So I have upgraded my audio interface and can now (in theory) record at 24/32bit up to 192khz -
There is an interesting read here on sample rates Sampling (signal processing) - Wikipedia and some interesting points in it seem to say that the human ear can’t determine a difference above 50khz - but that it higher rates do seem to have some kind of ‘impact’ -
So what’s your preference? do you record above 48khz? do you think it makes a difference once mixed down?
I reproduced a small recording I had done at 24bit 44.1khz with 24bit 192khz today and I seem to be able to notice a ‘richer sound’ - but is that just placebo? I’m not sure -
I’d welcome all of your views - thanks v much - JS
The human ear can’t really hear the difference between a 48Kh recording or a 88Kh.
If you don’t work with musicians or if you plan to do a lot of time/pitch process in your audios, the more sample rate you got, the less artifacts you will find.
Humans hearing 50Khz?? I think that when you are young you could hear up to 20 or 22 Khz, when you’re getting older it wil probably drop to around 16 or 18 Khz depending on how much volume voilence you had to bear your whole life.
What I heard from many People is that the frequencies above 20 Khz can influence the sound perception, more like I’m not hearing it but feel it.
The biggest problem with the theory that the frequencies you cannot hear affect the ones you can (anything above 20KHz). is that hardly any microphones or monitor speakers reproduce these frequencies anyway, so they are never there to perceive - whatever sampling frequency you use.
The only “sound difference” in high sampling rates is usually that of the D/A converter (quality of the aliasing filters), and that is in the end users CD player or computer (read - not very good).
How ever it is easier to design a cheap (poor) D/A running at 96KHz than one at 44.1KHz, as at 96KHz, you can heave really rubbish aliasing filters that do their job without affecting the audio.
So maybe the sampling rate is best higher for all the wrong reasons.
Because of where I live I often get clients using solo instruments.
Slack key guitars/bamboo flutes/ukes/a single drum for chants/solo vox(s) etc etc
In these situations when using higher rates I can actually ‘perceive’ a diff.
And often some of these plays are ‘purist’; in the sense that they want NO fx at all.
No comp no eq and no reverb/delay.
However when a ‘typical’ band comes in, it is 24/44.1 all the time.
I just can’t hear/percieve a diff at higher rates in those scenarios.
Nobody can. It is physically impossible to hear anything over 20kHz.
The human hearing has no mechanics, as in no haircells in our cochlea to convert soundwaves in the frequencies above 20kHz (after a few years only 16-18 kHz) into electric signals (this is the best A to D converter ever excisted - our hearing system).
We don’t actually “hear”, as the brain receives electric signals. Then we are (our brain is) interperating what’s coming in. Ask any hearing doctor.
So 24/44.1 should be more than fine. If not… if not your converters have bad filtering (Read andyjh’s post again about filtering in 44.1 vs 96Khz - he is spot on).
If you hear a difference on 96k against 44.1k, it is not better in 96kHz. It is way worse in 44.1, due to bad filtering artefacts entering our hearing frequency range.
Then watch this video:
After that we can discuss.
I have read thousands of pages on this topic, attended seminars with the head master of TC. Electronics’ digital research and developement, Thomas Lund. And I have read my share on the human hearing system.
The most important I have learned, is that we can’t fight nature, pure math and physics. It is just physically impossible to hear anything over 20kHz.
To the other guys who took the time to reply - thanks. I’m definitely talking about sample rate rather than frequency - I know Humans can’t hear a frequency of 192khz ;o) -
No. Not possible, within a musical context.
Think of how much energy it would take to make that “tickle” your skin… What happens when you feed (play back) a full range frequency signal with that amout of power. Your head would explode, literary.
There are no amplifiers able to feed that much power, to any speaker system, and to make anyone feel those upper frequencies.
We are obviously talking within a music production/playback context, and both speakers and eardrums, and probably our skull would blow to pieces.
If we are talking Ultra-sound and Roentgen in medical environment/tests (we MAY feel it), you have to have a High-pass filter in those medical equipment set very high (otherwise you would burn to death), to not break down the power supply for your whole city.
Said in another way. When you are at the physician taking Ultrasound treatment, you may feel a bit hot on your skin/muscle. If it were possible to remove the highpass filter in that equipment with a flick of a switch, you would blow up the building.
