever you want to use.

Kirk: Okay. All right. More bands isn't always better. It depends on what you're trying to accomplish.

Hans: Yes.

Kirk: Okay.

Hans: That's why you want it to be configurable. You can even just set every band and the steepness of the band. You can control everything. That's really useful if you want to use it for something completely different like, as I said, a microphone processor. Then you just have to tweak everything and in the end, you get a good sound out of it.

Then there are some standard things like a band pass filter, a simple bass boost filter, a single-band compressor, which I'm planning to replace by a two or three band compressor, by the way, because I think that will be better. Then the last step is the clipper. The clipper has some very specific ways to remove any audible distortion. To my knowledge, it's the loudest clipper that exists at this moment. That's the final stage. You can also use it in composite. You can go even a few dB louder without going composite. That's the whole thing.

Kirk: Wow. In this part of the show, we've been talking about the Stereo Tool software that Hans makes. By the way, the website is StereoTool.com. This software is used by serious broadcasters as well as by hobby broadcasters, pirate radio stations, internet broadcasters.

We've got to take a break here in just a second, Hans. I guess you have presets that you can use your processor for, for a stream that's bit-rate reduced. That's one kind of processing. You don't want to do clipping on that. You can also use it for AM or for FM broadcasting. You implement the correct parameters for those transmission methods as well, yes?

Hans: Yes. It has also AM support including C-QUAM stereo, and basically everything I can throw in there, I'm throwing in there. Yeah, it can do everything, at least I hope so. If it can't, tell me and I'll add it.

Kirk: All right. We're going to talk when we come back about the hardware that you might use to implement Stereo Tool if you want to do this yourself. There is a cost for Stereo Tool. It starts at a low price and goes up depending on what features that you want. It's StereoTool.com. Check it out.

It's a great way to learn about processing, but it's also a great way to actually have some audio processing, or mic processing, or whether it's full blown FM processing for your station.

Our show is brought to you in part by the folks at Telos. By the way, I should mention if you've just joined us, you're listening or watching This Week in Radio Tech. It's our 248th episode. Our guest is Hans van Zutphen. He is the designer of the Stereo Tool software audio processor. I'm Kirk Harnack along with you today. Glad to have you along with us on the show.

Our show is brought to you in part by Telos and the Telos Z/IP ONE IP Codec. Here's Andrew showing you a picture of the Z/IP ONE from the website. I encourage you to go to Telos-Systems.com or to our new website at TelosAlliance.com and have a look at the Z/IP ONE. I'm going to take my little camera here. There we go. We're live in front of the Z/IP ONE. What you're seeing right there is the - let me scoot my mic over here.

What you're seeing there is the screensaver on the new beta software for the Z/IP ONE. If I touch a knob, there we go. We're back to the normal screen for a Z/IP ONE. I'm going to hit the "Auto" button. There's Auto. The first one that's highlighted here, I'm going to call this. This is Hope Media off in Australia.

Let's give them a call. It says they do not have a port forward on their router. We're negotiating with the router right now to form a connection. It usually takes under ten seconds for the router. There we go. Bam. Let me let you hear that music. It's coming from Australia to Nashville, Tennessee.

[Music 0:43:42 - 0:43:54]

That is pretty cool. Let's see, I wonder what bitrate they're sending. We can have a look at that. They are sending me HEAAC at 96 kilobits. I'm sending them HEAAC at 96 kilobits. We can hit "Disconnect". It disconnects just that easily. You can call places all over the world. I'm going to put my camera back here, Andrew. There we go. What a way to run a commercial.

If you need to get audio from here to there, whether it's from Nashville, Tennessee to Sydney, Australia, like we just did over the public internet, or whether you just need to get it across town or from the local ballpark back to your radio station to do a live, remote broadcast, or from the car dealer, or whatever, you can easily do it with the Telos Z/IP ONE.

