artificial reverberation: from mono to true stereo

“True stereo” is a term used in audio processing to describe a stereo recording, processing or playback technique that accurately represents the spatial location of sound sources in the stereo field. In true stereo, the left and right channels of a stereo recording contain distinct and separate audio information that accurately reflects the spatial location of sound sources in the recording environment.

This is in contrast to fake/pseudo stereo, where the stereo image is created through artificial means, such as by applying phase shifting techniques to create the impression of stereo. True stereo is generally considered to be superior to fake stereo, as it provides a more natural and immersive listening experience, allowing the listener to better locate and identify sound sources within the stereo field. In the domain of acoustic reverberation, this is essential for the perception of envelopment.

Artificial reverberation has come a long way since its early beginnings. The first mechanical devices for generating artificial reverberation, such as spring or plate reverberation, were initially only available as mono devices. Even when two-channel variants emerged, they usually did summing to mono internally or did processing in separate signal paths, known as dual mono processing. Typically, in a plate reverb, a two-channel output signal was achieved simply by mounting two transducers on the very same reverb plate.

The first digital implementations of artificial reverberation did not differ much from the mechanical ones regarding this principle. Quite common was summing the inputs to mono and the independent tap of two signals from a single reverb tank to obtain a two-channel output. Then, explicit early reflection models were added, which were typically processed for left and right separately and merged into the outputs later to preserve a basic representation of spatial information. Sometimes, also the first reflections were just taken from a (summed) mono signal. The Ursa Major 8×32 from 1981 is a good example for this design pattern. Later, the designs became more sophisticated, and even today it is common to distinguish between early and late reverberation in order to create a convincing impression of immersion.

However, ensuring proper sound localisation through early reflection models is a delicate matter. First and foremost, a real room does not have a single reflection pattern, but a vast variety of ones that depend on the actual location of the sound source and the listening position in that room. A true-to-life representation of this would, therefore, have to be represented by a whole set of individual reflection patterns per sound source and listening position in the virtual room. As far as I know, the VSL MIR solution is the only one that currently takes advantage of this, and with an enormous technical effort.

Another problem is that first reflections can also be detrimental to the sound experience. Depending on their frequency and delay in relation to the direct signal, the direct signal can be masked and affected in terms of phase coherence so that the overall sound becomes muddy and lacks clarity. This is one of the reasons why a real plate reverb is loved so much for its clarity and immediacy: it simply has no initial reflections in this range. As a side note, in the epicPLATE implementation, this behaviour is accurately modeled by utilizing a reverberation technique that completely avoids reflections (delays).

Last but not least, in a real room there is no clear separation between the first reflections and the late reverberation. It is all part of the same reverberation that gradually develops over time, starting with just an auditory event. This also means that there is no clear distinction between events that can be located in space and those that can no longer be identified – this also continuously evolves over time.

A good example of how to realise digital reverb without this kind of separation between early and late reverberation and at the same time in “true stereo” was impressively demonstrated by the Quantec QRS back in the early 80s already. Its ability to accurately reproduce stereo was one of the reasons why it became an all-time favourite not only in the music production scene, but also in post-production and broadcasting.

Artificial reverberation is full of subtleties and details and one might wonder why we can perceive them at all. In the end, it comes down to the fact that in the course of evolution there was a need for such fine-tuning of our sensory system. It was a matter of survival and important for all animal species to immediately recognise at all times: What is it and where is it? The entire sensory system is designed for this and even combines the different sensory channels to always answer these two questions. Fun Fact: This is exactly why some visual cues can have a significant impact on what is heard and why blind tests (in both meanings) are so important for assessing certain audio qualities. See also the “McGurk Effect” if you are interested.

Have fun listening!

something epic is coming

Stay tuned!

the twisted world of guitar pedals III

According to urban legends, the Inuit have more than a dozen words for one and the same thing: snow. But forget that, it’s nothing, really. The ambitious modern guitarist is guaranteed to know a multiple of different words for one and the very same phenomenon: distortion. Seriously, guys, who comes up with terms like “clean distortion” or “transparent overdrive”? The other day, I was searching the online store of my trusted dealer for distortion pedals and got 632 hits.

