Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Brian Campbell]

While the discussion of high sample rates rages on it occurred to me that some of the most cherished microphones (U47, U67, C12) have a limited high frequency response. I'm fairly certain my U67s roll off dramatically after 20khz. It seems to me recording at 96khz, 192khz with these microphones would be pointless. I would like some clarification.
regards

[klaus]

Hello Brian,
High frequency response is but a small, and I would say largely insignificant, aspect of a microphone's value as a recording tool for music.

What truly matters is how much detail-some call it resolution-of the original musical event gets passed along. Some of the most intimate, emotionally rich (what else matters?) recordings are and were made with ribbon mics whose high frequency response beyond 12kHz is usually very poor. The architecture of the mic's processor and its circuitry components play a huge role in the ultimate value of a mic as a recording device that can deliver truth or pleasure (what else matters?).

If you want to concentrate on what frequency response alone contributes to the value of a mic as a superior recording tool, it may be more rewarding to look at the ups and downs - where they happen and how large they are- within the response curve: certain boosts and attenuation throughout the curve, rather than ruler-flat response, make the difference between a very good and poor microphone.

Put another way: There are Chinese-made condenser mics retailing for less than $100 whose frequency response looks ruler flat, up to 20kHz, on paper. Are they any good? Doubtful.

[Piedpiper]

And re: sampling rate, the relevance of high rates has more to do with the integrity of the signal, as well as processing of that signal, within the normal accepted hearing range than supersonic response.

[Brian Campbell]

Thank you Klaus. I should have clarified my post by saying that I'm not an advocate for extreme sample rates. I'm happily recording at 24bit 48khz. I'm seeking confirmation that the microphones mentioned are not extended into the ultra sonic above 20khz and certainly not 48khz or 96khz.
It's always a pleasure to use such microphones, they are icons for good reason.

[Brian Campbell]

And re: sampling rate, the relevance of high rates has more to do with the integrity of the signal, as well as processing of that signal, within the normal accepted hearing range than supersonic response.

Yes and that's another discussion (can of worms)

[Kai]

I'm fairly certain my U67s roll off dramatically after 20khz. It seems to me recording at 96khz, 192khz with these microphones would be pointless.

There's a big difference in the way a mic rolls off (12-18 dB/octave) and an ADC/DAC cuts off the high frequency band (ca. 1000 dB/octave !).
The very steep cutoff a ADC working at 44.1 or 48 kHz needs to avoid aliasing is causing time domain problems in the upper octave of the usable band, known as "ringing".
Same applies to the reconstruction filter in a DAC, and I can asure you, decades of experience has show, each of those sound different.
An example: once I did a (very carefully set up) private listening comparison of 6 high class CD players.
I could clearly determin between all six.
Funny sidenote: the one that sounded the most brilliant and transparent had a measurable treble rolloff of 3 dB @ 20 kHz! The smoother filter resulted in (measurable) better time response, which by far overrided the HF energy loss in the audible result.
In fact it sounded as if it had more treble!
Other sidenote: when MP3 stuck it's ugly head out of the dirt it never should have left I wasn't very good in identifying it in listening tests.
I just did listen to spectral differences where higher rate MP3's don't show much difference to the original.
Today I don't have any problem identifying even 320 Mbit/s MP3's, partly even without the original as comparison. The artifacts often are so obvious, if you have a quality speaker or headphone system.
Probably it's what Neil Young had already discovered long time ago.

[Timjag]

When I'm not recording I work in a music venue, and believe me MP3s are the bane of my life, the high end is so poor - even at the higher rates that it sounds like white noise - which of course it is.

Sorry to derail the thread.just had to get that off my chest

[Brian Campbell]

It was not my intention to start a sampling rate discussion. Suffice to say that a U67 does not reproduce any significant signals in the ultra sonic range, say for the sake of argument above 25khz?

[Kai]

To come back to your primary question:
The U67 is not band limited and therefore has signal content above 20 kHz.
So if you want to keep everything the microphone catches you need to record at least at 96 kHz  sampling rate.
On the other side, if your recording will end up at 44.1 kHz it's questionable if not the final sample rate conversion will eat up this advantage.

But If you do intermediate processing this will benefit from the higher sample rate, specially equalizing sounds completely different in the higher bands.

[boz6906]

Recording at 96kHz or 192kHz samplimg rate doesn't mean you're recording audio at 96 kHz, just that the signal is being sampled at that rate.
IIRC, at 96kHz the UFX upper spec is -1dB at 45kHz.  A big advantage of higher sampling rates is moving the anti-aliasing filter's cutoff far above human hering response.

Regarding a mic's high freq. performance, I think smooth HF response is more important than extended HF output.  A good example is the RCA ribbons.  The 77DX is down about -10dB at 18kHz (no 'air' there!) but all agree they aound great.

[Jim Williams]

Higher sample rate/bandwidth benefits include reduced phase shift of the filters, plus lower noise and better THD in some cases. Double the sample rate and you have double the "slicing" of the waveforms for easier "re-assembly".

Even with 20K hz bandwidth limited mic designs the added transparency is easily heard. You don't need 45k hz bandwidths for human voice but it will sound more natural. Use extended bandwidth 1/2" condenser mics on instruments like brass or percussion and you will hear/feel the closeness to the source. That stuff has large harmonics past 50k hz so any 22k hz bandwidth digital recording chain will low pass that sonic information away permanently at -100 db per octave. Even modified analog recorders can do 32k hz flat at 30 IPS, seems a shame to throw all that extra bandwidth away but that's what we have done for decades.

[Brian Campbell]

To come back to your primary question:
The U67 is not band limited and therefore has signal content above 20 kHz.
So if you want to keep everything the microphone catches you need to record at least at 96 kHz  sampling rate.
On the other side, if your recording will end up at 44.1 kHz it's questionable if not the final sample rate conversion will eat up this advantage.

But If you do intermediate processing this will benefit from the higher sample rate, specially equalizing sounds completely different in the higher bands.

Has anyone here measured the top end response of a U67? What would the upper end extend to? Is this roll off a function of the capsule or the electronics?

[klaus]

You can find the U67's frequency response on the Neumann website:
http://www.neumann.com/?lang=en&id=hist_microphones&cid=u67_publications

then look here:
Brochure U 67, 03/1966, English

[boz6906]

Thanks Klaus, freq response is 30 - 16kHz, looks to be around -6dB at 20kHz with a steep slope, no 'air' there.

The U67 HF response is similar to the RCA ribbons.

http://www.coutant.org/rca77dx/

[Kai]

I would not exactly call this steep.
If you extrapolate the curve it looks like a 12 dB/octave slope.
This means 40 kHz is only 18 dB down, this could be brought back with an equalizer.
It's exactly what Schoeps does with their "CMC 6 Ug xt" -type mics.
If you want to see a really steep slope look at the brick wall antialiasing filter of any ADC.
It goes down to at least -96 dB or even -140 dB within the range of, e.g., >20 kHz to 22 kHz (~1/10th octave).
This is about 1000 dB/octave!