HD Video, what about the audio?

farss wrote on 12/1/2005, 2:54 AM
So we've had all these thought provoking discussions about what id HD, and cameras and lenses and 24p and a whole host of things about the image but what about that rather more important part of our production, the soundtrack?
Why do I think this is important. One things for certain, a HiDef image on a large screen says "Stop, watch me!!!" but along with that we also stop and listen. So what our productions will be targetting isn't where we're mostly at today, something that's heard with a high ambient noise, instead the lights will be dimmed and everyone will be focussed on our production with way less distractions than SD permits. If nothing else this means our audio needs way more dynamic range which means for most of us much better kit.

Sorry if I'm being a bit obtuse here but with all the discussion about HD video this issue seems to be totally ignored and my gut feeling is it deserves way more attention than it's getting.
Bob.

Comments

Edin1 wrote on 12/1/2005, 3:26 AM
I don't know what to say here, except that using MPEG Layer II for HD audio isn't the right choice. I mean, with SD MiniDV tapes we had uncompressed PCM 48kHz, 16-bit sound, and now we get MPEG Layer II compression instead of pure PCM.
So, we moved on to higher video quality, and at the same time to lower audio quality. Stupid choice! And I wouldn't be so angry if the MPEG audio didn't have audible difference and artifacts. They would have been better off with MP3 or AC3 at 256-320kbps.

And please, don't give me that same old argument about audio compression saving space!
The ratio between 1.4Mps and 27.46Mbps is too high to be taken into consideration, so when you save about 1Mbps which makes it easily audible, but not so easily visible, it just doesn't add up.

I am not completely sure if you are talking about audio recording in the field, or post-editing, probably both, but I would say that 48kHz, 16-bit Stereo still has a fair amount of room if well recorded and edited. The preamp and conversion circutry, as well as the choice and placement of microphones is more important to me.
If the audio you are bringing to the A/D converter doesn't have the noise floor lower than, and the dynamic range higher than 16-bit, then using 24-bit doesn't help much.
I did notice a few times where I could benefit from the 24-bit audio, and I agree that it is a good thing to have, but I also noticed how more often the analog circuitry could have been better, and that its noise and dynamic range were the limiting factors in the final audio, not the 16-bit resolution.
farss wrote on 12/1/2005, 3:54 AM
I'd kind of assumed that the audio that goes with the HDV cameras is only there as guide track if you know what I mean :)

I was more thinking of how we work in post and how the overall quality of our audio needs to be much higher to match the larger hi resolution images, particularly given the likely different viewing experience.
One thing that I'd just thought of is surround sound, ever noticed how those rear speakers end up almost in the ears of the viewer, that means they're getting the sound straight into their ears so any defects will be more noticable.
I guess the other angle to this is that whilst most of will not be able to afford the very best of the HD cameras and lenses for some time to come, the very best audio kit is sort of within reach.
Bob.
Spot|DSE wrote on 12/1/2005, 4:09 AM
You've got the right concept exactly, Bob. It's not the MPEG1/Layer 2 audio that people need to concern themselves with. For dialog, it's simply fine. The converters (in the Z1/A1 are the best Sony's ever used (which isn't saying much) but if PD 170's with PCM were fine, I challenge anyone to honestly hear the difference in a blindfold listening test.
But...with bigger screen productions making their way into smaller rooms, the audio carries with it a certain expectation. Music, Foley, elementary, SFX, Sound Design, and dialog all carry a higher expectation than DV has carried, simply because the visual has greater resolution. This means better monitors, better pre's, better everything down the line for many folks. Video guys that have great audio chops probably have most if not all, of the necessary tools needed to achieve this, except now much more time and attention must be planned for the project so that it stays on par with other "high def" projects.
I would agree wholeheartedly with you on that.
ottowr wrote on 12/1/2005, 5:28 PM
The biggest difference between pro audio and audio for video is bit depth. No-one in pro audio would consider working with 16 bit audio until a cd mix is required, and then it would be converted as late in the process as possible.

Seeing that the DVD-Video spec allows for stereo PCM audio at 48/96kHz at 16, 20 and 24 bit, I'm surprised that more product doesn't use the better bit depths available. I have only one DVD (James Taylor Live at the Beacon Theatre) that has 48kHz 24bit audio, and it sounds fantastic.

Using 24 bit (or higher) gives a lot more resolution, meaning the levels don't need to go near 0dBFS to prevent distortion in the lower level passages. Seeing that the reference level for video soundtracks is at a low level (-10 or -20), it is no wonder that the majority of 16 bit audio-for-video sounds like rubbish. (to an audio engineer).