Anyway, somewhere in the sheer mass of pedals and terminology, the really interesting concepts and devices are lurking. And from my humble explorations in the twisted world of guitar pedals, I mostly found them in the realm of boutique manufacturers. Small companies that have often gained a large following among guitarists and are looking for new approaches and distinctive sounds that cannot be found in the mass-produced products of the big manufacturers.

They not only experiment with new ideas and designs that larger companies may overlook, but also take the risk of being successful with niche products. They skillfully combine analog and digital technology to reinterpret classic effects or bring entirely new concepts to the table. You can literally tell that many of these companies are run by musicians who know the needs and wishes of guitarists or just are willing to work closely with customers to develop very individual or extraordinary pedals.

For me personally, some of these boutique devices and manufacturers are a true source of inspiration, whether as a musician or plugin designer. And I can only hope that they continue on their chosen path, survive difficult economic times, and above all, don’t get swallowed up by the big boys.

 

the twisted world of guitar pedals I

the twisted world of guitar pedals II

sustaining trends in audio land, 2023 edition

So, the year 2023 is slowly getting underway – time to take another look at the sustaining trends in audio land. Two of the 2022 themes have already been further confirmed and manifested – so let’s take a quick look. A third topic, however, has developed at an incredible speed and in an unbelievable way, to the surprise of all of us. But one thing at a time.

The (over-) saturated audio plugin market

A continuing trend towards market consolidation was to be expected as a result of a constantly oversaturated market, and indeed last year saw a whole series of merger and acquisition activities as well as new alliances. Involved in such activities were brands such as Slate Digital, Sonnox, Focusrite, Brainworx, Plugin Alliance, NI, iZotope and many more.

This is something quite normal in saturated markets and not a bad thing per se, but we might worry about a lack of innovation and diversity as a result. Alongside this, we will continue to see many companies late to the party offering “me too” products and retro brands gilding their HW brands with yesterday’s SW technology. The smarter companies will continue their efforts to successfully establish leading business platforms.

The future of HW DSP as a plugin platform

Since the HW DSP market has not succeeded in creating such a competitive (plug-in) business platform, we are currently witnessing the decline of this domain and in the long run everything will be offered natively. Last year, we’ve seen some late movers also starting such transformations, e.g. UA.

The emergence of AI in audio production

Of course, this was not only predictable but also announced, but no one had ever expected the extent and speed of its emergence over the past year. This applies first and foremost to its appearance in general, but also its impact to the whole music domain in particular. This impact will be immense and dramatic, affecting not only tools and work processes, but also music culture and its economy. The effects will be very, very profound, similar to the way the internet entered all areas of our lives.

The current trend of emulating effect devices with deep learning seems less exciting in this context, as it is just yet another form of effect sampling where we might see little innovation. Much more exciting will be the impact on areas such as composition, mixing and mastering, but also music distribution and value creation in general. But that will be the subject of another detailed article in this Blog.

We live in exciting times.
Stay tuned!

---

sustaining trends in audio land, 2022 edition

sustaining trends in audio land, 2021 edition

 

 

Korg Retrospective

sidechain linking techniques

How an audio compressor responds to stereo content depends largely on how the channel linking is implemented in the sidechain. This has a major influence on how the spatial representation of a stereo signal is preserved or even enhanced. The task of the compressor designer is to decide which technical design is most suitable for a given overall concept and to what extent the user can control the linkage when using the device.

In analog compressor designs, in addition to unlinked “dual mono” operation, one usually finds simple techniques such as summing both stereo channels (corresponding to the center of the stereo signal) or the extraction of the maximum levels of both channels using a comparator circuit implementing the mathematical term max(L,R).

More sophisticated designs improve this by making the linking itself frequency dependent, e.g. by linking the channels only within a certain frequency range. It is also common to adjust the amount of coupling from 0 to 100%, and the API 2500 hardware compressor serves as a good example of such frequency dependent implementation. For the low and mid frequency range, simple summing often works slightly better in terms of good stereo imaging, while for the mid to high frequency range, decoupling to some degree often proves to be a better choice.

