24 bit recording Location

[Phaseshifter27]

I was wondering if anyone knows why camera's do not support 24 bit audio? only 16 bit.

[John Willett]

Originally Posted by Phaseshifter27

I was wondering if anyone knows why camera's do not support 24 bit audio? only 16 bit.

Because picture people think that sound is not important and only the picture matters.

[Gerax]

Originally Posted by John Willett

Because picture people think that sound is not important and only the picture matters.

Besides...it could be that 24 bits is a lot more data to store...and according to the above statement..."it's not worthwile", picture has to look best....

[rumleymusic]

Any good videographer will not use the camera to record audio anyway. Most of the time they just leech off the sound guy who does not get paid a quarter of what they do and I don't get compensated at all for giving it to them...

[Roland]

Many camera's do support 24bit recording. Pretty much all of the Sony pro range and they represent the largest percentage of broadcast cameras. Obviously recording sound direct to camera often depends on what you are recording. If it's run and gun stuff, interviews, news conferences etc. most of the time they will record to the camera, albeit with a sound mixer attached by umbilical (sometimes by radio). This doesn't really present a problem as with the microphones being used (often radio's), and the limiting applied with an SQN the noise floor is most likely to be far worse than the camera's audio performance.

In multiple camera shoots, often the sound is recorded seperately. With music often being multitracked recording to the camera isn't an option.

Regards


Roland

[NorseHorse]

Originally Posted by John Willett

Because picture people think that sound is not important and only the picture matters.

I recorded a voice recital last weekend. A videographer was there shooting from the balcony. After the concert was done, I went up to them, introduced myself, and offered to email them an audio-for-video stream.

"No thanks."

Me:

I guess they are planning on using the shotgun mic on the camera? I went to their website later. Their "choral sample" is clipped and distorted. But then again, perhaps I shouldn't judge -- I don't really want to hear what the hardcore video folks have to say about my simple video shoots. To each his own!

[Yenrah]

I'm actually teaming up with a videographer I know to provide quality audio-visual solutions... Recital videos are the worst without good audio...plus you can't have all your eggs in one basket...

[hbphotoav]

One of my great frustrations in "video world" (especially around newbies and low-end events videographers) is the attitude that the picture is all that matters. It never has been for me. My first video camera (Toshiba 3-chip Hi8) had XLRs in and PCM audio that, when handled carefully, was nearly "CD" quality... good enough on the first attempt to land a 12-year relationship with an annual music festival. Of course, the gear has changed somewhat since 1997...

Feeding my Z1U cam with balanced line audio, and recording to -6dBfs on peaks at 48kHz/16bit, the pro video standard, (the balanced line being originated MKH8040s or Gefell M296s or DPA 4061s through a DAV BG8 through either an Ensemble or a Mackie Onyx 1220) has resulted in audio that has been perfectly satisfactory for DVD excerpts of concerts in USA, England, Switzerland and Italy. It's a bonus to have a Logic session tracked to HDD, and makes archival CDs a breeze.

But... on the consumer/prosumer side, it's telling that folks will spend big bucks on a HD camcorder, viewed on a 52" LCD or plasma... but heard through a $300 5.1 system purchased at Best Buys. It's quite possible that they actually CAN'T hear a difference... much less, care. The capability is there, through direct-in and 3rd party workarounds (Sound Devices to BeachTek) to record good audio. But the will to do so (or to spend a like amount on transducers, mics, speakers, and electronics) is lacking. My playback, BTW, is Tannoy Series 800As (with the matching sub) in the edit room (with Large Advents as a secondary "real world" check) and AR 9 towers flanking the Bravia LCD in the living room.


Add to that the proliferation of über-compressed video formats and a YouTube quality expectation... and... there you go. Education, and will. That's all that's missing.

My $0.02US.

[LX3]

I wouldn't fret about the lack of 24-bit on cameras.

While we all know it's more flexible and safer to record 24-bit and leave lots of headroom, the reality is that the audio is often getting played with by non-audio people in an edit suite. If you gave them 24-bit with lots of headroom, they wouldn't know where to start.

