Welcome to our community

Be a part of something great, join today!

Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

24-bit audio = no riding the levels needed?

C

combatentropy

Banned
Joined
Apr 22, 2007
Messages
139
Reaction score
0
Points
0
Website
www.combatentropy.com
Red One's manual says that because audio is recorded at 24 bits, no input limiting is needed -- that is, like how the video sensor records the picture raw, so you record the sound raw and adjust levels in post.

This intrigues me. By my reckoning, 24 bits = 144 dB of dynamic range. 144 dB is indeed deafening. It would be great not to have to ride the levels throughout a shoot. It does allow setting the amplification at first for your microphone, but it seems to say you don't have to watch the levels as with lower-resolution sound digitization.

Has anyone tried this out?
 
Pretty much all location audio pros are used to having some kind of limiting functionality available, and just about all pro-level gear has it. Red's audio is still a very new feature... perhaps time will tell?

Note, I am not aware of any audio bandwidth ADC converter ever made which has a true 24-bit SNR, the last few bits are just noise. Around 22 bits is as good as you can hope for (but that is still pretty good).

I have yet to see real measured SNR, DNR range specs posted for the Red system.
 
I think that any limiting should be done to the signal before it enters the camera. Camera limiters are very often noisy aren't always the best solution for recording the cleanest signal.

David
 
The analog front end of most audio gear is where signal really get messed up bad. A digital AES EBU optical input would be awesome to have on the camera.
 
I begged for a digital input for the RED. . . to no avail.

:-(

Stephen
 
audio seems like it is always at the bottom of the list with manufacturers. I wish it were different.
 
it's dissapointing that RED was unable to install audio technology that has been there for the past decade in it's camera... it's a shame how silly little $2000 prosumer video cameras can do hissless audio with their xlr plugs.

No digital cinema camera today can be complete without audio. from RED, we aren't asking for revolutionary audio... it could've been if they did what the 744T does, as in 48 bit audio - 4 channel w/std xlr mounts... we are simply and "Kindly" asking for something we can use.

Please Red, audio is 50% of what we see.

for all the guys who are going to go like.. oh well get a sound crew... (i'd advise you to read the post again and define the argument raised in this post before you open up your mouth)
________
Buy vapir oxygen
 
There is NO device now, nor will there ever be one, that achieves 48 bit quantization! This would represent a dynamic range that can not be approached at any temperature above absolute zero as thermal noise would dominate the spectra. 24 bit dynamic range approaches the theoretical limits with the signal magnitudes that are used in audio applications.

Regardless of the quantizer width, in order to extract the maximum performance from any digital acquisition system, the input signal's peak value must come close to the FS range of the A->D. If not, the signal will not exercise all available levels. Setups where the full dynamic range is utilized is not a trivial task to configure and will, in real world situations, only be partially achieved.

The audio performance that can be obtained with the RED ONE depends upon how well your ancillary equipment comes to the meeting the conditions that the camera requires. Microphone choices is one area that will have the greatest influence on the obtainable outcome. If paired with the right equipment, the results will nearly equal any that which can be achieved with a secondary sound system, but only if you choose very well.

Kevin Halverson
 
khmuse said:
There is NO device now, nor will there ever be one, that achieves 48 bit quantization! This would represent a dynamic range that can not be approached at any temperature above absolute zero as thermal noise would dominate the spectra. 24 bit dynamic range approaches the theoretical limits with the signal magnitudes that are used in audio applications.
Click to expand...

I expect he meant to say 48KHz.
 
No i actually meant 48-bit.. my bad.. khmuse is definately more knowledgable than I, it's just that i think i read that the 744t was able to do 48-bit recording... i re-checked their website it seems that it only does 24-bit... again my bad.

and regarding the red one camera having only certain mics being compatible with it.. or certain setups being compatible with it is just not right.. especially when smaller prosumer cameras can get it done.
________
Ford Gpa Specifications
 
Note also that you don't confuse the 144db "dynamic range" of 24-bit audio with "loudness" in your "deafening"-link. Dynamic range describes the "gap" between the noisefloor (quatization noise, tape hiss etc) in an audio device and the top level where clipping occurs, where loudness is a measure in db with reference to 0= 20mPascal wich is the lowest level a normal human can hear. It describes how loud a sound is to the average listener. Basically you should still watch your levels and avoid standing next to a spaceshuttle taking off unless you are wearing hearing protection :blink:

Morten
 
I misspoke earlier saying 22 bits was the best. I believe the Cirrus Logic CS5361 "24-bit ADC" is representative of current state-of-the-art for audio performance. http://www.cirrus.com/en/products/pro/detail/P1001.html

The CS5351 is advertised to have 114 dB dynamic range, which is 114/6 = 19 bits. That is the highest advertised SNR for any ADC chip that I have seen, and you will get the datasheet SNR only if your surrounding circuitry is flawless. I am not criticizing any particular unit, but just explaining that you should not expect anything near "24 bit" (144 dB) performance from any audio input device.
 