The channel coupling can also be considered as RMS (or vector) summing, which can be easily realized by sqrt(L^2+R^2). As an added sugar, this also elegantly solves the rectification problem and results in very consistent gain reduction across the actual level distributions that occur between two channels.

If, on the other hand, one wants to focus attention on correlated and uncorrelated signal components individually (both of which together make up a true stereo signal), then a mid/side decomposition in the sidechain is the ticket: A straight forward max(mid(L,R), side(L,R)) on the already rectified channels L and R is able to respond to any kind of correlated signal not only in a very balanced way but also to enhance its spatial representation.

More advanced techniques usually combine the methods already described.

that unique plate reverb sound

Unlike digital reverberation, the plate reverb is one of the true analog attempts in recreating convincing reverberation build right into a studio device. It is basically an electro-mechanical device containing a plate of steel, transducers and a contact microphone to pickup the induced vibrations from that plate.

The sound is basically determined by the physical properties of the plate and its mechanical damping. Its not about reflecting waves from the plates surface but about the propagation of waves within the plate. While the plate itself has a fixed, regular shaped size and can be seen as a flat (two dimensional) room itself it actually does not produce early reflection patterns as we are used to from real rooms with solid walls. In fact there are no such reflections distinguishable by human hearing. On the other hand there appears to be a rather instant onset and the reverb build-up has a very high modal density already.

Also reverb diffusion appears to be quite unique within the plate. The wave propagation through metal performs different compared to air (e.g. speed/frequency wise) and also the plate itself – being a rather regular shape with a uniform surface and material – defines the sound. This typically results in a very uniform reverb tail although the higher frequencies tend to resonate a little bit more. Also due to the physics and the damping of the plate, we usually do not see hear very long decay times.

All in all, the fast and consistent reverb build up combined with its distinct tonality defines that specific plate reverb sound and explains why it is still so much beloved even after decades. The lack of early reflections can be easily compensated for just by adding some upfront delay lines to improve stereo localization if a mix demands it. The other way around, the plate reverb makes a perfect companion for all kinds of delay effects.

everything just fades into noise at the end

When I faced artificial reverberation algorithms to the very first time I just thought why not just dissolve the audio into noise over time to generate the reverb tail but it turned out to be not that easy, at least when just having the DSP knowledge and tools of that time. Today, digital reverb generation has come a long way and the research and toolsets available are quite impressive and diverse.

While the classic feedback delay network approaches got way more refined by improved diffusion generation, todays computational power increase can smooth things out further just by brute force as well. Still some HW vendors are going this route. Sampling impulse responses from real spaces also evolved over time and some DSP convolution drawbacks like latency management has been successfully addressed and can be handled more easily given todays CPUs.

Also, convolution is still the ticket whenever modeling a specific analog device (e.g. a plate or spring reverb) appears to be difficult, as long as the modeled part of the system is linear time invariant. To achieve even more accurate results there is still no way around physical modeling but this usually requires a very sophisticated modeling effort. As in practise everything appears to be a tradeoff its not that much unusual to just combine different approaches, e.g. a reverb onset gets sampled/convoluted but the reverb tail gets computed conventionally or – the other way around – early reflections are modeled but the tail just resolves into convoluted noise.

So, as we’ve learned now that everything just fades into noise at the end it comes to no surprise that the almost 15 years old epicVerb plugin becomes legacy now. However, it remains available to download for some (additional reverb) time. Go grab your copy as long as its not competely decayed, you’ll find it in the downloads legacy section here. There won’t be a MkII version but something new is already in the making and probably see the light of day in the not so far future. Stay tuned.

BootEQ mkIII released

BootEQ mkIII – a musical sounding Preamp/EQ

BootEQ mkIII is a musical sounding mixing EQ and pre-amplifier simulation. With its
four parametric and independent EQ bands it offers special selected and musical
sounding asymmetric and proportional EQ curves capable of reproducing several
‘classic’ EQ curves and tones accordingly.

