You might’ve noticed on our site, or read in our newsletter, we started offering some new file formats on the website. Not only are we re-releasing albums in three flavors of DSD (2.8, 5.6 and 11.2 MHz, or DSD 64, 128, or 256 respectively), but we also introduced – as far as we know as a world’s first – the DXD 32bit format. So what’s this new format all about, what do (or don’t) we do with it, and why you should you get it? Let’s dive deeper, shall we? Let’s start with DXD So let’s start with the easiest part: DXD. As we’ve mentioned in our previous More = More? blog, DXD (Digital eXtreme Definition) is a format developed by Merging Technologies that allows you to record for DSD masters in the highest quality possible. How that works, we’ll explain in the next More = More? blog, by the way. In short, DXD is audio sampled at 352.8 kHz, so 352.800 samples per second. A lot more than CD quality, which is sampled at 44.1 kHz (eight times less). But DXD is not just great for DSD masters, it’s actually a great format in itself as well! Why? As you can read in our previous blog, at such extremely high sample rates, the need for filtering is pretty much non-existent, making it possible to use very slow filters in your AD and DA converters. This, in turn, results in much better time-domain response, meaning that more percussive elements in the music will be recorded and played back much snappier and more naturally. Music recorded – and, preferably, played back – in DXD will sound more natural, dynamic, and overall will have more “slam and snap” to it than other formats. Great, right? So how can we improve? For that, we have to take a look at ones and zeroes, or, in fancier terms, bits. All the right bits in all the right places Digital audio is stored in bits, representing the amplitude of the signal at a specific moment in time. The sample rate (e.g. 352.8 kHz) determines how many of those bits are recorded and played back every second. So how many bits is stored for every sample? Well, that depends. At CD quality, each sample carries 16 bits of fixed-point data (I can already hear you say: “wait, what’s fixed-point?!” but we’ll get to that), which basically means that it can store any integer value between 0 and 65.535. Technically (we’re not getting into noise shaping just yet, we will in our next More = More? blog), you can fit 96 dB of dynamic range in just these 16 bits. If we go one step higher, we reach 24 bits per sample. This is the resolution almost all (let’s say 99.9%) of recordings are made at. And up until a couple of years ago, we did as well. Now, the 24bit format has a lot of interesting upsides for both recording and playback. 24 bits of fixed-point data can store any integer value between 0 and 16.777.216, or 256 times as much as the 16bit format can, reaching a theoretical dynamic range of 144 dB. Fantastic, right? Yes! But… Introducing the 32bit format. I can already see you grabbing your calculator and discovering that 32 bits can hold over 4.2 billion integer values, or reach 192 dB of dynamic range… And yes! It does, as a fixed-point value. But, we told you before we’ll talk about what fixed-point means, because it gets interesting here… Floating your points all over the place As we discussed before, the values we talked about are all fixed-point, meaning that they basically just store their data in a binary fashion (for example, the number 13 as a 16-bit fixed-point integer looks a bit like this: 0000000000001101). But there’s a more efficient way of storing way bigger numbers, and it’s called floating point values. We’re not getting into the nitty-gritty mathematics of things here, but here’s the ELI5 of it: floating-point arithmetic was created as a perfect trade-off between range and precision. It uses three components: a significand, a base, and an exponent. See the image below to get a feeling of how that works: So what’s this useful for? Well, actually for a lot of things. You can use the same format to store the distance between two galaxies as well as the distance between two atoms in a molecule. It’s that flexible, and the scale can be huge. Then, what does this all mean for audio? We’ll get more into why it could be useful in just a bit, but simply put: you can store data in it that has both extreme precision as well as giant scale! Sounds like it can be very useful for audio, right? Right. Basically put, you can be both extremely precise in your calculations on the audio, as well as having an extremely huge dynamic range. And when I say huge, I mean huge. Like blade-of-grass-falling-on-snow to supernova explosion huge. To be precise, a 32bit floating point audio stream can reach a dynamic range of 1.528 dB. And while these numbers are definitely overkill for any scenario life can throw at a mortal sound engineer, they are great for the fact that you have a huge headroom for DSP. A DSP’s native tongue DSP, you hear the term everywhere. Digital Signal Processor for long. What a DSP does is pretty simple: it processes digital signals such as audio. This can be anything from a digital volume control to an