In the early days of digital audio, 44.1kHz sampling with 16 bits of resolution was hard. It was so hard that when the first two CD players were released, one from Sony and one from Philips – the two companies had collaborated in the development of the CD – the Philips one only ran at 14 bits. It just ignored the two least significant bits.*
I purchased the first Sony one, the CDP-101. It did the full 16 bits, but used one DAC for both channels, staggering them. It cost me $1,200. According to the Reserve Bank of Australia’s inflation calculator, that’s just short of $4,000 in 2020 dollars.
That was back in the early 1980s. These days for less than $250 you can purchase a digital to analogue converter that can decode – without having to rapidly alternate between channels – digital audio sampled at 768kHz with 32 bits of resolution. Nowadays, digital to analogue conversion is easy.
How about some test material?
But I had – note the past tense – a problem. With assiduous searching of the Internet I’d found tracks with resolutions all the way up to 384kHz and 24 bits, and in Direct Stream Digital all the way up to DSD512 (that is, eight times the resolution of the original DSD format, the Super Audio CD). I wanted these tracks not so much for glorious high fidelity listening, but just as mundane tests to make sure manufacturers’ claims about the capabilities of their equipment were correct. I could spin up some 352.kHz or 384kHz, 24-bit PCM, or 22.5MHz 1-bit DSD from the server, send it to the DAC, make sure that reasonable sound came out, and tick that box. Many test resolutions – and as a bonus, the music and recording quality is excellent – are available for free here.
(Why not 32-bit? Well, for one thing, I don’t think there is anything recorded with 32 bits of resolution. And it seems particularly useless. The noise floor of 24-bit audio is -144dB. There’s not much point in extending that to the -192dB available from 32 bits.)
But I really wanted some PCM audio with 768kHz sampling so that I could test the new breed of DAC that supports the kind-of insanely high 768kHz sampling. (Do I think that 768kHz is necessary or useful? No. But if a claim is made, I try as best I can to check it.)
With my assiduous searching across all the Internet I could find but one 768kHz track. I think. A thread on www.head-fi.org from 2017 was from someone seeking 768kHz-sampled files. Of course, the early responses were along the lines of, “why bother?” Hey, the facility is there. Of course you want to check it out!
A couple of years later someone responded with two links. I thought I had my answer. The second link took me to a Japanese web page which Google thought was already in English. But an override revealed the following titles: “The world's first?! Beagle Kick starts delivering high-resolution sound sources with 768kHz/32bit integers”. But following the links (I’m not putting them here – if you want to explore, use the link near the start of this paragraph … with care) brought me to a purchase page which always responded with “Error occurred” when I tried buying the track, even when I used a VPN to place myself virtually in Japan.
The first link took me to one of those hosting sites where you can download stuff, sometimes slowly for free, but for a price at regular speeds. The site soon informed me that the track wasn’t available for slow download. I have a burner credit card for this kind of thing, so I paid around $50 for a month’s access and downloaded the gigabyte+ file.
It turned out to be some random video stuff, jammed into a zip file to consume a plausible amount of space. And even the video didn’t work. Burned. Me.
It was the towards the end of writing an email to the boss here at Addicted To Audio that a possible solution came to me. I’d been about to suggest that since he dealt with Topping and iFi Audio, both of which offer products which are said to be capable of decoding 768kHz PCM, he might perhaps ask them for a test track.
The possible solution was: could I make one myself? I make no claims to be an expert in high-end recording technology. Nonetheless, I have an RME ADI-2 PRO FS R Black Edition analogue to digital converter, DAC and headphone amplifier. Most of my software tops out at support for 192kHz PCM. But I do have relatively low cost ($US60 if you’re not big), but excellent, audio production software Reaper, a licence for which I bought some years ago. I gave it a whirl.
I used the interface for the ASIO driver for the RME ADC to set the sampling frequency to 768kHz. Reaper showed it was working at this rate. I fed the output of a Blue Ember condenser microphone to a small Yamaha mixer – the required phantom power on – and thence to the RME unit. Then I recorded me chatting about this and that, primarily the recording itself, for 54 seconds.
The resulting WAV file is indeed encoded in 24 bits at 768kHz. A CD-standard WAV file runs at a bitrate of 1,411.2kbps (where “k” is 1,000, not 1,024). My new test track runs at a bitrate of 36,864kbps. I used JRiver Media Center to send it to the Topping E30 DAC, and it duly displayed “768.0 PCM” on its front panel display (see the photo up top). That worked using the ASIO driver, the WASAPI driver and even Windows Kernel Streaming. It would not work using Direct Sound.
It also played fine with JRiver on a Mac using Core Audio.
But don’t expect 768kHz tracks to proliferate. This 54 second track is 240MB in size, almost a quarter of a gigabyte. My usual FLAC conversion software won’t work on it, but it’d probably still be around 100MB in size with lossless compression.
And, really, that’s a complete waste of space.
And all that said, do you have a DAC you’d like to test? Well, here’s the 768kHz test track which you’re free to download for testing purposes (believe me, entertainment value is nil). Remember, this is a 240MB download, so it’ll take a while. And, for completeness, here’s another, shorter, one recorded with 705.6kHz sampling. That’s a whole multiple of 44.1kHz.
* Tech note: the difference between 16 and 14 bits is manifested in one specification only: noise level. Theoretical noise level with 16-bit audio is -96dB. With 14 bits, it’s -84dB. That -84dB was so much better than all then-existing analogue systems, that Philips must have figured was that it was worth the trade-off. My memory from the time was that the hifi press tended to prefer the first Philips player to the first Sony player.