Moqtev said:
don't confuse the 144db "dynamic range" of 24-bit audio with "loudness" in your "deafening"-link.
Click to expand...

To qualify my question, if the camera's audio circuits were quiet enough, would you set your peak lower? Say, if you used to set the peak at -20 dB with a 16-bit camera, would you set it at -60 on a camera with 24-bit audio? Again, if the noise floor was low enough. By the way, is there audio equipment anywhere (such as by Sound Devices) where the noise is low enough to set the field recording peak at -60 or even -80 dB?
 
Moqtev said:
Note also that you don't confuse the 144db "dynamic range" of 24-bit audio with "loudness" in your "deafening"-link. Dynamic range describes the "gap" between the noisefloor (quatization noise, tape hiss etc) in an audio device and the top level where clipping occurs, where loudness is a measure in db with reference to 0= 20mPascal wich is the lowest level a normal human can hear. It describes how loud a sound is to the average listener. Basically you should still watch your levels and avoid standing next to a spaceshuttle taking off unless you are wearing hearing protection :blink:

Morten
Click to expand...

The threshold for human is much lower than "20mPascal" by a large factor! In fact, 20 milli Pascal is a SPL of 68 dB which is near the level of a quiet conversation. The generally accepted (typical) threshold of sensitivity is actually 1000 times lower (near 20 uPA for a midband stimulus). This corresponds to a SPL of near 0 dB.

There is another aspect to human hearing that should be understood when discussing the dynamic range of our perception. Our hearing system has an AGC like behavior (as does our vision system) that allows us to adapt to different environmental conditions. Yes, under ideal conditions for a person with excellent hearing we can just detect a mid band sound source of 0 dB SPL, but only after adapting to a quiet condition for a period of time first. We can not be exposed to a loud source and then immediately detect a very quiet one. The same for vision, we can see in very low illumination conditions once we reach the night adapted vision condition. In vision, the mechanism is our iris but in hearing its not as readily as apparent what causes the adaptive (AGC) behavior, but it is still part of a humans hearing system.

A final point to make is that it is not safe to interchange S/N ratio and dynamic range. Though both are typically expressed in the unit of dB, they are measurements of different phenomena. Consider a situation where a qualitative evaluation of a sound system states that the acceptable noise floor is -60 dB (.1%) for any given condition. Now, if an acquisition system exhibits a noise floor of -96.3 dB (2^16) its usable dynamic range would be reduced by the minimum acceptable noise condition (96.3 -60 or 36.3 dB). This gives you a range of just of 36 dB (about 65:1) that meets the criteria. This is far less than the 16b performance of 65,536:1 that the quantizer is capable of. The engineering specification which best expresses this is one rarely used known as equivalent input noise. This measurement gives one a great tool as it takes into account the different gain conditions, but unfortunately its not often used in consumer specifications.

Kevin Halverson
 
I am having a very hard time understanding the choices made by RED regarding audio...

24bit 48khz are specs which are only relevant to the digitization, and have nothing to do with the quality of the soundrecordings as such. 24bit, simply means that it doesn´t "clip" digitally eventhough the preamped signal is distorted as hell, and what is so marvelous about that? How does that improve the soundquality?
Secondly, everybody knows that it is important to have a "hot" level, to hit the sweetspot of the preamp compared to the loudness of the soundsource, which often require some small audio adjustments while shooting... how are these adjustments made?
thirdly... I know people love to scream about hiring a soundguy to ride the levels, and beeing professional etc... well, either RED intended REDONE to be able to make soundrecordings or they didn´t, and4 channels 24bit 48khz + phantompower doesn´t seem like your average reference mic to me.
 
I also opted for a digital sound input on the red, simply because i knew, that it would have given us much better options. Consider for instance a 4mic from coresound just going in to red through a toslink. That way you would be able to ride the levels on the 4mic AND you would have your preamp nicely separated from the roaring horsepower of the REDONE monster... this is too late now, but a digital input module would definately make my day.
 
Does Red have plans to make thier own mike?:detective2:
 
Everyone should experiment and find their own working methods, but on the audio side we've been recording at 24 bit 48k for years and riding gain the whole time. The dynamic range of real life is still much greater than can be "described" by 24/48 recording without a great loss of detail and the addition of noise unless you optimise your gain structure from moment to moment. The Red is a video camera, a nice camera, but a camera and not a device optimised for sound recording. For most things that I see get shot on video the audio seems like it will be fine--equivalent to many other pro video cameras working now. For higher-fi audio or greater dynamic range going double system would be a good idea.

Philip Perkins CAS
 
You must log in or register to reply here.
Back
Top