equalizer, or pretty much anything else. A lot of things have DSPs in them. Your favorite DAC, for example, might make use of a DSP to perform calculations on the incoming audio stream to optimize the outgoing one, or as mentioned above, a digital volume control. But your DAC’s not the only thing with a DSP in it. In fact, the entire recording chain we (or any other studio, for that matter) use, has DSPs all over the place. And while we use as little DSP as we can (per our recording philosophy), using DSP is unavoidable in any digital recording. But a DSP doesn’t have to be all that bad, or destructive to your audio! In fact, if it uses floating-point signals/values rather than fixed-point, there’s ample headroom to do so, and as long as the conversion between fixed-point values and floating-point values happens perfectly, then nothing’s wrong. This is why pretty much all modern DSPs, from software ones that run on your computer’s processor, to dedicated pro-audio ones, all run on floating-point values. 32bit floating-point values are, to put it very simply, the DSP’s native tongue. It’s what it likes to speak and is able to speak it best. If a DAC with a DSP with an internal (32bit) floating-point architecture (which, again, applies to most DACs) receives a 32bit floating-point signal, it doesn’t have to convert anything at all, right until the moment it’s converted into the analog domain. If you throw a (24bit) fixed-point signal into it, it has to first convert that signal into a 32bit floating point one before it can work its magic, and then back again right when it converts into analog. Also, and this is mostly interesting for us sound engineers, 32bit floating-point recording, editing, mixing, and mastering allows us to use the full potential of the audio and its dynamic range, use equalization or other tools, and never ever have to worry about the dynamic range. Every single bit that was there in the recording is still there. Goodbye, dithering Without getting all technical on you (we’ll save that for a later More = More? blog), dithering is what a mastering engineer has to do to the audio in order for the master to sound right at 16bit or 24bit. This dithering is a tiny bit of noise (yeah, really. You add noise to make it sound better. Weird, right? Just read our next blog in the More = More? series and you’ll get why) to help against rounding errors on the so-called least significant bit of the audio stream. With 32bit floating point audio, there’s no need for dithering because, well, there are practically no rounding errors in the first place. This means that from the first bit entering the recording system, to the last bit leaving the mastering system, nothing is added or destroyed. Which is exactly our recording philosophy here at TRPTK. When (not) to get the 32bit floating point DXD master at TRPTK.com So, all in all, 32bit floating point DXD sounds like a blessing, right? Well, yeah. To us, absolutely. To you, your mileage may actually vary. Which is why we’ll list some things to consider before making your purchase on our platform. We want you to enjoy our music in the best possible format, and sometimes that simply is the 24bit DXD master instead of the 32bit floating point one. You should definitely get the 32bit floating point DXD master if: you’re an audiophile who just wants the best, clearest, most transparent signal path possible;AND your DAC has a floating-point signal path or DSP built in (this actually applies to a lot of high-end DACs, even though it’s seldom mentioned in the manual);OR if you’re planning to do some of your own DSP, such as room correction, up or downsampling, equalizing, or other. Conversely, you should probably choose the 24bit regular DXD master if: your streamer or streaming software doesn’t support 32bit floating point audio;OR your DAC has a 16bit or 24bit fixed-point architecture all throughout;OR your DAC or streamer adds its own dithering you don’t like;OR you just prefer FLAC files (floating-point audio isn’t supported in FLAC, which is why the 32bit DXD masters are supplied as WAV files). What albums should I get in 32bit DXD? Great question! Right now, we have four albums and an EP available as 32bit floating point DXD masters, but we’re adding new ones regularly, as well as re-releasing older titles in the format. So stay tuned, or just click this link to view all 32bit floating point DXD masters. Questions? Just ask us. We can imagine, after reading this massive blog post, you might still have some questions. Not to worry! Our team is always happy to answer any of your questions you might have. Just write us at firstname.lastname@example.org and we’ll try to get back to you within a day or two. Happy listening! Posted by Brendon Heinst. Brendon is the founder and senior recording & mastering engineer at TRPTK. He gained a Bachelor's and Master's Degree in audio engineering at the University of the Arts Utrecht and Open University of London. Brendon was involved in more than 200 recordings to date, focusing heavily on ultra-high-resolution and multichannel immersive recordings.