frequencies problem ?

[karxix]

hmm
i've recorded a full song with cubase sx and emu 1212m and mixed it down
am just wondering where the high frequncies are gone
when i play in winamp a song usually i have the last 3-4 level meters jumping up and down
while our song just doesn't have those high frequencies, i even tried boosting them it didn't work
i would expect those frequencies to be in the recording in the 1st place but it's not there, what should be done ?

 

[Blue Bear Sound]

You must have mixed down to a lo-resolution format.......

 

[SEDstar]

tis mught be a nyquist thing....

this nyquist guy discovered that at x sampling rate, you can (at best) only capture x/2 frequency...

in laymans terms, 44,100 theoretical sampling rate means you can at best only represent 22,050hz signal...

at firast i was recording tracks in 22000 or so, then wondered why my spectral analysis display showed nothing above 10k.....

whats the mic capable of? if it doesnt have response at the freq you want...

and while you can boost or cut eq at a certain frequency, it doesnt work miracles...the frequencies have to exist before you can work with them...you cant create frequencies with an eq

 

[boingoman]

I think he's talking about level, not frequency response. He said his meters weren't peaking, and seems to think this means he's not getting hf response.

It sounds like you need to get more level in your final mix, not high frequencies. Boosting the highs will actually make it harder to get a louder overall level, as it will eat up headroom. Try a little stereo compression to tame the peaks in your songs. This will let you have a higher average level on your final mix.

 

[boingoman]

this nyquist guy discovered that at x sampling rate, you can (at best) only capture x/2 frequency...

in laymans terms, 44,100 theoretical sampling rate means you can at best only represent 22,050hz signal...

at firast i was recording tracks in 22000 or so, then wondered why my spectral analysis display showed nothing above 10k.....

whats the mic capable of? if it doesnt have response at the freq you want...

and while you can boost or cut eq at a certain frequency, it doesnt work miracles...the frequencies have to exist before you can work with them...you cant create frequencies with an eq

22050= 0-20k + 2k needed to roll off the signal + 50hz video lock signal.

We use 44.1 not so we can sample to 22050, but so we can cleanly sample to 20k and also lock digital sound to video.

 

[cornjchob2]

22050= 0-20k + 2k needed to roll off the signal + 50hz video lock signal.

We use 44.1 not so we can sample to 22050, but so we can cleanly sample to 20k and also lock digital sound to video.

50hz? If NTSC is 29.93fps, then how could the lock be 50hz? Are you talking about PAL? That's 25fps, if I remember correctly, because it's used on tvs getting AC at 50hz instead of the 60hz we Americans are used to. You've piqued my curiosity, hope to hear you elaborate.

 

[boingoman]

I guess it's different in digital. It's not a 50hz signal, it's 22050hz. And it also seems to be different than SMPTE time code. I found this while in another thread, we were talking about sample rates and someone asked why we use 44.1khz as a sample rate instead of 40khz, which is required to sample 20khz following the Nyquist rule. It's cool info.

Edit: I looked around some more. Many digital master clocks will sync to NTSC or PAL, generating their own 50 or 60 hz signals. Or I should say, they will sync NTSC or PAL material to their clock, and the digital audio as well, using that 22050hz video lock signal. So it's really the master clock that does all the work. I googled "44.1 video lock". I also found a great article by Roger Nichols (digital guru) about being able to fix a master perfectly that had been downsampled because the multitrack at 48khz and the mixdown DAT at 44.1 had both been referenced to video sync at mixdown. Roger Nichols
This seems to imply that a video sync tone is not something standard on a CD. Hmmmm.... need to do more research.

Back to the main story!!!!!!! Here is what I found:


You can capture a 20 kHz simply by sampling at 40 kHz to satisfy Nyquist, plus 10% more for the guard band, plus 100 Hz to lock to video. 40 + 4(10%) + 100 Hz = 44.1.

"Now we have to build these anti-aliasing filters [low pass filter] to cleanly pass 20 kHz, but be out (-90 dB) by 22 kHz".

" Truth is you can't dump that much level in that little frequency band without huge phase problems in the analog or digital domain. Therefore phase shift and high frequency ringing are common. 48K is smoother than 44k because of the extra headroom (10%). The problem with 48 k is it uses more media and is another standard"

Stephen St. Croix, "Deep Down Digital - Lessons from the anesthesia front," Mix, vol. 10, no. 10, October, 1996

http://csunix1.lvc.edu/~snyder/2ch11.html

It doesn't seem as though it is a fixed frequency for syncing, though I need to do a bit more research. In 44.1, the sync frequency was kept out of the audio range, as opposed to being a seperate track. I'm assuming in a 48khz sampling situation, you don't get the full 23k range (48k-2k for the filter=46k/2=23k) but instead get an effective upper frequency limit of 22950hz after you subtract the video lock frequency.

I'm really glad you said something. I like this stuff. You, sir, have piqued my curiosity.....