Naked Science Forum
Non Life Sciences => Technology => Topic started by: engrByDayPianstByNight on 25/08/2011 03:58:06
-
I have a Sony Cybershot DSC-T700 digital camera. As an amateur pianist, I often use its video-recording function to record me playing the piano. In a piano piece, you play different dynamics at different passages, sometimes pianissimo, sometimes a little more forte, that sort of things.
I'd like to think I do a fairly good job playing these different dynamics. However, after recording a piece and playing it back, I often find that the recorded audio does not faithfully capture the various dynamics.
It could certainly be that I'm not playing the dynamics well enough, but I'm also wondering whether it could be that the audio recorder in the camera is too rudimentary to do that. Anybody have any ideas?
And, what about a digital camcorder? Would it have a better audio-recording system built in to do a better job of capturing the dynamics?
Thanks.
-
I don't know about the specific camera you have, but I suspect it is trying to be too clever! It is probably automatically adjusting the audio recording level. You don't want it to do that because it will wipe out the dynamic range of your recording.
What you need is an audio recording device that allows you to manually set the recording level and leave it fixed. In days of yore, tape recorders had dials and some sort of VU meter for that purpose.
I'm sure there is someone on TNS with much more up to date knowledge on this subject [:D]
-
I think that's probably the case. Is there any mention of "automatic gain control" or AGC in the settings/manual?
-
I’d agree about AGC being the cause.
IF the AGC can't be switched off it is possible to expand the dynamic range (http://en.wikipedia.org/wiki/Dynamic_range_compression#Types) of the sound recording, (e.g. make quiet bits quieter),
using DAW (http://en.wikipedia.org/wiki/Digital_audio_workstation#Free_and_open_source_software) software (Audacity (http://en.wikipedia.org/wiki/Audacity) is free) which will reverse the AGC compression effect.
-
I believe, as Ben suggests above, those cameras have in-built automatic gain controls (AGCs), which are normally there to achieve a consistent sound level. This is because the majority of what will be recorded on video cameras is speech, meaning the subjects are going to be at variable distances from the camera (and hence microphone) and speaking at different levels of loudness. Consequently, to prevent clipping (digital overshoot), which sounds very unpleasant on a recording, or inaudible audio, the equipment has a compressor limiter and an average gain function circuit.
This looks at the average energy density of the incoming sound and turns itself up and down accordingly. It's rather like people scurrying across the deck to opposite side of a boat to balance things out every time they see a big wave or a gust of wind coming. The problem is that when you actually "want" the sound to vary for artistic effect, as in your case, this system works very hard to reverse all of the dynamics that you introduce!
It's also very destructive in soundscapes where there might be a lot of bass - because the low notes pack a big energy punch and cause the gain control to crush the signal, punching a hole through the higher frequency part of the spectrum. This is why most high-end equipment these days uses multi-band compression to get around this.
Our solution, in house, is to shut off the bloody AGC, record at a sufficiently low level to ensure that you have plenty of "headroom" - ie during the loudest parts it never clips, and then you can compress and amp up the signal later in post production.
I'd check the instructions and see if you can disable the AGC; it will be very hard - and laborious - to reverse the AGC effects in post production.
-
Thanks guys for all your informative replies. I looked up the camera's manual, and there's no mention of AGC, nor are there any settings hat let you tweak any video-recording parameters except resolution. I can only guess that the AGC is built in, and they don't want users to temper with it. Anyways, I'll try with Audacity. Thanks again. [;D]
-
On some cameras you can set the recording level "manually" as opposed to "automatic" (which is really AGC). This is usually found on the better cameras (like my Canon G9) - but of course (as Chris says) the user is then responsible for setting the level low enough that the loud bits don't distort and sound ghastly.
-
It's slightly ironic that manual recording level control is now considered a "feature" on high end equipment.
My old Grundig reel-to-reel had an high-tek device called a "magic eye" that showed you the recording level. It was supposed to be better than a moving coil VU meter because it had a very fast response, however, perhaps the best thing about it was that it looked really cool [:D]
-
If the problem is not due to the AGC then it is likely to be due to the recording format your camera uses. It is probably over compressing the dynamics. This is common in commercially available music cds. There is software available to increase the dynamic range but I don't know if it is available for video.
For audio I use MP3 Gain an excellent small free program that will do just what you want. It's also reversible.
-
Cameras are for recording video. The audio is a secondary requirement therefore will be heavily compressed as spoken word is what most people expect on a video camera.
I suspect this is where you are loosing the dynamics.
Why not buy a professional audio recorder and re-synch with the video using a PC later.