Topic: 96kHz recording with JB3

[Karl]

I've been thinking about it, and it can be done, but it would take a lot of work and be a PITA.  But I thought I would bring up the thoughts anyways.  You would need two JB3's (which I don't have, but I know a lot of people do). 

Thus far, we know that the JB3 will take a 96k signal.  The file header is written wrong, that needs to be changed, but apparently that's easy.  The problem is, is that 96k is too fast for the JB3 processor, so it results in glitches and dropouts.  Now I'm going to go on the theory that lightning never strikes the same place twice.  So, if you were recording with two JB3's at the same time with the same signal (optical splitter), then both recordings would end up with random dropouts, but not in the same place.  Here's where the hard work comes in--you would need to load both files into some sort of wav editor where you can do two tracks of stereo at the same time, and be able to line the files up against each other (I can do it in Samplitude).  You would then have to painstakingly go through and take all of the good parts and throw away the bad parts, save them as one file, then there you go!  A clean 96k recording.

Food for thought.

[jaguaracer]

interesting...
[dumb question] Are we sure the drop outs be random though?  Maybe the dropouts are at the same time marks for all JB3 because of the physical limitations? ie could be that after 5 seconds of 96 recording, the physical limitations of the JB3 cause all JB3s to drop out at exactly that point, then 15 seconds later, all JB3 drop out again[/dumb question] This is maybe far-fetched, but I'm there are people out there with a couple JB3s who could test this out. I am interested too. 
edit: although I guess one JB3 is all you would need to test this. Just run two separate 96 recordings and see exactly where the drop off occurs. Although if the drop outs do occur at exactly the same spots that wouldn't be conclusive evidence as two JB3 could produce different drop out points.

[Karl]

But then you would just start the recordings at two different times.

[jaguaracer]

yeah, but you would only be testing where the drop outs are on both.
If they both drop out at say exactly 0:05.85 then there might be a problem but if they drop out at completely different spots, you are in business then.

[admkrk]

that all sounds like a lot of work. i'm not sure yet, but i don't understand why recording at a higher frequency is any better? i don't know of any mics that do better than 20-20, so it seams to me that you aren't gaining anything above 44.1.  the only advantage i can see is in increasing the bit rate.

the sampling therem states that a continuous band-limited signal can be replaced by a discrete sequence of samples w/out loss of any information and describes how the original continuous signal can be reconstructed from the samplles; furthermore, the therem specifies that the sampling frequency must be at least twice the highest signal frequency.

the way i read that, unless you have a mic capable of picking up a 48khz signal, what's the points.

[creekfreak]

that's crazy talk

[admkrk]

like it says,        i'm an idiot!    i understand raising the bit rate, but just haven't figured out any advantage to raising the frequency?

[BC]

that all sounds like a lot of work. i'm not sure yet, but i don't understand why recording at a higher frequency is any better? i don't know of any mics that do better than 20-20, so it seams to me that you aren't gaining anything above 44.1.  the only advantage i can see is in increasing the bit rate.

the sampling therem states that a continuous band-limited signal can be replaced by a discrete sequence of samples w/out loss of any information and describes how the original continuous signal can be reconstructed from the samplles; furthermore, the therem specifies that the sampling frequency must be at least twice the highest signal frequency.

the way i read that, unless you have a mic capable of picking up a 48khz signal, what's the points.

some mics have extended freq. response up to 40 or 50 KHz, Schoeps CMC6xt, Sennheiser MKH800 come to mind.


Although regarding the original post it seems lile too much work to just gain some extra frequency response when most PA systems are not putting out anything over 15KHz anyway.

[mmmatt]

What would it take to record @ 24bit with a jb3?  I don't know enough to even guess what would have to be done.

Matt
 ***edit*** ...and does a 48K recording mean that is the highest audible frequency it can record?  I thought there was more to it than that.  Isn't that the sample rate... like how often the signal needs to refresh itself before it saves or something like that?  Like I said I have very limeted knowlege of this kind of thing.

Matt

[Brian Skalinder]

What would it take to record @ 24bit with a jb3?

Can't be done.

and does a 48K recording mean that is the highest audible frequency it can record?

Highest capturable frequency at a given sample rate = sample rate / 2.  So 24 kHz is the highest we can achieve at 48 kHz sample rate.  At 44.1 kHz sample rate, highest frequency would be 22.5 kHz.

