Nuendo Sample Rates over 300kHz: what are they used for?

Nuendo shows some insanely high sample rates… what are they used for in commercial productions?

I know that in the music world, the highest we normally go is 192 KHz, but there are options over 300KHz…
What kind of project would require or benefit from these?

-176.400 Hz (44.1 kHz times 4)
-192.000 Hz (48 kHz times 4)
-352.800 Hz (44.1 kHz times 8)
-384.000 Hz (48 kHz times 8)

I know that the format Digital Extreme Definition works at 352.8 kHz, but I couldnt find any recent references for it being used in commercial products.
I mean… does anybody nowadays launch products in SACDs? I don´t think I´ve seen an SACD player in years.

I’m just wondering if the audio quality improvement is even perceptible at these crazy high rates…

This is always a topic of much discussion. Also on these forums. For example here in the Nuendo 7 Sticky Post:
https://www.steinberg.net/forums/viewtopic.php?f=240&t=88517

My personal opinion - from all I’ve read and researched - is that the frequencies you want to capture are mathematically dependent on the sample rate. Meaning the sample rate has to be twice the value of the highest frequency you want to capture, otherwise those frequencies cannot be recorded. Example: Humans can hear up to 20kHz in their best teenage years, so you need 40kHz sample rate to capture that full spectrum - and play back all frequencies from 0-20kHz without any information loss.

This has been mathematically proven. There are people though who say sample rates above 40kHz increase audio quality. And that also ultrasonic frequencies matter and add subconsciously to the “feeling”. This might be true. But if those ultrasonic frequencies matter, you need the gear to record those super high frequencies above 20kHz. And the gear to play them back, too. And here, if you look at the spec of most mics, even expensive Neumann U87 mics, they record frequencies from 20Hz - 20kHz. They are not spec’d to record past our hearing range. And in my opinion, you don’t need a higher sample rate to capture the output of those mics. Also, many studio monitors and hi-fi speakers cannot reproduce it. Genelec’s 8050B go up to 20kHz. And their big mastering speakers, too. I would assume they know why.

There are microphones, though, that can record up to 100kHz. Those are useful for all kinds of situations. Scientific for example. Recording bats, pitching them down so we humans can hear and study their vocalizations, and those of other animals in the ultrasonic realm. Or sound design, record all kinds of weird frequency content, pitch it way down, and make strange audio experiences audible! This is how Godzilla in the 2014 Godzilla movie was done. Look at the first few minutes: http://soundworkscollection.com/videos/the-sound-of-godzilla

So for me, the sample rate goes together with the signal you record and the gear you have. If you have a mic that records up to 100kHz, you need a sample rate of 192kHz (or 200kHz to be correct) to record it, and later pitch it down for crazy sound design or scientific uses. If you have a crazy mic that records up to 200kHz, then 384kHz sample rate is needed. Otherwise, if you capture “normal” stuff, 48kHz are fine.

I actually think it’s a bit more to it than that, theoretically. It’s not just about capturing high frequencies but about how we encode signals in digital and what tradeoffs we end up making. I’m too tired to go finding the information now but I think that most modern converters operate at a high rate and then down-sample to whatever we’re actually setting manually. In other words they’re all operating in the MHz range or whatever and then it gets converted down to 44.1 kHz or whatever we want. I believe part of the decision to use SACD was to preserve more of the originally converted digital stream, as well as pushing conversion noise into a different range.

And on that note I think the actual definition of needed sample rate is that sample rate X can capture any frequency lower than X/2. So a 10kHz sample rate can accurately describe a signal that is lower than 5kHz, but not 5kHz. It’s of little consequence of course, and I could remember either of these things incorrectly…

Anyway, best recording I ever heard was on SACD. It was a while ago though so who knows if I still would feel that way.

Thanks guys! I´m no scientist or physics expert, but what you two are saying makes much sense:

  1. Frequency response of your gear, most notably mics and monitors determines the validity of setting up Nuendo to work on higher sample rates
  2. AD/DA performance must also make a difference

So, leaving aside the recording of bats or other scientific stuff… How do these sample rates get used for commercial applications, like music or movies? Or do they at all?

I actually think there’s a fair amount of gear that relays high frequencies, so you surely can record it and “use” it in the mix. Whether or not anyone can hear or feel a difference I have no idea. The argument that processing might benefit from it isn’t entirely illogical, but I think there’s been very little good practical research on just how many people actually can perceive a difference. And that’s all for music.

For post I have come across high frequency content once, last year actually, and it was for what would be a “radio soap opera”!!! I immediately told the producer that it was a complete waste of resources and they subsequently recorded the actors at the normal 48kHz, which is what I see all the time without exception in post (i.e. TV & film). Only reason high rates might be useful in post is if there’s sound design that benefits from pitching down for example. But for dialog and everything else it’s just more work for the computer and more storage needed… pretty much pointless.

MattiasNYC,

I agree with you. I´ve read the article ChrisPolus kindly shared, and it boils down to what you´re saying.

There is scientific evidence that the highest you need to go in order to “get more quality” is between 50 and 60kHz. Above from that, the amount of data may even tend to create imperfections (depending on the specs and build quality of your ADDA convertors), which can be identified as the “hey! that sounds better” effect. But all this happens within the subjective realm of the listening experience.

If I understand correctly, there’s also a benefit to latency times using higher sample rates? The higher the rate, the lower the latency?

Thanks!

Darren “Latency” Ingram

Of course. With higher sample rates, more samples pass your DSP plugins in a given time. That means i.e. a brickwall limiter has enough information in shorter time to “know” how to process your audio. The same is true for any plugin that uses upsampling (which is probably any plugin these days?). But don’t mix latency and performance! Higher sample rates usually require higher processing power, so you might need to increase audio buffers which in turn leads to higher latencies.