By the way, the new version, the new hardware version of the Z/IP ONE is now shipping, finally. It's the AES version. Now the Z/IP ONE for audio connections has Live Wire. That's how mine is connected to the rest of my system here. It also has analog inputs, as it always had, and now it has AES digital inputs and outputs right there on the back of it.

The audio you're receiving from somewhere else, it comes out going out all those outputs, Live Wire, analog, and AES. Then on your input, you select which input that you want to use. The Z/IP ONE, it's an easy to use IP codec. In most cases, with most routers, you literally can plug it into the router, get a DHCP - IP address automatically, it registers with our Z/IP server if you want it to, but by default it does, and you have instant access to call the people that are in your directory, the buddies that you set up in your phonebook.

If you have a router that needs a little help, you just do a port forward through that router, any odd port you want to pick out, and there you go. You can connect easily that way too. I think it's the easiest IP codec that there is out there in the market to use. In fact, personally I wrote the quick start directions for the Z/IP ONE.

To make sure those directions were good, I had my then 16-year-old daughter, who had never seen one before, read the directions and plug in the Z/IP ONE and make it work, and she did, in four minutes. Four minutes later, she had audio coming from the Telos test line and had never seen this box before. So proud of her. I guess the directions were okay.

Check it out if you would. If you need to get audio from here to there, remote broadcast, people are using this for studio-transmitter links. If you take a few precautions over the public internet, you can do that. One of our guests on the show, Dave Anderson, is doing that on a bunch of transmitter sites with excellent results. Check it out on the web at TelosAlliance.com and look for the Z/IP ONE IP audio codec. Thanks to Telos for sponsoring This Week in Radio Tech.

All right, we are here with Hans van Zutphen. He's the designer of the Stereo Tool audio processor. Hans, we don't have a lot of time left - maybe another 15 minutes or so. Let's talk about how you implement... I'll tell you what. Was there a picture or another video that you wanted to show before we start talking about hardware and actually making this work?

Hans: Let's do the hardware first. If we have time left, maybe. Let's do this first.

Kirk: Okay, good idea. If you want to implement Stereo Tool, it's designed to run on a PC running Windows, right?

Hans: Yes.

Kirk: Okay.

Hans: It should run on anything from Windows XP to Windows 10. Yeah. What you need is you need a PC which is fast enough, and you need a decent sound card. To be honest, even many of the on-board Realtek cards are good enough to create a very good composite signal because you can calibrate it inside the software if there are some deviations in the card itself.

Even with those Realtek cards, with calibrations I can make them flat from five hertz square wave to 60 kHz. If it's flat, your signal should be good.

Kirk: I have a question. If you're looking at the specs of a sound card, can you tell by the specs if it will be able to create and pass an FM composite signal?

Hans: You should be able to tell, but in some cases, some cards don't really do what they promise. Basically, any 192 kHz card should be okay, especially if it is DC coupled, as they call it. That means that if you send out a DC offset, it will just stay there and not fall back to zero. Again, that can also be calibrated if there's not too bad of a drop to zero there.

However, there are some cards that do match the specs according to their official specifications but if you measure what comes out. I had one person who measured one card and he only sent out a 19 kHz pilot tone and he measured audio up to one megahertz with a scope. In that case, it's really not useful. You won't hear it, so the card will sound perfectly fine. But obviously, it's completely unusable if you put it on FM.

Kirk: Let me see if I understand something correctly. If you're going to use Stereo Tool and put it on a PC and come out of the sound card and go into an FM exciter or FM transmitter via the composite signal, and this is a very good way to do it. You're properly limited and clipped, and you're stereo pilot is going to be perfect and pure. This is a great way to do it.

You're talking about coming out of the sound card's output, which may be on a 3.5 millimeter plug and going into probably BNC connector on the exciter, right? This is going to carry the composite signal which is base band from zero hertz, or close to zero hertz, up to what? About 57, 59, or 60 kHz?

Hans: 60 kHz. It also has an RDS encoder built in, so it's going to about 60.

Kirk: What? Stereo Tool has an RDS encoder built in?