[admkrk]

that all sounds like a lot of work. i'm not sure yet, but i don't understand why recording at a higher frequency is any better? i don't know of any mics that do better than 20-20, so it seams to me that you aren't gaining anything above 44.1.  the only advantage i can see is in increasing the bit rate.

the sampling therem states that a continuous band-limited signal can be replaced by a discrete sequence of samples w/out loss of any information and describes how the original continuous signal can be reconstructed from the samplles; furthermore, the therem specifies that the sampling frequency must be at least twice the highest signal frequency.

the way i read that, unless you have a mic capable of picking up a 48khz signal, what's the points.

some mics have extended freq. response up to 40 or 50 KHz, Schoeps CMC6xt, Sennheiser MKH800 come to mind.


Although regarding the original post it seems lile too much work to just gain some extra frequency response when most PA systems are not putting out anything over 15KHz anyway.

i didn't say there weren't any mics that could do it, just that i wasn't aware of them.

and, exactly my point.

24 bit compared to 16 bit,      that's a different story

[mmmatt]

so all I'm getting by recording at 48k instead of 44.1 is a nicer sounding glass clink by the bar?
That just doesn't seem right.

Matt

[admkrk]

so all I'm getting by recording at 48k instead of 44.1 is a nicer sounding glass clink by the bar?
That just doesn't seem right.

Matt

not nessessarily,   as i understand it, people don't hear frequencies above 20khz to begin w/, so unless you're recording a dog wistle for the pooch, you won't "hear" a difference either way. that doesn't mean you can't "feel" the sound. i'm sertain that although i can't hear it i can feel sounds below 20hz, so why shouldn't it work the same way at the other end? it's more of a subliminal deal. it's the bit rate that needs to be higher to get a fuller sound.  unless someone can convince me otherwise. 

[mirth]

I thought that sample rate has nothing to do with the frequency response of your mics... Meaning, it can be compared to frame rate when making a video - higher frame rates mean that more data is 'captured'.

this has me cornfuzed...

[nic]

I thought that sample rate has nothing to do with the frequency response of your mics... Meaning, it can be compared to frame rate when making a video - higher frame rates mean that more data is 'captured'.

this has me cornfuzed...

in your analogy, the frame rate would be comparable to bit-rate in audio, not sample size

[BJ]

What would it take to record @ 24bit with a jb3?

Can't be done.

what is the reason for this...is it firmware that wont allow it?  or is it a hardware issue??

[mirth]

I thought that sample rate has nothing to do with the frequency response of your mics... Meaning, it can be compared to frame rate when making a video - higher frame rates mean that more data is 'captured'.

this has me cornfuzed...

in your analogy, the frame rate would be comparable to bit-rate in audio, not sample size

Thought bit rate is the amount of data captured and sample size is how many times per second its captured.... I know that higher can be considered 'better' but I'm still not clear on all of this. I thought I was, but apparently not.

[nic]

Thought bit rate is the amount of data captured and sample size is how many times per second its captured....

this is correct.

continuing with the video analogy, think of audio samples as the screen size.
an example(my video knowledge is sorely lacking in details):
(video)24FPS@640x480 == 16/44.1(audio)
(video)29FPS@1024x768 == 24/96(audio)

[mmmatt]

Thought bit rate is the amount of data captured and sample size is how many times per second its captured....

this is correct.

continuing with the video analogy, think of audio samples as the screen size.
an example(my video knowledge is sorely lacking in details):
(video)24FPS@640x480 == 16/44.1(audio)
(video)29FPS@1024x768 == 24/96(audio)

So is 16/48 the same as 800x600 or 640x480 with less screen flicker?   (Needless to say, I'm still lost)
Matt

[nic]

Thought bit rate is the amount of data captured and sample size is how many times per second its captured....

this is correct.

continuing with the video analogy, think of audio samples as the screen size.
an example(my video knowledge is sorely lacking in details):
(video)24FPS@640x480 == 16/44.1(audio)
(video)29FPS@1024x768 == 24/96(audio)

So is 16/48 the same as 800x600 or 640x480 with less screen flicker?   (Needless to say, I'm still lost)
Matt

staying with the analogy, I would say that 24FPS@800x600 == 16/48
remember, this is just an expression, the really do not correlate at all

[mirth]

how would a faster sample rate then allow for higher frequencies to be captured? wouldn't that be more the domain of increased bit depth and the larger dynamic range it provides?

or is it because you're taking more samples per second that the possibilities exist for capturing fleeting high frequencies?

