Operating in 192k

[3dot...]

http://www.adam-audio.com/en/pro-audio/ ... nical-data
http://www.adam-audio.com/en/technology/x-art
http://www.enjoythemusic.com/magazine/e ... n_tlr1.htm
(scroll down for freq response)

Super tweeter
From Wikipedia, the free encyclopedia

A super tweeter is a speaker driver intended to produce ultra high frequencies in a multi-driver loudspeaker system. Its purpose is to recreate a more realistic sound field, often characterized as "airy-ness". Super tweeters are sometimes found in high fidelity speaker systems and sometimes even in home theater systems. They are used to supplement the sound of tweeters by reproducing frequencies which the tweeter may produce only with a narrow polar output, or perhaps with distortion.

An early example of a super tweeter used in a commercial design is the microphone element derived 4001 from STC, used in the Bowers & Wilkins DM3 and several subsequent models. These were basically two-way designs using the STC 4001 as a super tweeter. Ohm, the Brooklyn-based speaker manufacturer, uses super tweeters in their Walsh and MicroWalsh speaker lines, which several reviewers have concluded offer a more pristine high-end response.[citation needed]

A super tweeter is generally intended to respond well into ultrasonic frequencies over 20 kHz, the commonly accepted upper frequency limit of human hearing. One super tweeter has been designed to respond up to 100 kHz.[1] This super tweeter was designed to eliminate the problem of the normal tweeter diaphragm continuing to move after a signal has ceased.

en.wikipedia.org/wiki/Tweeter

AMT tweeter

The Air Motion Transformer tweeter works by pushing air out perpendicularly from the pleated diaphragm. Its diaphragm is the folded pleats of film (typically PET film) around aluminium struts held in a strong magnetic field. In past decades, ESS of California produced a series of hybrid loudspeakers using such tweeters, along with conventional woofers, referring to them as Heil transducers after their inventor, Oskar Heil. They are capable of considerable output levels and are rather more sturdy than electrostatics or ribbons, but have similar low-mass moving elements.

Most of the current AMT drivers in use today are similar in efficiency and frequency response to the original Oskar Heil designs of the 1970s.

[3dot...]

I know RME has those soundcards, I've been through this discussion many times before, someone always brings that up. I don't think 3dot has that sound card.

nope.. mine goes up to 96KHz... (I use 44KHz btw)
http://www.rme-audio.de/en_products_multiface_2.php#5

as for stretching.. what I meant...
we convert archives of analog tapes where I work..
so we modded the playhead to run at the uncomprehensible x8 speed..
(so we can more tapes/time)
before we did that we did a number of tests with 'time-stretching' algos..interfaces... etc.
and we figured..
if we record at a higher samplerates...
then we can get better quality 'time-stretch' (on conversion to 44KHz)..
that turned out to be right...

[twisted-space]

doesn't matter, your sound card's output is band limited to 20kHz.

Specs from my RME Babyface.

http://www.rme-audio.de/en_products_babyface.php#5


Frequency response DA, -0.5 dB: 10 Hz - 22 kHz (sf 44.1 kHz)
Frequency response DA, -0.5 dB: 10 Hz – 45 kHz (sf 96 kHz)
Frequency response DA, -1 dB: 5 Hz - 80 kHz (sf 192 kHz)

Are you saying that these are incorrect?

those are correct. now go check your monitors. THE WHOLE SIGNAL CHAIN NEEDS TO PASS ULTRA HIGH FREQUENCIES.

I know RME has those soundcards, I've been through this discussion many times before, someone always brings that up. I don't think 3dot has that sound card.

I'm going to quote you again, just to be clear what it is I'm talking about

doesn't matter, your sound card's output is band limited to 20kHz.

You didn't mention his monitors frequency response there.

3dot uses a multiface:

Frequency Response
D/A: 5Hz to 20.9kHz, -0.5dB @ 44,1kHz
D/A: 5Hz to 35.0kHz, -0.5dB @ 96kHz

So saying his cards output is bandlimited to 20Khz is simply incorrect.

[Der_Makrophag]

My Genelecs has max 18kHz output. So is the only benefit of higher kHz that you just add frequencies abowe the range of what humans can hear, and most monitors can output?

NO, that would be stupid of course, as mentioned before. I find it rather shocking, that the main reason for higher samplerate is not mentioned here yet, and everyone is talking about bandlimited output. This is not the reason to switch to higher samplerates, your are right of course. The reason is aliasing. You better get some information about this, if you want to know, why some people work above the minimum of 44,1 kHz. Ok , nowadays, it becomes less important, as many plugins provide oversampling of their own, so I think 192kHz is too much. But I think 96kHz is quite ok.
Necessary? Don't know. I am working with 48khz right now and I do no treally have problems, but I will get a new interface soon and plan to work at 96kHz.