Hans: Yeah.

Kirk: Oh my goodness. That's great. Okay. That's cool.

Hans: Since six years actually, already. I think I was one of the first or maybe even the first to add this. Yeah.

Kirk: My goodness.

Hans: As I said, I throw everything in I can.

Kirk: I had no idea that a sound card, especially a cheap built-in one could do this. I thought you had to buy a sound card capable of at least 96 kHz sampling. Does the sampling rate of the card make a difference?

Hans: Yeah. You need 192. Honestly, I've been at the NAB and at the IBC. At the NAB, I didn't have a transmitter with me. At the IBC, I did. I was feeding the transmitter from the headphone output of my laptop. That was going into the transmitter. I had this nice analyzer which was perfectly straight at 75 kHz with [inaudible 0:51:12] blinking. That was it. It works.

Kirk: Oh my gosh. I've got to try this. I've just got to. I had no idea that the headphone output of a laptop…

Hans: Neither did I.

Kirk: Yeah. Can you tell us which laptop this was? Was it a PC like a Dell or something?

Hans: No. This was an ASUS laptop with Bang and Olufsen. I have it here. Let me see.

Kirk: It had a good sound card system in it?

Hans: Yeah. It's just Realtek onboard sound card. Nothing special, actually. It is 192 kHz. As long as you can keep it flat, it's okay. I had expected that the amplifier for the headphone would really screw things up, but I don't know why it doesn't, but it doesn't.

Kirk: I hear you. I don't know why it doesn't screw it up, but it doesn't. Okay. If you did want to use a high quality sound card, you could put one into your PC.

Hans: You really should. Also, with these onboard cards, you don't get low latency, etc. If you want to use this on the real station and not just on some test transmitter or something, or a pirate station, then you really need to get a good sound card.

Kirk: Okay. The sampling, if the sound card will do 192 kHz sampling, then it should have no trouble putting out the composite signal for FM.

Hans: Yeah. Again, it should not, but some do.

Kirk: For AM, if you want to put your Stereo Tool in AM mode, you're not doing composite there. You're just putting out, typically, a mono, clipped, limited, perhaps asymmetrical signal, for an AM transmitter, right?

Hans: Yes.

Kirk: Okay. Then for streaming, tell me, can you send the audio from the Stereo Tool to an encoder program in the same computer, or do you need to come out a sound card?

Hans: There are several ways of doing it. What most people do and what usually gets the best results is to indeed stream it to another program on the same PC, which you can do using some virtual audio cables. There's a program called Virtual Audio Cable. There's another one called VB Audio Cable. You can use either of those. I have recently added VOC integration so you can actually, just from inside Stereo Tool, stream it and then it will find VLC lip [sounds like 00:53:40] on your system if you have a VLC player installed and stream through that.

At least for MP3 streaming, it's not very usable because the MP3 encoder in VLC turned out to be pretty bad. If you use AUK 4, or something like that, then it's fine. Don't use it for MP3. Then you really should use the virtual audio cable solution. Of course, you can also still go through your sound card.

Kirk: If I want to use Stereo Tool for mic processing, you talked about that earlier, what kind of latency would I expect? Can I listen to the output live, or will it be too delayed?

Hans: The minimum latency that Stereo Tool has at this moment is about 16 milliseconds. That would be somewhere close to what's acceptable for most people, not for all.

Kirk: Yeah.

Hans: That's, at this moment, the minimum latency.

Kirk: Does Stereo Tool work, and I may not be asking the right question here, it's my understanding that audio going through the Windows sound mixer may typically receive additional latency. Does it go through the Windows sound mixer? If it does, does that cause additional latency?

Hans: You have several modes. If you use AGO, then it doesn't go through there. The AGO latency is actually, in some PCs, even less than one millisecond. There's no problem at all there. If you do go through the Windows mixer, you get an additional, you can easily get 100, 200, or even 300 milliseconds of extra latency, and it will, if you have bad luck also, resample the audio using a very bad re-sampler. Then you get, if you send out one pilot, you'll analyze it and see ten pilots. That's really not a good idea.