[BC]

how would a faster sample rate then allow for higher frequencies to be captured? wouldn't that be more the domain of increased bit depth and the larger dynamic range it provides?

As Bri posted earlier, the Nyquist theorem of digital recording says that the highest freq we can record is 1/2 of the sample rate.   So at 48KHz the highest freq we can record is 24KHz.

Just to remember. in digital recording

bit depth => dynamic range (theoretical of 6dB/bit)
sampling freq => frequency response. (Nyquist)

or is it because you're taking more samples per second that the possibilities exist for capturing fleeting high frequencies?

basically, yep!

[mirth]

eeeeeeeeinteresting... Thanks Ben

[jk labs]

how would a faster sample rate then allow for higher frequencies to be captured? wouldn't that be more the domain of increased bit depth and the larger dynamic range it provides?

or is it because you're taking more samples per second that the possibilities exist for capturing fleeting high frequencies?

Ever seen a video of a forward moving car with the wheels spinning backwards?

Think of it this way. You have a nice clean 48 kHz sinusoidal coming towards you with tops and crests. Like waves in the ocean. 
And you have a 48 kHz sampler. The sampler takes snapshots (16 or 24 bit) of the wave and hits say the top every time.
The sampler sees no wave. It outputs DC!

To see the 48 kHz you have to sample fast enough to see the crest in between. If you increase the rate to 49 kHz you see a slowly varying wave. At 95 kHz you see a fast varying wave.  And finally, at 96 kHz (and over), you see the full 48 kHz.

[mirth]

+t, I get it now.

Thanks Jon (that's the J in JK, right?)

[Ed.]

this thread just made my head hurt a bit.

[creekfreak]

this thread just made my head hurt a bit.

I was thinking the same thing

[mmmatt]

this thread just made my head hurt a bit.

I was thinking the same thing

yeah... ouch

[mirth]

Mine too, but at least now I have a better understanding why folks want to record at 96Khz & beyond.

[dklein]

Once your heads stop hurting...take that example (sample rate=96k, wheel spinning at 48k) and contemplate what you would see if the wheel sped up to 49 kHz.

It would appear to move slowly, an illusion known as aliasing!  On playback we would be creating something that wasn't there - a by product of our sampling rate.  This is the reason for 'brickwall' filters that don't let any frequencies beyond 1/2 the sampling rate into the ADC - to avoid aliasing.

[taktheride]

not nessessarily,   as i understand it, people don't hear frequencies above 20khz to begin w/, so unless you're recording a dog wistle for the pooch, you won't "hear" a difference either way. that doesn't mean you can't "feel" the sound. i'm sertain that although i can't hear it i can feel sounds below 20hz, so why shouldn't it work the same way at the other end? it's more of a subliminal deal. it's the bit rate that needs to be higher to get a fuller sound.  unless someone can convince me otherwise. 

So do you record 44.1k or 48k?     I ask because I'm not sure its theres really much point in having the convert to burn to cd later.  Does that make sense?

[admkrk]

not nessessarily,   as i understand it, people don't hear frequencies above 20khz to begin w/, so unless you're recording a dog wistle for the pooch, you won't "hear" a difference either way. that doesn't mean you can't "feel" the sound. i'm sertain that although i can't hear it i can feel sounds below 20hz, so why shouldn't it work the same way at the other end? it's more of a subliminal deal. it's the bit rate that needs to be higher to get a fuller sound.  unless someone can convince me otherwise. 

So do you record 44.1k or 48k?     I ask because I'm not sure its theres really much point in having the convert to burn to cd later.  Does that make sense?

for a nice timely answer  ,

i do 44.1. i'm pretty sure that somewere in my chain it would get converted automaticly(i haven't tried). i don't think i would hear any inprovement in sound if there is one. when i upgade to a lappy and can do 24 bit, i'll probly switch to 48 or even 96, but i just don't see the point until then.

not trying to bring back any head aches, just answer a ?. i understand a lot better now and since i currently run dat(16 bit), i see no real advantage in going to 48k.

[Ray76]

I may get good at recording someday, but somehow I think I will never understand what I know, if you get what i mean. theorems , theorys, frequencies and all that mathematical stuff make me queasy. I keep reading that folks with music degrees are supposed to be adept at math and engineering talk, since basically music is math based, but I mustve missed that day.

 

+T to yall that can comprehend that stuff.