To many, the experience of listening to speakers isn’t just different to headphones, for most people listening to good speakers in a good room is better but while acceptably good speakers aren’t difficult to get, a good sounding room can be more of a challenge. This is where binaural rendering seems to offer an ideal solution - the advantages of speakers with the convenience of headphones. However many people who have tried these generic binaural renderers, which offer virtualised mix rooms, come away underwhelmed with the experience.

The issue with binaural is that the experience is so inconsistent between listeners. This is because of the HRTF or head related transfer function, a way to capture the directional cues perceived by a listener which help them perceive the direction from which sounds arrive. In most products this is a generalised HRTF based on an average ear and body shape. However in the real world the colourations which provide this essential extra information which helps us locate sounds are unique to our own physiology. The ideal way to get the best possible experience from a binaural render is to use it with an HRTF based on the precise physiology of the listener. I wrote an article on one such solution - Genelec’s Aural ID, which is a way to create personalised HRTFs from photographs and video of the listener’s head, shoulders and ears.

The question was always that, while that was fine in theory, what did it sound like in practice? Well I’ve now tried it. This is going to be one of those times when you’ll just have to take my word for it as there is no way you can experience a personalised HRTF without having one created based on your unique physiology. But I have tried it and if you’re in a hurry we’ll just get this out of the way right now - it works!

Genelec Aural ID

I’ve been testing Genelec’s Aural ID for a couple of days, it’s an interesting product as unlike almost every plugin there is no demo available. There can’t be because you have to have a custom HRTF created to hear it.

The HRTF creation process is straightforward and relatively easy. Sign up for an account after downloading the Aural ID app to your phone and you’ll be guided through the process which involves taking a photo of each of your ears holding a ruler over them and then you’ll need a friend to help create a video of your head, shoulders and ears. The process is guided but keep in mind that you do need quite a lot of space to create this video. A typical UK house might not have a clear space large enough for this walkround, 360 degree video. Mine certainly doesn’t and, rather than move the furniture around, my neighbours probably wondered why I was sitting on a stool in the garden while my ever-patient partner created the video.

After reviewing the photos and video you upload them and when the magic is complete you’ll get a notification and you can download your personalised HRTF. Of course there is the small matter of payment and it has to be said that Aural ID isn’t cheap. Pricing below.

The plugin looks much like the GLM4 speaker calibration software with the dark grey hexagons of the former carried over into this UI. The two can work together with information about a speaker array capable of being imported into the Aural ID plugin so a virtual rendering of your monitors can be available via headphones.

Setting up the plugin is simple enough with the custom HRTF file, a .aid file, being imported into and parsed into the plugin from the cloud or from a local copy. I was a little disappointed that I couldn’t inspect data about my personalised HRTF but I’m not sure seeing such data would have been particularly useful. The plugin works in greater than stereo channel widths, up to 7.1.2 in Pro Tools but because of the nature of Objects in Dolby Atmos workflows you’ll have to make sure your workflow accommodates using a plugin like this. I used it principally in stereo, with a few experiments with 5.1 and Atmos, and once I’d made something of a mental leap I got on with the Aural ID extremely well. 

Aural ID In Use

In use it’s straightforward enough. You instantiate the plugin in an appropriate place for headphone monitoring, you select the appropriate channel width and load you PHRTF. The first time you do this you have to download it but you should store it locally after that for convenience. There is a list of models of professional headphones which have been measured and I was pleased to see I had one of those models, the Neumann NHD20s. I selected those and got busy auditioning.

The localisation was simply unlike anything I’d experienced before, with pin sharp accuracy revealing where each sound was. None of the crowding round the ears or just “somewhere inside my head” of regular headphones. I expected toggling bypass would be suitably pleasing but I became very aware that with this excellent localisation came a dramatic timbral change which was just too much.

I contacted Genelec and after discussing it with them I think I had been being too inflexible in my approach to the setup. I’d noticed that using my regular Austrian Audio Hi-X65s with the ‘Other’ option selected in the headphone model sound better and on talking to the wise people at Genelec they said that this was very possible and recommended that I tweak the response of the binaural render using the built in EQ. This took a little processing on my part but actually there’s loads of clever stuff going on in this render and if it sounds peaky in the midrange to your ears then just EQ it.

Now feeling like I had permission to tweak the response I quickly settled on a few moves which made things much better and I got on with mixing.

Mixing with Aural ID

This is where things got more interesting. The frequency response is different using Aural ID with my, and presumably anyone else’s, PHRTF compared to running the same headphones without Aural ID. However as soon as you put some monitors, with their ruler-flat response as measured in an anechoic chamber, in an actual room things change. The thing which matters is whether or not you can mix on them and whether your mixes are better using it.

After a couple of revisits to troublesome mixes I found that actually things turned out pretty well. I really dislike mixing on headphones and I found I dislike mixing with headphones using Aural ID a lot less. It’s an odd experience which you get used to reasonably quickly and for this reason I suggest that however tempting it is you don’t get too far into questioning what it is doing. You’ll end up questioning everything! The accurate localisation really reveals the details and all of those things I associate with strikingly good monitoring, like immersion in the stereo field and reverb tails being laid bare, are all there.

That is in stereo though, what about surround? I spent a happy couple of hours remixing a stereo session to 5.1 and then doing a fake upmix from stereo to 7.1.2 using Liquidsonics’ Cinematic Rooms. I’ve tried binaural renders from the Dolby Atmos Renderer, though I should say I am yet to try the recently introduced PHRTF feature for the Dolby Atmos Renderer, but comparing the localisation achieved with Aural ID I have to say that I have never before heard sounds unambiguously sound like they are coming from behind me. With a suitable cue to tell me that a sound is behind rather than in front of me vanilla binaural renders work fine but on their own, without a visual or contextual cue to prompt my perception they have never been reliable. Not only does Aural ID communicate front/back position clearly but it also communicated height very effectively. I’m impressed.

Is Aural ID Good Value?

Aural ID is aimed at the serious user and its interface is based around GLM4. The ideal use I can see for this, which would easily justify the costs, is as a carefully set up alternative to a surround or Atmos room which would allow much of the work to be done away from the mix room. Because of the way Objects work in an Atmos project, the practical way to use Aural ID in at atmos project would be to insert it in a re-rendering path.

Is there is a downside to Aural ID? Firstly it is undeniably not a cheap product. It is a subscription product which costs €490 for the first year, €249 for the second year and €149 thereafter. However, if it offers a way for a business to get more work done than their available mix rooms permit, it is easy to see how it could be seen as excellent value, it just depends on whether Aural ID fixes a problem you actually have. I heard about an in-house solution similar to this which allowed the owners of a big mix stage to do far more work than their mix spaces allowed by virtualising them in the same way as Aural ID allows. For them it made sense to create their own!

The experience of Using Aural ID can be, as previously mentioned, disorientating to begin with. Stick with it and don’t AB against regular headphones - it makes it worse! Listen and if you feel you need to tweak the EQ, just do it, it’s why it’s there. If you feel you need a reference compare it to your monitors, not to regular headphones. In the final analysis I thought mixing with Aural ID created better mixes than mixing just on regular headphones and that’s ultimately the point.

The whole area of PHRTF binaural rendering is definitely an area which is seeing a burst of activity. Genelec have been in this area longer than anyone I’m aware of and while this is an area which is by its nature frustratingly unique, the detail involved in the capture of the subject’s ears and body shape suggest to me that in a typically Genelec way they are doing it properly. There are other alternatives out there but if Aural ID does something you think you need, I suspect no one is doing it better.

Photo by Johnny Mckane