AF-101 16 Bit Audio Recording ?

[TheReverend]

Thanks for the informative information Barry,

I come from a music background and haven't seen 16Bit in modern equipment for a while now, I'll hang on to my H4N and test the Audio side by side when my AF-101 arrives in Dec/Jan.

Anyway, always better to have more than one audio recording device.

I've done significant audio work as well with video being the last 2-3years. 16bit 48khz is great for single mic recordings and the inputs on the DVX100b and the HPX170s I've used have had phenomenal sound quality. 24/96 is really prorecord album territory or promovies where everything is rerecorded in the studio. (foley, voices etc)

 

[Jade Lane]

Thanks for the info Petros,

I wish these things done what they actually said, I've used the H4N with my 5DMK2 but get fed up with the cage and gear everywhere. I really hope the AF-101 sounds good and then I'll lose the H4N forever.

 

[Petros Kolyvas]

Oh and there's this: http://www.martin-doppelbauer.de/fieldrecorder/distortions.html

Certainly doesn't inspire any real confidence in the H4n!

But, FYI, I have one because it's a $300 audio recorder I can put directly in harm's way and not be worried - so everything has its place.

 

[Petros Kolyvas]

Thanks for the info Petros,

I wish these things done what they actually said, I've used the H4N with my 5DMK2 but get fed up with the cage and gear everywhere. I really hope the AF-101 sounds good and then I'll lose the H4N forever.

I feel like such a party pooper some times. The funny thing is that it can "record" 24 bit, it just can't resolve it. 24-bit audio has become a necessary audio marketing tool - just not a camera marketing tool.

Don't feel bad though - take the internal recorder of the SD 552 - a $3000 mixer/recorder. SD does publish specs and that recorder can resolve about 103dB (A-weighted) of dynamic range. At 6dB of dynamic range per bit in PCM audio, that's, well only 17.2 bits. Since .2 of a bit is impossible and no one tracks at 20-bit, even with that recorder you need track 24bit audio to get about 1 extra real bit of dynamic range above a 16bit signal.

Now, for the real soundies out there, of course there's some maths and dithering going on down there but, as usual, I'm painting with some broad brushes here.

My FF800, arguably a decent middle of the road audio interface, does about 117 on the AD side (what's that 19bits?). Even the very best converters will only do about 124dB of dynamic range (24-bit is 144dB max).

Now, of course there's a lot more to audio than dynamic range: frequency response, transient response, etc. etc.

So, when the H4n doesn't resolve 24-bits of dynamic range that's ok. It still sounds more than decent to my ears!

And ultimately, what that means is that with all these devices you can track at 16-bit in many cases, save some disk space, and still get very very high quality audio.

If your signal is peaking up in the -12dBFS range when recorded and your recorder can't resolve below about -96dB anyway (not to mention once you add mic self-noise, pre-amp noise, etc.etc.) why track 48 extra bits of data?

I'll stop flogging the horse now.

 

[Yak]

Barry when you get the camera for testing could please do a check on the audio including recording, there was a problem with the HVX 200 where some cameras had a slight beeping durning recording Panasonic was good in fixing it but it is better to "nip it in the bud" and Sony's PD-150 DVCAM had very noise preamps, I am sure that Panasonic's will be good but we would still like to know just in case, Thanks.

 

[Postfade]

16 bit gives 98dB of dynamic range against I think 108db for 24 bit. Yes I record in 24bit in my sound dubbing suite....but it's not worth 'throwing the toys out of the pram' because it's only 16 bit. The real thing is how quiet are the mic amps, as it's not worth going for ultra dynamic recording if the input signal to noise isn't up to it.

 

[trez]

16 bit gives 98dB of dynamic range against I think 108db for 24 bit. Yes I record in 24bit in my sound dubbing suite....but it's not worth 'throwing the toys out of the pram' because it's only 16 bit. The real thing is how quiet are the mic amps, as it's not worth going for ultra dynamic recording if the input signal to noise isn't up to it.

24bit is 144dB, not 108.

 

[Darren Levine]

i second the notion to use better recording practices before counting bits. you could have 64bit/256kh audio hooked up to a $60 shotgun pointed off axsis. the sound it's picking up of a squirrel next to the actor will be lovely, but a lovely fail nonetheless

 

[cowpunk52]

I feel like such a party pooper some times. The funny thing is that it can "record" 24 bit, it just can't resolve it. 24-bit audio has become a necessary audio marketing tool - just not a camera marketing tool.