Plus, correct me if I'm wrong, but I'm told that Avid video systems truncate any 24-bit audio to 16-bit anyway. So your carefully-considered 24-bit headroom ends up just resulting in noisy 16-bit audio.

If you want happy video people, it's best to try and deliver solid-sounding ready-to-wear 16-bit audio.

[hbphotoav]

Originally Posted by LX3

...

If you want happy video people, it's best to try and deliver solid-sounding ready-to-wear 16-bit audio.

48kHz, 16bit, and no rendering will be required by FCP, Avid or the rest... until it's muxed down for DVD release, that is. Or compressed and converted for a web stream or YouTube submission. Different kettle of fish, those.

[ghellquist]

You could sort of ask why no recording systems records true 24bits. Yes, I know, the file format is 24 bits, but the last 4 bits are generally only noise.

In a camera it might be difficult to set up an analog chain so that it truly goes above 96dB in S/N-ratio *) . And if the customers does not ask for it ...

// Gunnar

*) shorthand "rule-of-thumb" :: one bit of info corresponds to 6dB of Signal-to-Noise ratio. This would give 96dB on a CD that is 16 bit.
If you want to go to high S/N-ratios you need to work through the whol analog chain. This generally means either transformers (heavy, expensive) or using more power and/or discreet circuits ( more power draw and larger ).

[Petrus]

Come on guys, you all (should) know that even the best recorders get only 19-20 bit real resolutiion even with nominally 24 bit files. Tape/memory card capacity and data transfer speeds are bottlenecks as they are and CD-quality is as good as anybody really needs for delivery medium. If CDs are good enough for most audiophiles with $$$$$$ sound systems, what would tv programs and DVDs gain from 24 bit audio, really? NOBODY in the world would notice.

Old miniDV had 16/48, newer HDV and memory card MPG4 formats have lossy audio: a step backwards.

I have measured my Canon XH-A1 HDV-cam to have over 90 dB dynamic range when fed with good mixer (SD302 in this case). How much better is that than the venerable Nagra used in great movies untill the nineties? Maybe 4 bits better...

Serious videographers use outside recorders anyway, and many fashionable cameras (like RED and EOS-D5II) have no audio to mention. We just shot a short company video for 3 days with Canon EOS-5DII, sound was captured to SD722 at 24 bits just for safety. That was not the problem, but converting the files for editing (that alone took 6 hours) and just copying over 200 GB of data for a few minute video takes several hours...

Longing after better than 16/48 PCM for video is just number chasing with no real benefit (exept maybe headroom safety).

[Gerax]

BTW, who cares for 24 bit capable cameras if their onboard audio converters are lame?

[Petrus]

24 bits is just a catchword, even most small "24 bit recorders" are strained to get 16 bit real resolution. Who would expect a video camera costing about the same to have better specs, when optics and video capture/storage has to be included in the price?

[LX3]

Originally Posted by Petrus

If CDs are good enough for most audiophiles with $$$$$$ sound systems, what would tv programs and DVDs gain from 24 bit audio, really?

It depends what you're doing, but 24-bit can be very useful at the capture stage. Speaking voices can be amazingly dynamic, going from a whisper to a full-on shout without warning. With 16-bit you usually need to compress and/or limit as you go to tape, otherwise you can find yourself in quantisation hell on the quiet sections. 24-bit solves that... but tends to introduce post-production issues when it lands in the hands of non-audio people. Especially Avid users. (FCP as far as I can tell is fine with 24-bit audio).

[kstrauss]

Originally Posted by hbphotoav

One of my great frustrations in "video world" (especially around newbies and low-end events videographers) is the attitude that the picture is all that matters.

A little tidbit from the educational world. I have found that my students take to video much more easily, I think it's from having watched TV and movies which generally have very high production standards. They seem to innately know how to frame a shot, color balance images, and switch multiple cameras; so I find that the video looks very good with a minimum of effort. Audio, on the other hand, is much more difficult. It takes a lot of work and effort on my part to teach students what a good recording sounds like and how to blend microphones into a cohesive mix. It seems they have little experience with good sound and even if they do, they don't know how to achieve it without assistance.

