What is the difference between 44.1khz and 96khz

This article will cover the basics and best practices for setting sample rates. Sample rate tells us how many times per second we take a measurement of an analog audio waveform as it is converted to a digital signal. Since sample rate has a speed, or frequency, the sample rate defines the frequency response of an audio recording. Specifically, the Nyquist Theorem states that the highest frequency we can record is half of the sampling rate.

This means a sample rate of Accordingly, a 96 kHz sample rate allows for 48 kHz of audio bandwidth. If we attempt to record above half the sample rate, or the Nyquist limit, audible artifacts called aliases occur. Analog to digital converters eliminate aliasing by low pass filtering the analog signal at half the sample rate.

This low pass filter is referred to as an anti-aliasing filter. In practice, the low pass filter requires a range to operate, so we state 20 kHz as the practical upper limit for We know that human hearing covers from about 20Hz to 20 kHz, so why would we need sampling rates above One answer is that many people, including scientists, claim that humans can perceive sounds as high as 50 kHz through bone conduction.

That claim may theoretically be correct, but through air humans only hear up to about 20 kHz. The second reason is a more practical one. The low pass anti-aliasing filter is not a perfect filter, so it creates some of its own distortions.

There is a design trade-off between how steep a filter can be vs. We just learned that sample rates above In other words, For this and other reasons, it is recommended that we produce and mix pop music at 48 kHz. First, 48 kHz allows for better sounding anti-aliasing filters than Is there any advantage to recording at 48 kHz?

Recording at a higher sample rate offers a combination of pros and cons, depending on the output format. It is important to note that humans cannot hear the difference between While some people claim that they notice a slight improvement in audio quality when selecting the higher audio rate, research indicates that 20 kHz is the limit to human hearing.

This theorem stated that the sample rate needs to be double the highest frequency that you want to reproduce. To ensure that the audio CD covered the entire frequency spectrum that is audible to humans, engineers used the The main benefit of sticking with a The file sizes also tend to be smaller, which may be a factor when sharing audio files with collaborators over the Internet or saving space on your hard drive.

As humans cannot hear the difference between You should also consider the format that you plan to use when delivering your final mix. If you record at a higher sample rate, the sample rate needs to be converted to Older conversion software would produce distortion or a loss of quality when converting integers. These same issues do not apply when converting However, when converting from 48 kHz to Basically, if you are planning to burn your music to CD, While you cannot hear the difference between Using a 48 kHz sample rate offers slightly more headroom for tweaking your mix.

If you decide to go back and edit the master, the 48 kHz sample rate offers more flexibility, especially when working with high frequencies such as the sounds produced by cymbals and high hats. A higher sample rate also reduces the risk of aliasing. Aliasing occurs when the different frequencies become less distinguishable due to artifacts not getting filtered out.

You also get lower latency rates with higher sample rates. Basically, a higher sample rate helps to produce a cleaner sound. However, the difference will not be noticeable in the final output. You may also use a higher sample rate for burning audio to CDs without using 48 kHz. My opinion is that bit is nowhere near "plenty". I've been involved with developing musical-instrument-related digital audio since the early 80s, bit was barely enough for the optimized case where you've fit the audio to use all 12 bits and added compression to keep the level up.

Particularly in exponential decays of note and drum tails. Usable yes, plenty no, clearly a compromise. In real music you have the added issue that you may well want the bulk of the music at dB from full scale. A perceived doubling of loudness is 10 dB aka 1 bel , that's not at all out of the question. Sorry to ask, but I was very interested in the original discussion. Posting my 2c because i am bored and ive been scratching my head about this I think most of us can agree to that.

Any identical 96khz session is going to sound immediately better than a 48khz session. But when it comes to comparing mixed down and bounced material that are sampled down to And i am having large amounts of trouble trying to pin point exactly what the difference is. All i know is, i have a song processed at 96khz and 48khz, identical processing both are bounced to Its driving me insane but ive basically settled with mixing in 96khz and just completely steering clear of before i drive myself into a psych ward.

I used to be like that back when i was getting into listening to music and i was a newb and for example, i liked songs that had a kbps bit rate more than i liked songs that had kbps bit rate, but id never really know why Personally, this went as far as avoiding anything low quality back in my younger days like radio rips, leaks, etc and always running for higher quality when i finally figured out the difference simply because i knew the technical quality of the song may sway my opinion.

Reason: Added last part. Bob Olhsson. It is most noticeable in the rhythm. My guess is that it may be due to less distortion in dynamics processing. Mix at 48 kHz and upsample the mix to 96khz with ant aliasing filter Here's a little Christmas present that might help people make up their mind about what difference there really is between sampling rates.

Is the difference to be attributed to the design of the digital filter which can be made arbitrarily good at the expense of more latency or is it to be attributed to the inherent smaller sample interval, characteristic of higher sampling rates?

What we have here are 3 brickwall linear phase filters for each sample rate 48 and 96 kHz. The main difference being rejection, respectively at , and dB. These are the equivalent of the brickwall filter one would find when using a DAC at Wav files are bit integer. There are also txt files with raw double precision FP coefficients. Full rejection is reached at Latency is at most 1.

What these filters simulate is the change in signal that happens when using a DAC at Or rather, at least, since they operate at the same sampling rate of the original files one wants to test 48 or 96 kHz , the issue of inherent smaller sampling interval is taken out of the equation. Only thing needed to test is a convolution plug-in at the end of the chain. An interesting test to do, for example, is to put just one filter at the end of the chain, or many.

The first filter will mainly take out the ultrasonic content. That should allow to test for audibility of amp and speaker extra IMD caused by ultrasonic content. The following filters should give one an idea if there really is something inherently 'evil' about brickwalling a signal. If there was, every extra filter added should audibly alter the signal. At the very least, these can be used as anti-ultrasonic filters in the chain, to be used after non-linear plug-ins to keep IMD at bay if latency is not an concern.

These filters were very kindly provided by the same person that makes FinalCD. He suggested a donation to your local food bank if you happen to download them and find them useful come on people, it's Christmas!