So in a fullrange frequency musical piece, to “feel” those upper frequencies, our heads had exploded a long long time before we could feel some content at 40kHz.
This is actually basic math and physics, and the human hearings limitations.
PS. The hard part in explaining math and physics related to music production, useful frequency range and the “law of the nature”, is to do it as easy as possible. It’s really hard to do it really easy
I’ve done tests with vsti’s and recorded them at 96k and 48khz. At 96(my card won’t go any higher) you get a nice shimmering high end that is not so present on the 48k but the thing is it’s not really a big deal unless you are doing blu ray stuff. If you producing music that’s inevitably going to be 44100 in the end then it won’t make a difference. I recorded a track at 96 and then bounced it down to 441. Then I compared the two and I still lost the high end shimmer that the original 96k file had. If I remember correctly nothing really was lost on the low end. So all in all you can eq to get some more high end if you need it. Sometimes it can be to much and you need to shave it off the 96k stuff.
As for 32 bit,well you get way more headroom to play with and a bunch of other crap that I don’t care about.
If you ever hear the difference between 96K vs 44.1K, it is due to bad filter design of your converters (Bad filtering artefacts is creeping down in the hearing range < 20 Khz). The cheaper the converter/interface the more likely it will sound WORSE at 44.1K, NOT better at 96K (due to bad filtering.)
In better/very good converters human hearing will not hear or feel the difference.
It’s impossible to hear anything over 20kHz.
The mechanics of the human hearing (auditory) system has no way to mechanically “convert” soundwaves into electric signals (sent through nerves to our brain for interpretation).
That is the the auditory hair cells job. They are located within the organ of Corti on a thin basilar membrane in the cochlea of the inner ear.
There are simply no hair-cells to capture anything over 20kHz (when newly born), ie no mechanical moving parts up for the task.
The frequencies above 20kHz can therefore NOT be converted to electric signals, and therefore will NOT reach our brain, because there is nothing for the nerves to send to our brain.
If you hear a difference between 44.1kHz and 96kHz, it’s either bad converter filter design or imagination.
Somebody saying otherwise has NOT studied the Human Auditory System.
Just to put this in context: most gigging musicians I know have severely damaged hearing, and I suspect most of them can’t hear a thing above 10-12kHz, certainly those who are now in their 40s and older, who’ve been gigging for years beside a drummer without wearing ear protection. This fact doesn’t stop them from making great music that sounds fantastic. For example, I find that most great electric guitar sounds can endure a steep cut above 5kHz without any loss in perceptible quality (and with younger guitarists: with a noticeable improvement!).
Just saying … if the music doesn’t sound so good, perhaps the first place to start looking might not be in the area of Nyquist’s theorem, sampling rate, or D/A converters.
Anyway, back on topic: I use 44.1kHz/24-bit as a default, or 48kHz/24-bit if I know it’s going to video.
The main thing to consider here is that recording at higher bitrates and playback are two very different things. If you’re recording a 100 piece orchestra you’re better off to record it at the highest sample rate / bitrate as possible because summing up 100 channels of different data (the waves adding together to a stereo track) you get a higher resolution and more accurate rendering.
So when you render those 100 channels down to CD quality (44.1/16bit) you will have a more accurate reproduction.
Playing back at 192KHz/32bit is absolutely pointless and a waste of disk space but recording is important to capture as much data as possible.
The finished result will probably be an MP3 or CD anyway.
Not so at all when mixing and doing real time processing. All internal mixing/processing is done in the 32 bit floating point realm anyway (1500 dB headroom).
But… When doing a lot of OFFLINE processing (mathimatically altering the source file), it is an advantage with 32 bit fp and 24 bit vs 16 bit.
If you just mix perfectly good source files without a lot of offline processing, 24/44.1 is more than good enough.
As a sidenote: Bon Jovi recorded their last record in 24/44.1. Why, when they can buy and use whatever equipment they want? Beacause their highly “educated” and trained engineers know that our ears don’t need anything more (can’t hear the difference).
And of corse. They have greatly recorded source files to work with.
So don’t waste your time on technically non-issues