Don't feel bad though - take the internal recorder of the SD 552 - a $3000 mixer/recorder. SD does publish specs and that recorder can resolve about 103dB (A-weighted) of dynamic range. At 6dB of dynamic range per bit in PCM audio, that's, well only 17.2 bits. Since .2 of a bit is impossible and no one tracks at 20-bit, even with that recorder you need track 24bit audio to get about 1 extra real bit of dynamic range above a 16bit signal.

Now, for the real soundies out there, of course there's some maths and dithering going on down there but, as usual, I'm painting with some broad brushes here.

My FF800, arguably a decent middle of the road audio interface, does about 117 on the AD side (what's that 19bits?). Even the very best converters will only do about 124dB of dynamic range (24-bit is 144dB max).

Now, of course there's a lot more to audio than dynamic range: frequency response, transient response, etc. etc.

So, when the H4n doesn't resolve 24-bits of dynamic range that's ok. It still sounds more than decent to my ears!

And ultimately, what that means is that with all these devices you can track at 16-bit in many cases, save some disk space, and still get very very high quality audio.

If your signal is peaking up in the -12dBFS range when recorded and your recorder can't resolve below about -96dB anyway (not to mention once you add mic self-noise, pre-amp noise, etc.etc.) why track 48 extra bits of data?

I'll stop flogging the horse now.

Very informative - thanks for taking to time to post this!

 

[Petros Kolyvas]

16 bit gives 98dB of dynamic range against I think 108db for 24 bit. Yes I record in 24bit in my sound dubbing suite....but it's not worth 'throwing the toys out of the pram' because it's only 16 bit. The real thing is how quiet are the mic amps, as it's not worth going for ultra dynamic recording if the input signal to noise isn't up to it.

The math for theoretical maximum dynamic range with uncompressed PCM audio is simple: 6dB per bit. 16-bit is 96dB and 24-bit is 144dB as noted by others. Additionally, like a lot of audio related theory, the scale is not linear (ie the returns per-increased bit of resolution are very good if the system is well designed) and power doubles every 3db. But now we're getting into parts unknown and not particularly pertinent to analyzing the potential drawbacks (or conversely benefits) of tracking in 16-bit vs 24-bit. (Note: I'm happy to have any of my knowledge here questioned and challenged. I could be wrong about some of this and I'm open to that - don't believe everything you read on these here internets - even this - aka Wikipedia, or your local sound engineer, is your friend!)

In many ways it can help to think that 24-bit converters are perfect for oversampling... that is to say that even if the system can't really resolve more than 16-bits, using 24-bit converters is essential to getting the best out of the 16-bits you can use. In many cases if you used only 16-bit converters you'd only end up with a 14-15bit useable signal thanks to the shortcomings of the 16-bit converter.

So in the case of the H4n the 24-bit converters aren't wasted, just useful in a way you might not expect at first glance.

And, back to the camera - I've always found Panasonic's audio sub-system quite decent and (since quite clearly I'm a "sound guy") have been more than happy with the result of tracking single system audio into HVX-200s, HPX-500s and the old, but not worn out, DVX-100.

Will a Zaxcom Deva or Aaton CantarX sound better? Sure. Will they make the difference for me (and I'm only talking about my case here), no, the AF100 (I too have a pre-order in) will suit my clients and small-time workflow just fine!

 

[Barry_Green]

That brings up an interesting question. Is it possible to disable the audio in the camera entirely and devote all 24Mbps to picture and get slightly better picture quality?

No, there's no section for disabling it (although, when overcranking, the audio tracks aren't included, so perhaps they are using more space. In fact, I know they are...)

I asked them the question about audio because there are two different types of audio recording in the PH mode, there's uncompressed LPCM and there's compressed AC3 at 384kbps. I asked if the LPCM was taking away video bandwidth and was told "no" -- the video bandwidth is one dedicated path, and the audio adds onto it on its own path, so video quality doesn't suffer by choosing LPCM (and, conversely, video quality doesn't improve by choosing AC3).

 

[Barry_Green]

Just as many cameras can "record" 1080p but can't "resolve" anywhere near 1080 lines of resolution the H4n can record 24bit audio but can barely resolve 16bits of dynamic range.

Thanks for saying it more clearly. This is what I was trying to get at. Just because the recording format has higher specs, doesn't mean that the pre-amps and audio circuitry can take advantage of it. Better pre-amps and hardware, recording to 16-bit, could very well sound better as a final product than worse hardware recording in 24-bit.