[ze2be]

Sometimes I dont bother to bring a laptop+soundcard, and I DJ my own tracks from CD (at clubs)
One of my studio mates told me the kick was sounding muddy. The track was pitched down on the CD deck -4 or -6. And he thought that it was the reason. This made me think. If CDs are the absolute minimum resolution for music, it might not be such a good idea to pitch it down.

Is this where the kHz gets important, and what about 16bit vs 24bit?

(I ask simple/stupid on purpose because it helps the comunity to get a clear answare. Wiki didnt help me much)

[Der_Makrophag]

OK, lets talk in absolute limits:
CDs have a samplerate of 44.1 kHz. That gives you a maximum of 22.05 kHz achievable frequency response (according to Nyquist). Pitching it down results in a lower frequency range, because you also pitch the maximum frequency down. The calculation here is more complex as pitch has a logharitmic frequency scale. If you pitch down by 12 semitones (an aoctave), the maximum frequency is halved, that is 11.025 kHz. That IS hearable and will sound awful. For pitches inbetween you can think of freuqencies inbetween.
But the kick can also sound muddy because you pitch the bass down, maybe out of the range of your playback system (thats from 40 to 20 Hz maybe).

[ze2be]

OK, lets talk in absolute limits:
CDs have a samplerate of 44.1 kHz. That gives you a maximum of 22.05 kHz achievable frequency response (according to Nyquist). Pitching it down results in a lower frequency range, because you also pitch the maximum frequency down. The calculation here is more complex as pitch has a logharitmic frequency scale. If you pitch down by 12 semitones (an aoctave), the maximum frequency is halved, that is 11.025 kHz. That IS hearable and will sound awful. For pitches inbetween you can think of freuqencies inbetween.
But the kick can also sound muddy because you pitch the bass down, maybe out of the range of your playback system (thats from 40 to 20 Hz maybe).

I see.. Thanks! So then, for dramatic shifts in bpm, its better to use beats mode of the warp modes in Live, instead of repitch? As long as the song is strongly rythmic.. (beats mode is similare to a rex file, in that it slices the song/clip on the transients, and stretch them out, keeping the original pitch.)

[Forge.]

wish I could afford a system that could make 192k actually make any difference whatsoever

I have ears no way near pristine enough to really appreciate any of it anyway. And the evidence seems to suggest that not really many other people at all do either.

The theory seems to weigh solely on the idea that what we can't /hear/ we might be able to /feel/.....

Don't worry about shit like that.

Work at 44.1/24bit if you are just making music, for yourself/labels etc.. 48k if you are working for video

beyond that, unless you have some amazing gig where you can go into the amazon and record a concert of weird creatures that make noises well above human hearing and you want to archive it more accurately, then it's not likely to concern you. Assuming you have a mic and AD converter that can reproduce it properly anyway.

I have admittedly recorded instruments like Sitar at 88.2 just because there are a lot of harmonics going on with a sitar and I wanted to catch those upper ones in case there are other people out there with better hearing who might appreciate it one day, but I'm not sure my mics or A/D converters were really up to it and I've never really noticed the difference. At least, not during my normal listening on CD/MP3/AIFF through iTunes or whatever... remember that's where it will end up in most cases.

Even on video stuff. More than likely it will be viewed on a Telly. And even with the new super-telly's and 5.1 systems, it's not really going to notice.

But I'm pretty certain my NS10s don't go above 20kHz anyway, and my headphones are probably about the same.

This is the territory of people with more dollars than sense.

[ze2be]

So, there really is not nessesary to go above 44.1kHz if the music isnt going to be pitced very much. How much of a problem is the aliasing, with or without pitching?

[nuxnamon]

i mainly work at 44.. but for vocal work that i record that gets mixed by another in the big studios, the engineers always ask for 96. maybe cuz the project or all the other stems were recorded at 96??. i honestly can't tell much difference even with an adequate monitoring system and good ad/da converters.. but from the big studios i've visited, they all record in 96.. haven't seen 192 in use yet..

[icedsushi]

but from the big studios i've visited, they all record in 96..

Wouldn't 88.2 make more sense for audio tracks if the end result is 44.1? Unless they're doing video, 48 is more common for video, and 96 is an even multiple of 48.

I'm thinking of experimenting with sample rate, perhaps 88.2 & not 96 because I don't do video. To produce a 22,050k waveform/harmonic at 44.1 it's basically going to look like a triangle wave after it's recorded/rendered because there's only 2 samples per cycle, is this correct? Technically at 88.2 you have 4 samples per cycle so that waveform still within the range of human hearing is going to look (sound) more accurate?

I agree that 192 seems like overkill, too much CPU use without any audible benefit from the other options.

