96 khz and normal used 44.1 khz...

[Scottxx]

Yesterday I was using some amp sims and configured the project in reaper to 96 Khz and 24 bits. The sound was so clear and crispy! I can't understand guys why 44.1 khz is the standard that most of us use. I know 96 khz use more than double resources than 44.1 but one thing I can't understand is that too much people says that the difference is subtle. If you record a voice for example, using 44.1 the dsp is taking 44.100 samples per second (sqare shaped wave) and using 96 khz takes the source to a much "rounded" and similar to the original incoming signal. So if it is so clear to the ears of any of us, why don't we use it ? I know that then if you must to export it to a cd you must downsample it. But working all the time in a greater resolution and just at the final stage downsample it to the needed sample rate.

 

[Zozobra]

First a 44.1 sampled wave is not square. Each sample point is convolved with a sync function to restore the waves shape. See nyquist.

However, its plausible (but I dont really know one way or another; I think I've heard the difference in some things) that processing and effects are affected (for the better) by the extra resolution/availability of usefull audio information, and I've heard somewhere that near-inaudiable HF information can affect how we hear sound well below that range. Dunno about that personally, but I have heard quite distinct differences between 96 and 44.1. Whether there were other explanations for the differences I heard I'm not sure.

Long story short: its not as simple as youre saying it

Default fallback answer: Nyquist > .

 

[Scottxx]

Zozobra, I'm not and audio professional. I know 44.1 khz is not a sqare wave, I want to say that when you digitalized a sound wave it gets in 1 and 0 and you have a more "sqared" sine wave because it is just zeros and ones. So if you get more samples like 96 khz, you obtain a more rounded and similar to the original one. That is why it sounds so much better and clear. So I want to know what all the experienced people out there, if you are working at 96 khz or 88.2 and what do you think about all this.

THank you all out there!

 

[syns]

I think you should record the same thing at 96khz and 44.1khz, mix it abit, and mixdown the 96khz to 44.1 and compare the result, I want to hear it too!

 

[SocialNumb]

CD/MP3

Fuck 96

 

[deLuther]

Is impulse used at 96k recorded at 96k or it`s just resampled from usual 44.1k? There is some difference due to impulse resampling.
Additionally most samples recorded also at 44.1k... and final format (CD).

 

[Scottxx]

For example, Pod Farm 2. If you get that cristal sound, using 96 khz, in each source (voice, guitar, bass, etc...) and if then you need to downsample it to a cd I think It will have better sound than getting each source at 44.1 khz.

 

[schuldiner]

There is another thing....when you sample an analogic signal the maximum frequency that you can record is equal to (sample frequency)/2.
Well...when you have a sample rate at 44.1 Khz you can record all frequecy from 0 to about 22Khz, at 96Khz you can record a 48Khz sound.
Then:

1. you can't hear a sound with this higher frequency
2. I think that a normal speaker it can't reproduce a sound at 48Khz and in any case you can't hear it.
3. If you use a too higher sample rate you have the risk to record some unexpected sound.

Istead is another thing when you talking about quantization levels (16bit or 24bit)

At 24bit you have more level for quantize the signal, and then you have more resolution for rebuild the original waveform

But when you step up your sample frequency it means that you record a higher frequency sound, anything else.

Sorry for bad english but I hope that you understand what I mean.

I post this link that can explain better :

http://en.wikipedia.org/wiki/Sampling_rate

 

[TheDarkProject]

Zozobra, I'm not and audio professional. I know 44.1 khz is not a sqare wave, I want to say that when you digitalized a sound wave it gets in 1 and 0 and you have a more "sqared" sine wave because it is just zeros and ones.

No, you get a number between zero and 16,277,216, if you're recording at 24 bits per sample.

So if you get more samples like 96 khz, you obtain a more rounded and similar to the original one.

It's not so much that it's more rounded, it's that it's more straight! If you drop the precision down then the values stored are worse approximations of the actual measurement which means the wave form acquires extra jaggedness, and that manifests itself as aliasing and distortion.