Kirk: That's why it's best to use the, and I know in different countries we pronounce this differently. I would say the AZiO drivers. You say the AZiO drivers. This is the sound card...

Hans: I don't know.

Kirk:... these are the sound card drivers that are designed for professional applications. They bypass a lot of the consumer-ish stuff inside Windows.

Hans: That is exactly one of the reasons why the onboard sound card will give you more latency, because at least these Realtek cards don't have an AZiO driver. You need some sort of thing on top of it, which will at least have about ten milliseconds of extra latency. Ten milliseconds is quite a lot if you're talking about 16.

Kirk: Can Stereo Tool receive its input audio from any recognized Windows source? What I'm getting at is, could I run an Axia IP audio driver in the same PC, receive audio from the Live Wire network, and have Stereo Tool process that audio?

Hans: I would guess so. I've never tried it. If you can send it through a sound card, you can send it to virtual audio cable, and you can also route it back into Stereo Tool. Then, yes.

Kirk: I don't see any reason why it wouldn't work. Any other application sees the IP audio driver from Axia as a Windows sound device.

Hans: Okay. Then it will also work, yeah.

Kirk: That means you can also send it back out to the Live Wire network if you want to.

Hans: I suppose so.

Kirk: I'm using Live Wire as an example. If there are other IP drivers for other standards, than it should work there as well as long as it looks like a Windows device. Wow. Cool.

Hans, when people are building up, let's imagine that, hey, I'm a radio station. I want to play with Stereo Tool. I purchase the software. I download it. Or maybe I take the trial version. I know you said it will run on any hardware, but is there some hardware that's popular? What might you consider important for the PC that's running Windows and running Stereo Tool? What's important for that to work reliably and without any glitches of any kind?

Hans: Most importantly, I would say to get a CPU that's fast enough so the CPU doesn't go too high. If you have a PC that's constantly running at 90% CPU power, then it will get hot. It will break down easier than it would otherwise do. Actually, most hardware will run it fine. I have many people who have been running it for years without a single glitch.

Get some decent hardware, which is not the very cheapest that you can find. Also, as I said, put in a good sound card. For example, Marian TRACE ALPHA works very well. ESI Juli@ works pretty well, but you need to calibrate it a bit.

Kirk: You mentioned the sound card the Marian TRACE ALPHA. I'm not familiar with this one, but you are. This is a German sound card, is that right?

Hans: Yes.

Kirk: Okay. We'll put a link to that in the show notes for the program so people can easily click on it and find the Marian TRACE ALPHA sound card, if you really want to get the good one.

Hans: Yeah. That's the best one that I know of, at least for a decent price. There are other cards of course that are E800 or E1,000 or something. The Marian is around E200 And it's as good as those expensive ones. That's why most people use it.

Kirk: Okay. Tell me about multitasking. Is it really best for Stereo Tool to be the only thing running on a computer? Does it play nice with other programs?

Hans: It depends on what you do. If you allow a lot of latency, than you can get away with a lot more. If you really want low latency and you're talking about one millisecond AZiO latency, then you really need to do nothing else on it. Also, Stereo Tool was made to never allocate memory, never touch the hard disk when it's running, never swap anything to the hard disk, never connect to an internet server unless you're using VOC integration.

Basically, it just does the same thing over and over. It grabs a few bites of data from a sound card, processes it, and sends it out again. That's all it does. If it runs for one second, it will probably run forever. If you're starting to run other software, you do run the risk, and it also depends on which Windows version you use, that the system will become unstable at some point. If possible, I would really advise to not do anything else on it.

Kirk: Okay, yeah.

Hans: You could, by the way, easily run it together with automation, for example, because if the automation goes down, your station is done anyway.

Kirk: Yeah. Okay.

Hans: But I wouldn't go browsing or reading my email if it's on the air.

