Vienna MIR: secret tips from its creator – part 1

On a warm day in the summer of 2021 I had the lovliest zoom call with Dietz Tinhof. Dietz is none other than the inventor of Vienna MIR Pro, the super powerful holistic mixing and reverberation engine for Mac and PC. During our chat we talked about the origin and development of MIR, we discussed several amazing secret tips and Dietz even shared some info about the upcoming evolution of MIR, MIR 3D. I hope you’ll enjoy reading this tips-filled chat as much as I enjoyed having it!

From concept to reality

Giovanni: Can you share the story of how you had the idea for MIR and what it was like to transform that idea into reality?

Dietz: That the basic idea of MIR is actually quite simple. Back in the days when Vienna Symphonic Library was founded by Herb Tucmandl, it became very clear pretty soon that there was one instrument missing from the whole collection: the room. In fact, there is hardly an orchestra playing without the actual surroundings of a great orchestral hall. This “meta-instrument” deserved the same attention we put on the actual instruments. Like our solo violins, with several ten-thousands of individual samples just for the main patches. Back in the years 1999 or 2000, impulse response based reverbs existed already, so you were actually able to make samples from rooms. But it was still very basic stuff, much like taking the middle C of a piano and stretching it across the whole six, eight octaves and saying “Ok, now you can play Chopin with it”. You know very well that wouldn’t sound good at all.

Giovanni: That’s a great analogy actually.  

Dietz: [laughs] And that’s why we started to think about the possibility to do justice to such a great instrument. And then I came up with this idea that when taking samples from rooms we’d wouldn’t limit that to a single perspective, but record many. Furthermore, we wouldn’t capture just one position where the impulse is created, but many positions and maybe not just one direction, but many directions. In other words we were talking about multiple impulse responses and this is where the name MIR came from. 

Giovanni: Ahh I see

Dietz: The abbreviation was just a project name initially, but then we stuck to it.

Giovanni: Well I like it a lot, I must say. It has something. It reminds me of something space-bound.

Dietz: You are probably thinking of the Russian space station which was the core idea of the ISS, too.

Giovanni: That’s right. I think it’s a great name. I’m thinking there must be lots of science and acoustics behind creating MIR and an impulse response reverb where you can actually move the microphone. You must have done lots of homework.

Dietz: In the end it’s not so complicated as it might seem. It’s just audio engineering. You approach a room like an audio engineer. You go into the hall, you listen to it, maybe you clap your hands, you make funny noises, you listen to the reflections and then you think: Ok there is the audience, there is the orchestra, where do I put my microphones? I wanted to present it in a very playful way so that you can move around stuff easily and things that happen on your interface always relate to what you hear. I just threw together several well-known concepts and made MIR out of it. 

Giovanni: How long was the whole process from the initial idea to first commercial version of it. 

Dietz: Very long. I’m very very happy that Herb had the stamina to go through this because other people would have given up after two years. But we defined the idea around 2002, and we had a first proof of concept around 2005 I think, or maybe 2004. We had a first version of the engine (that didn’t actually work) around 2007. And the first version of MIR was released in 2009. So about six, seven years.

Giovanni: That’s amazing and it shows that you really had a clear vision. To go through all these years but keep going means you knew you had something precious.

Dietz: I had three very important partners, aside from Herb Tucmandl. One of them is no longer a part of VSL but still a very good friend of mine by the name of Martin Rajek. He is the chief architect of MIR. I had crazy ideas, lots of crazy ideas, and he put hem together. I had the vision, the audio engineering idea, but the basic concept is very much Martin’s. Then there was Florian Walter who developed both the convolution core that powers MIR and its Ambisonics innards. And finally there is another very important Martin, who is now the chief software developer for VSL, Martin Saleteg. He had this brilliant idea of overcoming the biggest hurdle of MIR: the fact that we needed thousands and thousands of convolutions, actually. But convolution is a mathematical process that is very heavy on the CPU. Martin found a way to bring down the number of those thousands of convolutions to a bare minimum. So nowadays we put the representation of a signal on the stage, and MIR immediately looks up for the closest impulse responses relative to the position. Then it creates a new set of impulse responses, but only four of them (in case of stereo) instead of thousands.

Giovanni: That’s amazing teamwork!

Dietz: These four people are the ones who made MIR happen more or less.

Buffer size

Giovanni: Now I understand how it is possible that MIR runs so well on my MacBook pro. It is a recent machine but my template is quite huge and I was a bit worried when I bought MIR and the room pack. It was a little bit jumping in the dark for me. I didn’t know if the computer would support it, but it works great.

Dietz: There’s one little known trick that MIR has a buffer size of it’s own in addition to the systems latency. When you are done with composing and you start mixing push it up to 2048 or even more and you will gain a lot from it. Convolution is much less demanding when the computer has time to go through it. This is something people forget: there is a dedicated buffer size for MIR alone, not for the system, but in addition to it. So if you run into a bottle neck just think about it. It’s in the preferences and easy to access, actually. 

MIR and third party instruments

Giovanni: What are the best practices when one is using third-party instruments with Vienna MIR?

Dietz: There’s just a few things to be kept in mind. One of them is that it’s quite important to send MIR a perfectly centered signal. As soon as there is signal inherent panning, MIR actually loses the possibility to position it. It will be somehow biased, you know, towards the inherent panning. The other thing has to do with something that you might have seen in MIR: We call it Instruments Profile. These Instruments Profiles are very important because they define how the signal is sent into the room. As an example: a trumpet would project mostly to the front, while a bass drum would be more or less an omni signal component.

We measured VSL instruments very carefully. We created detailed directivity profiles that are unique to these instruments. When you apply them to any other source, you might be actually losing information. People are sometimes tempted to use them on third-party instruments too, because they look so nice with their instrument pictures, but they should be used exclusively with Vienna Instruments. The main reason being: these profiles take into account the direction the instruments were recorded from and offset them to reflect an accurate direction according to the user’s settings.

Giovanni: This makes complete sense. I must, ok I confess I think I did use those profiles with non-Vienna Instruments…

Dietz: It’s not a sin, you know but you might have lost a little bit of definition in your mix. You’re probably blurring things unwillingly.

Giovanni: Yeah, and you know, the manual did a great job saying don’t use them but now that I know why I shouldn’t I’m gonna make sure I won’t use them in the future.

Check the next part for tips on adjusting the dry/wet balance, grouping, blending libraries and room EQ, and much more!