 

[TheDarkProject]

Istead is another thing when you talking about quantization levels (16bit or 24bit)

Recording at lower frequencies gives you more temporal quantization, whereas lower bitrates give you more amplitude quantization. Both can matter.

 

[schuldiner]

yes I you have 24bit of quantization level it means that you have 2^24 level for rapresent the sample in a sequence of bit, it means that you have more precision in the amplitude of the signal

 

[Trevoire520]

why don't we use it ?

Basically it comes down to computer processing power, hard drive space and the fact that the end product is going to be a 16 bit, 44.1k cd anyway. Which is then going to be ripped to a mp3 file.
You're recording and listening to frequency content that is not going to be present in the final product, and not going to be heard by the end user. So whats the point?

I'd be quite interested to listen to a comparison between something that was mixed at 192k and the same thing mixed at 44.1k. But listening to the final product at 16 bit 44.1k to see if there's a difference between the two.

 

[BLUElightCory]

High sample rates:

PROS -
- Higher resolution recording (duh) - 88.2/96 sounds noticeably better to my ears right out of the gate
- Improved converter performance - ADCs can use a gentler high-pass slope
- Better plug-in performance - less aliasing

CONS -
- Use more processing resources
- Use more disk space
- CDs end up at 44.1kHz anyway so there's always a question as to whether it's worth it or not

I'm a big believer in tracking at the highest sample rate you can reliably get away with.

 

[drope934]

So... technically... if we release songs only digitally... it could be accurate (if the CPU handle it) to work in 96khz?

 

[Scottxx]

I agree 100% with BLUElightCory. And you are right Trevoire520. I know that downsampling to cd formta or mp3 obviously you lose what you got with 96 khz. But the point is what you get before doing downsampling. The results to my ears are exceptional with 96 khz. Is an easy test just to do this.. use any amp sim plugin with the project at 44.1 khz and 24 bits. ANd play with a preset. Then just change the project to 96 khz and 24 bits, play a little with the same preset and hear that awsome sound!

 

[Max Morton]

In fact, I've noticed that some plugins do behave differently in 88 and 96, and in many cases they just work wrong. So it does not always mean that the sound is drastically different because of the samplerate.
When having some 70 tracks of drums, guitars, di's, backing vocals, keyboards and vsti's, it's getting a bit hard to work with a project if it's over 44k.
Time to buy the new computer

 

[Max Morton]

I agree 100% with BLUElightCory. And you are right Trevoire520. I know that downsampling to cd formta or mp3 obviously you lose what you got with 96 khz. But the point is what you get before doing downsampling. The results to my ears are exceptional with 96 khz. Is an easy test just to do this.. use any amp sim plugin with the project at 44.1 khz and 24 bits. ANd play with a preset. Then just change the project to 96 khz and 24 bits, play a little with the same preset and hear that awsome sound!

Don't forget that you're also getting a smaller latency (more then twice shorter) with 96 kHz. You're also listening through the converters operating in different mode. So it would be more correct to play and record your playing in 96. Bounce it to a wav. Then play and record in 44. Bounce it to a wav. Downsample the first wav. And compare it blindly.

 

[Fox Mulder]

Bormoleos Van Drill : I'm a Judas Priest freak and I love your signature

 

[Strachan of D]

well techically it should have less noise when down sampling beacause the noise will be shaped overr to the higher frequecys above the nyquist point that will then be cut out by the smoothing(econstruction) filter



im thinking of uploding my digital audio theory note from college this year when i get time.

 

[Max Morton]

Maamar Huq:
Yeah, they are the Gods. Saxon, too.

Getting back to 96/44 question. Let's keep in mind, that then it usually becomes mp3 that is approximated as hell. Yes, it's the right thing: if the sound benefits on the mixing stage, we must try to keep as much of it as possible even if it degrades later. But I did my tests long ago and I always got back to 44.