It provides further audio coloration capabilities utilizing pre-amplifier harmonic distortion as well as tube and transformer-style signal saturation. Within its mkIII incarnation, the Preamp itself contains an opto-style compression circuit providing a very distinct and consistent harmonic distortion profile over a wide range of input levels, all based now on a true stateful saturation model.

Also the EQ curve slopes has been revised, plugin calibration takes place for better gain-staging and metering and the plugin offers zero latency processing now.

Available for Windows VST in 32 and 64bit as freeware. Download your copy here.

sustaining trends in audio land, 2022 edition

Forecasts are difficult, especially when they concern the future – Mark Twain

In last years edition about sustaining trends in audio land I’ve covered pretty much everything from mobile and modular, DAW and DAW-less up to retro outboard and ITB production trends. From my point of view, all points made so far are still valid. However, I’ve neglected one or another topic which I’ll now just add here to that list.

The emergence of AI in audio production

What we can currently see already in the market is the ermergence of some clever mixing tools aiming to solve very specific mixing tasks, e.g. resonance smoothing and spectral balancing. Tools like that might be based on deep learning or other smart and sophisticated algorithms. There is no such common/strict “AI” definition and we will see an increasing use of the “AI” badge even only for the marketing claim to be superior.

Some other markets are ahead in this area, so it might be a good idea to just look into them. For example, AI applications in the digital photography domain are already ranging from smart assistance during taking a photo itself up to complete automated post processing. There is AI eye/face detection in-camera, skin retouching, sky replacement and even complete picture development. Available for all kinds of devices, assisted or fully automated and in all shades of quality and pricing.

Such technology not only shapes the production itself but a market and business as a whole. For example, traditional gate keepers might disappear because they are no longer necessary to create, edit and distribute things but also the market might get flooded with mediocre content. To some extend we can see this already in the audio domain and the emergence of AI within our production will just be an accelerator for all that.

The future of audio mastering

Audio Mastering demands shifted slightly over the recent years already. We’ve seen new requirements coming from streaming services, the album concept has become less relevant and there was (and still is) a strong demand for an increased loudness target. Also, the CD has been loosing relevance but Vinyl is back and has become a sustaining trend again, surprisingly. Currently Dolby Atmos gains some momentum, but the actual consumer market acceptance remains to be proven. I would not place my bet on that since this has way more implications (from a consumer point of view) than just introducing UHD as a new display standard.

Concerning the technical production, a complete ITB shift – as we’ve seen it in the mixing domain – has not been completed yet but the new digital possibilities like dynamic equalizing or full spectrum balancing are slowly adopted. All in all, audio mastering slowly evolves along the ever changing demands but remains surprisingly stable, sustaining as a business and this will probably continue for the next (few) years.

Social Media, your constant source of misinformation

How To Make Vocals Sound Analog? Using Clippers For Clean Transparent Loudness. Am I on drugs now? No, I’ve just entered the twisted realm of social media. The place where noobs advice you pro mixing tips and the reviews are paid. Everyone is an engineer here but its sooo entertaining. Only purpose: Attention. Currency: Clicks&Subs. Tiktok surpassed YT regarding reach. Content half-life measured in hours. That DISLIKE button is gone. THERE IS NO HOPE.

The (over-) saturated audio plugin market and the future of DSP

Over the years, a vast variety of vendors and products has been flooded the audio plugin market, offering literally hundreds of options to choose from. While this appears to be a good thing at first glance (increaed competition leads to lower retail prices) this has indeed a number of implications to look at. The issues we should be concerned the most about are the lack of innovation and the drop in quality. We will continue to see a lot of “me too” products as well as retro brands gilding their HW brands with yesterday SW tech.

Also, we can expect a trend of market consolidation which might appear in different shapes. Traditionally, this is about mergers and aquisitions but today its way more prominently about successfully establishing a leading business platform. And this is why HW DSP will be dead on the long run becuse those vendors just failed in creating competitive business platforms. Other players stepped in here already.