Kirk: If you had this on a dedicated PC to do your audio processing, or even on an automation PC, typically would you turn off Windows Updates so it just sits there and runs?

Hans: You probably don't want it to reboot at night, so yeah, you should turn that off.

Kirk: Okay, cool. I've got one more topic to talk about. Wait a minute. We had that video and the picture to look at. Why don't we hit that real quick?

Hans: Yeah. Let's take the picture first. It's an interesting one.

Kirk: Andrew, we've got a photo to look at. Let's have a look at that. What is this?

Hans: Okay, what you see here is, this is something that I really wanted to mention because I know that a lot of stations are sending their audio, which they process in the studio, as left right audio to the transmitter site. There, they just add a stereo encoder to it and maybe an RDS encoder. They send out that signal.

Now, what you're seeing here on the left is the composite output of Stereo Tool. What you see on the right is what you get when you decode one of the channels. What you can see here is that the output is a lot louder than the MPX signal is. We have 100% modulation. The audio can go to about 140%. That is something that, as far as I know, only the Omnia.9 and, I guess now the Omnia.7 and Stereo Tool do. No other processor does this.

Kirk: I've got to tell you, the first time I saw this with Leif Claesson, our mutual friend, I had always dreamed of having a room that is bigger on the inside than it is on the outside.

Hans: You have been watching "Dr. Who".

Kirk: Yeah, exactly. Actually, I watched "Land of the Lost" when I was a child. This is like, you have 100% total modulation, but when you demodulate it in a receiver, you get up to 140% peak audio.

Hans: Yes.

Kirk: You're going to have to explain this, over a beer or something, because I don't understand this and I really want to understand how this concept works.

Hans: Well, it's not so difficult. For example, let's say you have an audio peak that's going up in the left plus right channel. The left minus right channel happens to be such, that at that point, because you multiply it with a 38 kHz sine wave, then it goes down. Then the total might be below 100%. If you separate them again, it will be louder. That's basically the whole idea. In the same way, you can squeeze things into the holes that the stereo pilot and the RDS signal make.

That's just one important thing to say. Going to 140% or even higher may cause issues on some receivers. I generated it here in this image going to 140%, but it's actually safer to limit it at about 120%. That doesn't really make much difference to the audio anymore.

The cool thing about this, especially in the U.S. with 75 microsecond pre-emphasis, it's pretty important. All of this extra audio is high frequencies. You basically get about 2 dB or 3 dB of extra head room for the highs.

Kirk: Oh, okay.

Hans: That is really useful.

Kirk: Yeah, especially when you've been limited on the high frequencies by the 75 microsecond pre-emphasis curve. Now you can restore some of those highs that couldn't count earlier. Before we show the video, tell me what we're going to see. Give me a little taste of what we're going to look at.

Hans: Okay. This is what I mentioned earlier, the filter that takes out the percussion and boosts it separately from the rest.

Kirk: I'm going to stop you right there. We're not going to look at it yet because I have to pay for one more commercial. We're going to look at it right after the commercial, okay? We'll keep you hanging there. We're going to look at this dynamic percussion thing. It's natural dynamics. It's going to be cool.

We're talking to Hans van Zutphen. He's the developer of Stereo Tool. He has some great ideas that are now also in another product, the Omnia.9, the Leif Claesson hardware product, the Omnia.9 and then the Omnia.7 from the folks at Telos. This is cool stuff. I appreciate Hans explaining how some of this stuff works, especially as late an hour as it is now in the Netherlands.

Our show is brought to you in part by the folks at Lawo. You may not have heard of Lawo. They're a console company from Germany. They typically make really big, really expensive audio consoles for television, for sound trucks, for live events. Lawo also has a line of smaller consoles for radio broadcasters. They have a console that they've been shipping for a while now called the crystalCLEAR virtual audio console.

The crystalCLEAR audio console is really interesting. There's a DSP engine that sits off in a rack somewhere. That's where your audio inputs and outputs go to. There's some GPIO. There are some microphone inputs with mic pre-amps. There are headphone outputs on there. There is also an Ethernet connector so you can hook up to your network, and you can also get AES67 or RAVENNA AoIP audio in and out of the crystalCLEAR DSP engine.