Not saying definitively that the AF100 *will* sound better than the Zoom. Just saying that I wouldn't be surprised if it did. I would expect that the audio circuitry in a $5,000 professional camcorder could very well be on par with the circuitry in a dedicated audio recorder that only costs $300. I wouldn't put the AF100 up against an SD744T though!

 

[Barry_Green]

Barry when you get the camera for testing could please do a check on the audio including recording, there was a problem with the HVX 200 where some cameras had a slight beeping durning recording Panasonic was good in fixing it but it is better to "nip it in the bud" and Sony's PD-150 DVCAM had very noise preamps, I am sure that Panasonic's will be good but we would still like to know just in case, Thanks.

Certainly.

But don't worry about the beep -- that was only ever on P2 cameras and turned out to be some sort of shielding issue or something. It was never an issue on the AVCCAM products.

And the PD-150 was the worst. The DVX100 was substantially superior, and all subsequent Panasonics have maintained that higher level. I certainly expect the AF100 to have great audio (for a camcorder) but will attempt to check as well.

 

[Barry_Green]

i second the notion to use better recording practices before counting bits. you could have 64bit/256kh audio hooked up to a $60 shotgun pointed off axsis. the sound it's picking up of a squirrel next to the actor will be lovely, but a lovely fail nonetheless

Agreed 1,000%. In our Sound for Film & Television DVD, we demonstrate conclusively that great technique with a crappy mic, beats out lousy technique with a top-of-the-line mic, every time.

 

[Yak]

Certainly.

But don't worry about the beep -- that was only ever on P2 cameras and turned out to be some sort of shielding issue or something. It was never an issue on the AVCCAM products.

And the PD-150 was the worst. The DVX100 was substantially superior, and all subsequent Panasonics have maintained that higher level. I certainly expect the AF100 to have great audio (for a camcorder) but will attempt to check as well.

Thanks ones again for all your work and endless generosity.

 

[Darren Levine]

Agreed 1,000%. In our Sound for Film & Television DVD, we demonstrate conclusively that great technique with a crappy mic, beats out lousy technique with a top-of-the-line mic, every time.

i think the line that causes my brain the most damage, just after "leave it to post" is: " have that PA operate boom":furious3:

 

[Barry_Green]

Drives me batty. Would they say "hey, you, P.A. -- will you operate the camera for this take?" Suggest it, and people would be horrified -- but they have no problem handing the boom to the craft services person or the stills photographer or a p.a... sigh.

 

[Noel Evans]

Heres what I can say from my experience. I did a number of tests with the zoom H4N when I first got it, alongside direct recording to the HPX500. I had read at the time that the H4N couldnt fully resolve 24bits. I used the Sanken CS3e.

Take it however you like, but my tests concluded for me that the H4N at 24 bit was still flatter than the 500 at 16bit. I then changed the H4n to 16 bit and there was 0 difference. The result of this test is that I have never recorded audio at 24bit on the H4N.

I know nothing of the quality of the recording components in the AF100, but correct me if I am wrong Jan / Barry, but I have found that Panasonic uses similar components in all of its cameras. Go back to the original DVX and look at the audio control switches etc which are connected to the audio board inside - almost 100% exactly the same as the AF100 and for that matter the Panasonic HPX2700/3700.

As a side note I test every piece of equipment I get to check performance for personal reference. My results may vary, and are only what I concluded - they are not written in stone.

 

[DrDave]

Looks like I'll be hanging onto my Zoom H4N,

Why haven't Panasonic included 24bit/96Khz audio recording with the AF-101 ?

Would it really have cost that much more to include ?

Didn't think 16Bit was being implemented anymore.

I'll still be getting one but thought I'd bring this up for discussion as it doesn't seem anyone's mentioned it yet (unless I'm blind?) . . . .

Basically, ppl who record the audio "in camera" are not that picky about the sound.

 

[Andrew Stone]

It's pretty much all been said and there is nothing stopping you from using an outboard recorder. In point of fact in the video industry, generally once you get to a certain level, you hire a "Sound recordist that comes on a shoot or on the set with their own recorder, mics, poles and bag of tricks. 16/48, as has been said numerous times is and has been the de facto standard in the video world for a long time, not the least of which is consistency in the editing bay and for post sound people & sound designers, so when stuff comes in they expect it to be in 16/48. You start throwing in different sample rates and it can catch someone off guard. Ever try doing a midi sequence at 96K when the some of the existing sound material is at 48K? Standards are there for more than just sound quality, a lot of it has to do with workflow.