On a side note, how to you get a mix rendered at 88.2 down to 44.1 as the last step? Dithering is for bit depth, but do you have to do any special process when downsampling or just hit render in Live & select the lower sample rate in the menu? I'd like to compare a mix downsampled from 88.2 to 44.1 & compare it to the same thing in 44.1 throughout the whole process.

[trevox]

OK, lets talk in absolute limits:
CDs have a samplerate of 44.1 kHz. That gives you a maximum of 22.05 kHz achievable frequency response (according to Nyquist). Pitching it down results in a lower frequency range, because you also pitch the maximum frequency down. The calculation here is more complex as pitch has a logharitmic frequency scale. If you pitch down by 12 semitones (an aoctave), the maximum frequency is halved, that is 11.025 kHz. That IS hearable and will sound awful. For pitches inbetween you can think of freuqencies inbetween.
But the kick can also sound muddy because you pitch the bass down, maybe out of the range of your playback system (thats from 40 to 20 Hz maybe).

Exactly correct. Granted if you do not pitch things down severely, operating at really high sample rates does not matter too much. However, that is not to say that sometimes it is not necessary. As an example, I remember once recording various elastic bands to make an instrument and figured it would best suit being a bass instrument. I recorded at 48kHz and pitched down quite a lot which sounded grainy and shit. So I re-recorded the audio at 192kHz and it sounded incredible when pitched down (pretty much 2 octaves).

It stands to reason if you have 44,100 samples per second, if you pitch down one octave, you still have a total of 44,100 samples, but over 2 seconds. Which equals 22,050Hz (even if you resample, every second sample is duplicated). By recording at 192kHz, you can pitch down 2 full octaves and still have decent resolution (48kHz) which is the difference between my elastic bands sounding shit and really good.

Bottom line is recording at really high sample rates DOES have it's uses depending on how you intend on processing the audio. Listening to sound at those sample rates, not so much.

Also, similar for bit rate. It is always good practice to process 24 bit sound rather than 16 bit. You may not notice the difference in audio quality too much simply listening to it, but as audio gets processed and re-processed, gradually you would notice quite a difference in sound quality between the two. In my experience, that is not a myth.

[ze2be]

That was nice and clear trevox. So the djs that says vinyl sounds better then cds might be right after all, because a dj almost always pitch the record for the mix.

How much of a difference regarding cpu and ram load is it to run a 24bit/88.2 live set, then a 24/44.1? Say 16 tracks with audio clips, pluss effects.

[trevox]

That was nice and clear trevox. So the djs that says vinyl sounds better then cds might be right after all, because a dj almost always pitch the record for the mix.

How much of a difference regarding cpu and ram load is it to run a 24bit/88.2 live set, then a 24/44.1? Say 16 tracks with audio clips, pluss effects.

Well I don't think anyone can ever dispute that vinyl sounds better than CD's once they are mastered properly!! While someone can say technically if you record at 44,100kHz/16bit and make vinyl and CD copies that the latter is an exact duplicate, after the vinyl mastering process (if done properly), vinyl just sounds way warmer and has a lot more character. And as you say, you can pitch it down with no chance of that grainy effect.

While I said in my post that there are uses for higher sample rates, I generally work at 44,100. If I know I will be processing audio in such a way that a higher sample rate is necessary, I will do exactly that, but it is not too often I have to. I have worked in 24bit for so long that I have never really compared CPU/RAM usage between that and 16bit. Just do a test yourself. Record some audio files at 16 bit and some at 24 bit and play them in a blank set - doesn't have to be the same source or anything. Personally I don't care about the difference in CPU usage as it really does make a difference as to how it sounds - particularly when applying reverbs and during the mastering process. But that's just me...

[Der_Makrophag]

Wouldn't 88.2 make more sense for audio tracks if the end result is 44.1?

Indeed, thats something I also wondered. I think 88.2 kHz is pretty much the best way to go! As 96 kHz is even (unnecessarily) higher and IN THEORY downsampling to 44,1 kHz will be smoother (I have no idea of downsampling algorithms, so I might be wrong here).

I agree, that when recording instruments, the samplerate is not a big issue, unless you want to pitch it down. I was rather thinking of the processing. If you guys have NI Reaktor, you can EASILY hear the big difference in samplerates in some raw "circuits". Of course, the plugins you pay for are (at least they should be)much better designed and the effekt might not be hearable.
Why should every big plugin do oversampling (that is, increasing the internal samplerate), if it does not make a difference?
So THIS is the point where you might want to work at higher samplerates. I think you will benefit from it, if you use a lot software synths and effects (the same for 24bits, as stated above). And even there, 192 kHz is the end of line. Only downsample the whole file to 44,1 kHz / 16 bit, so that errors can't sum up.