Here's the cool part. There's a multi-touch touch screen that runs on a PC. You put it in the control room. You put it in front of your talent and you've got an eight fader audio console right there in software. It runs full screen. You really don't see that Windows is running in the background. You see this app and it's absolutely full screen.

It's multi-touch, so you can touch two of the faders at the same time, run them up and down, touch a button, run your speaker volume up and down. It's designed for finger size. They didn't just take some app and try to make it fit on a screen. They designed it for your fingers to fit. You can control the DSP engine with an app that looks like a console.

This is so cool. If you've been dreaming about this kind of a console in your studio, than you can certainly check out the crystalCLEAR from Lawo. It's got the usual face any console would have: program one, program two, it's got a record bus, preview bus. It shows you preview meters. It shows you peak program meters. All the stuff that you would normally expect an audio console to have, it's built in.

Since the console was built in software and the control interface is on a touch screen, they can do some very cool things. Like, when you touch a button to make a change in the source, to adjust the trim level, or the audio processing for a mic, it's context sensitive. It won't show you options that don't apply to that source. It shows you just things that apply to that source. It reduces operator confusion greatly. The operators don't have to ignore certain things and just pay attention to other things. Whatever you want to do on the console is contextual. It's a great idea in consoles.

I want you to check it out. Go to Lawo.com. That's L-A-W-O, dot com. Look for the radio products and the radio consoles. Look for the crystalCLEAR. On the product page for that console, you're going to find a video. It's Mike Dosch. He's explaining, he's doing a demo of the Lawo crystalCLEAR console. I think you'll find it very intriguing. You'll want to check it out. Lawo.com, look for the crystalCLEAR console and watch that video. It's about a ten minute demo by Mike Dosch showing the crystalCLEAR console's operation. I appreciate Lawo for sponsoring This Week in Radio Tech.

Glad you hung on through the break. Hans is going to show us, through a video here this natural dynamics. Did I say that right, Hans?

Hans: Yes.

Kirk: Okay.

Hans: Let's watch a few seconds. It should be clear what it does.

Kirk: Okay. Go ahead, Andrew.

Natural Dynamics demo. All right.

[Music 1:08:52 - 1:09:22]

That's cool.

Hans: That's it.

Kirk: I feel like I'm listening to the dance-mix version of the song. You can make the dance mix version right there, live.

Hans: Yeah. Basically, yeah. The scary part is that they actually released that CD sounding like this. Once you listen to the version with natural dynamics only, you turn it off, that's the reason why I put it in the video this way, it just doesn't sound right.

Kirk: Okay. You hear it your way and you don't want to go back to the original at all, right?

Hans: Yeah. Exactly. Which is weird. Why don't they release it like that? It is dance music, right?

Kirk: Maybe they need to download Stereo Tool and have it do it for them. That is so cool. Hans, we've got to go. We're about out of time. There's one more little subject I want to cover really quick. You and I talked about the philosophy of audio processing. Different audio processing people from Mike Dero [sounds like 01:10:25] to Bob Orban to Corny Gould to Jeff Keith to Frank Foti to Leif Claesson, they all have an idea of how they want to make an audio processor.

You are a bit more, shall we say, agnostic. You have worked with the brain trust of the users of Stereo Tool to give you the feedback of what they want. Can you tell me a bit about how that process has worked for you in designing and tweaking your audio processor.

Hans: Yes. Actually, if you look back a few versions, a few years, to the versions of Stereo Tool a few years ago, all kinds of names are completely wrong. For example, I didn't know what attack and release was, so I called those things up- speed and down-speed. That was because I had no idea what I was doing. I had some idea. I have to change the volume and do something, but I just didn't know what I was doing.

Kirk: Yeah.