[chris319]

Originally Posted by Petrus

24 bits is just a catchword, even most small "24 bit recorders" are strained to get 16 bit real resolution.

That's not true. I have owned several small recorders and they all got the full 24 bits.

[ghellquist]

Originally Posted by chris319

That's not true. I have owned several small recorders and they all got the full 24 bits.

No. You have missed the point.

The recorder might save 24 bit files, that is true. But several of the bits only contains noise.

The grim fact is that there is no portable recorder available at any price ( even very high ) that gives the full 24 bits in true resolution. Many so-called 24 bit converters have the equivalent of 6 or even 9 bits that only contains noise.

Check the published Signal-to-Noise ratio and use the rule-of-thumb *) and divide by 6 to get the number of bits containing information. The rest of the bits are simply noise.

We should definitely remember that well done 16 bits does sound really good. The extra bits above this most often only serves as an insurance -- they allow us to reduce the recording gain and therefore gives us more margin before distortion.

// Gunnar

*) Notice that this is only a quick assessment tool. Reality is a bit more complicated.

[Ozpeter]

The extra bits above this most often only serves as an insurance -- they allow us to reduce the recording gain and therefore gives us more margin before distortion.

Which is indeed the point when it comes to location recording. No debate about that. Debate about the worthwhile sonic benefits, or not, has been done to death all over the net - and ad nauseam here too.

[Petrus]

Originally Posted by Ozpeter

Which is indeed the point when it comes to location recording. No debate about that. Debate about the worthwhile sonic benefits, or not, has been done to death all over the net - and ad nauseam here too.

100% true.

One thing we should realize though, is that with cheap recorders the 24 bit thing is so much an illusion, that we get no extra headroom at all. Thinking " yes, now with 24 bits I can set levels 12 dB lower for safety" can mean, in the end, exactly the same result as you get with 16 bit recording and levels 12 dB too low, with noise floor exactly at the same place.

Just look at the specs of new Korg DSD pocket recorder, small Olympus (great little thing, by the way), Fostex FR-2 & LE etc etc. Dynamic range is around 90 to 96 dB, only enough for 16 bits, even though they wave a 24 bit or DSD banner.

[Ozpeter]

Yes, but you are getting more digital headroom - same end result but less risk of digital clipping.

[Petrus]

Originally Posted by Ozpeter

Yes, but you are getting more digital headroom - same end result but less risk of digital clipping.

?????

I would think the end result is identical, if you record, say, at 12 dB lower level than you should with true 16 bit system and "24 bit system" which has the same 96 dB dynamic range. "24 bits" does not help a bit, pun intended.

Both would have the system noise floor at the same level, both would have at least 2 wasted, unused bits at the top. Just that the 24 bit file would be 50% larger, containing only noise.

[pkautzsch]

Originally Posted by LX3

Plus, correct me if I'm wrong, but I'm told that Avid video systems truncate any 24-bit audio to 16-bit anyway. So your carefully-considered 24-bit headroom ends up just resulting in noisy 16-bit audio.

The latest Avid does 24 bits, as they finally seem to have found out that there are lots of 24 bits recorders out there.

Originally Posted by LX3

If you want happy video people, it's best to try and deliver solid-sounding ready-to-wear 16-bit audio.

Yes. And best use a tiny bit of that SQN limiter to avoid complaints that "the sound is too low". ENG stuff isn't really post-produced audio-wise.

Originally Posted by ghellquist

The recorder might save 24 bit files, that is true. But several of the bits only contains noise.

The grim fact is that there is no portable recorder available at any price ( even very high ) that gives the full 24 bits in true resolution. Many so-called 24 bit converters have the equivalent of 6 or even 9 bits that only contains noise.

Physics say that we can't actually get more than about 126 dB of dynamic range, as the thermic noise of very well designed electronics equals -126.
Actually, 126 dBSPL are REALLY loud anyway.