Hans: What I have done over the past years, basically, I have actually, probably because my software is so cheap, that's one reason. Also, because you can just download it on a PC and start using it, so there have been over one and a half million downloads by now. There are a lot of people who are using it.

Lots of them are giving me feedback and telling me, "Hey, can you add this? Hey, can you change that? Hey, this doesn't really sound right. Can you look into it?" Sometimes those are even things I don't even hear myself.

The good thing about this is, if I have, for example, something that I don't hear, then normally it's really problematic to fix it. If you have a $10,000 hardware box, which is processing FM right now which is running on your radio station, and someone says, "Hey, I made something. Can you check out if this helps?" That's just not possible. In my case, people can just grab an extra PC, install the software on it, compare the old to the new version, and say, "Yeah, this is going in the direction that I meant," or, "No, this is not at all what I meant."

Kirk: Yeah.

Hans: To me, the fact that it's so easy to install it, you can basically download and install it. I have a video that I recently published where you can get a complete FM station on air with a compliant signal within five minutes. Just installing it and trying it out is something you can do in a matter of a few seconds. That really makes it possible to, what I have often done in the past, come out with new, better versions every day, sometimes even two versions a day. Just to get people to respond to it and use whatever they give me as feedback.

Kirk: You know what's going through some people's minds? "Hans van Zutphen: enabling pirates with better signals every day."

Hans: Well, not just pirates.

Kirk: I know it's not just pirates. You've got professional radio stations using your software as well, and plenty of at home people, podcasters, and net-casters. I get that. There are people who are hobbyists who are using your software because the advantage is, it's cheap to get started. As you said, you can download a new version, try new things, give feedback to you. It sounds like a good ecosystem.

Hans: Yeah, exactly. That's what it is.

Kirk: Yeah, cool. Hans, I wish we had more time. We're out of time and we've got to go. We've kept Andrew up past his bedtime, I believe. I really appreciate you being on the show. I hope you'll come back and see us. If you'll be at NAB, maybe you can stop by. We'll try to do a show from the NAB floor. You can stop by and see us there, okay?

Hans: Okay, I will.

Kirk: All right, good. Hans van Zutphen, his website is StereoTool.com. You can download his software there, a trial version, and pay for it if you like. See if you like it or not. Our show, This Week in Radio Tech, has been brought to you by the folks at Axia, the Fusion console, Telos and the Z/IP ONE IP audio codec, and by the Lawo crystalCLEAR multi-touch touch screen interface console.

We've got, next week, Tom Morris is our guest. This is an interesting guy, I'm telling you. You're going to want to hear from him. He's got some interesting ideas about hacking. Tune in next week for This Week in Radio Tech. More guests are coming up during the month of March. It's going to be a great month for our show.

Thanks to Andrew Zarian back at the GFQ network, the network that is full of all kinds of interesting shows, including What the Tech? the Friday Free for All, Mat Men, and more. Stay tuned on the GFQ network for lots of interesting podcasts. Andrew, thank you for doing the show.

We'll see you next week on This Week in Radio Tech. Bye-bye, everybody.

Telos Alliance has led the audio industry’s innovation in Broadcast Audio, Digital Mixing & Mastering, Audio Processors & Compression, Broadcast Mixing Consoles, Audio Interfaces, AoIP & VoIP for over three decades. The Telos Alliance family of products include Telos® Systems, Omnia® Audio, Axia® Audio, Linear Acoustic®, 25-Seven® Systems, Minnetonka™ Audio and Jünger Audio. Covering all ranges of Audio Applications for Radio & Television from Telos Infinity IP Intercom Systems, Jünger Audio AIXpressor Audio Processor, Omnia 11 Radio Processors, Axia Networked Quasar Broadcast Mixing Consoles and Linear Acoustic AMS Audio Quality Loudness Monitoring and 25-Seven TVC-15 Watermark Analyzer & Monitor. Telos Alliance offers audio solutions for any and every Radio, Television, Live Events, Podcast & Live Streaming Studio With Telos Alliance “Broadcast Without Limits.”