audio bit depth question for dual sound audio

D

davidp158

Well-known member
I'd like to know what is the best practice for setting up my cameras and audio recorder so they match up in post (Adobe Premiere Pro).

I will be shooting a band with a GH2 and GH3. I will be recording 4 audio tracks with a Tascam DR680 (stereo mix off board, plus two room mics). I will only use the camera audio for sync reference via Plural Eyes.

For proper sync in post with minimal drift, I realize both the cameras and recorder must record at the same sample rate. The GH2 and GH3 audio files appear in Premiere Pro as 48kHz compressed audio, so I intend to record the dual sound audio at 48kHz. However, the Tascam has the option of recording in 16 or 24 bit, but the camera manuals don't state what bit depth the cameras record in.

Any suggestions would be appreciated.
 
Last edited:
davidp158 said:
I'd like to know what is the best practice for setting up my cameras and audio recorder so they match up in post (Adobe Premiere Pro).

I will be shooting a band with a GH2 and GH3. I will be recording 4 audio tracks with a Tascam DR280 (stereo mix off board, plus two room mics). I will only use the camera audio for sync reference via Plural Eyes.

For proper sync in post with minimal drift, I realize both the cameras and recorder must record at the same sample rate. The GH2 and GH3 audio files appear in Premiere Pro as 48kHz compressed audio, so I intend to record the dual sound audio at 48kHz. However, the Tascam has the option of recording in 16 or 24 bit, but the camera manuals don't state what bit depth the cameras record in.

Any suggestions would be appreciated.
Click to expand...

I would recommend using 24-bit recording. I use the old fashioned method of slating and then syncing the second sound recording manually. I do use the camera's audio as a guide track, but always record in 24-bit mode on my audio recorder.

I have a H4 for mobile recording, and have several USB type recorders for more fixed setups. In any case, all are 48k and 24-bits, no compression.
 
All video cameras that I have ever seen use 16-bit audio.
As you say, it is ideal that you can use the same sample-rate for your audio recorder as the cameras are using (48K by default)

HOWEVER! The bit-depth of the cameras OF NO IMPORTANCE if you are using them only for reference (whether manually, or with the help of software such as PluralEyes).

Whenever given the option, I ALWAYS use the larger bit-depth (24-bit) for LIVE recording because it gives you GREATER DYNAMIC RANGE, which will make it easier to get a good recording of live music. A wider dynamic range allows you to record at decent levels to avoid being "down in the mud" of the noise floor. But the 24-bits allows you to have more HEADROOM to accomodate volume peaks which will ALWAYS happen in a live performance. Of course, you should attempt to get the performers to play at max levels and set your recorder to peak at -9 to -12dB.
 
Richard - Thanks for the prompt and detailed reply. People like you make this forum a valuable resource.

I (sort of) thought that the camera audio could be a different bit depth, but I wanted to confirm so that I didn't run into sync problems in post.

Yes, live performances tend to have drastic dynamic range, and the 24bit option is what I intend to record with on the Tascam.

Thanks again for your help!
 
As said most TV and cameras are 16 bit 48k but if you have the option record at 24 bit 48k as it will give you more headroom, do not use 44.1 or 96k as you will just have to change it in the edit for most video work,
 
Richard Crowley said:
But the 24-bits allows you to have more HEADROOM to accomodate volume peaks which will ALWAYS happen in a live performance. Of course, you should attempt to get the performers to play at max levels and set your recorder to peak at -9 to -12dB.
Click to expand...
Make sure, as described above, that you set your recorder to max -9dB. That way, if it "peaks" it gives you extra dBs in post to adjusts. That, and by recording a little lower, you have the possibility to "normalize" in post to "raise" the levels of the audio track.
j1clark@ucsd.edu said:
I would recommend using 24-bit recording. I use the old fashioned method of slating and then syncing the second sound recording manually. I do use the camera's audio as a guide track, but always record in 24-bit mode on my audio recorder.
Click to expand...
Sounds familiar! ;) With the slate mark, it's easy to sync the audio.
 
GaryNattras and drapeama - Thanks for the comments. I will record in 48kHz/24bit on my Tascam DR680.

My primary concern was if the 16 bit depth of audio from the cameras would introduce any sync issues in post with Plural Eyes. I'm recording a live band performance, and a slate isn't a practical option, as roving cameras will be starting and stopping. Plural Eyes quickly syncs multiple cameras and to my continuous audio master tracks, and arranges everything into chronological order.
 
davidp158 said:
GaryNattras and drapeama - Thanks for the comments. I will record in 48kHz/24bit on my Tascam DR680.

My primary concern was if the 16 bit depth of audio from the cameras would introduce any sync issues in post with Plural Eyes. I'm recording a live band performance, and a slate isn't a practical option, as roving cameras will be starting and stopping. Plural Eyes quickly syncs multiple cameras and to my continuous audio master tracks, and arranges everything into chronological order.
Click to expand...

The 'random' thought I have about the camera guide track, and 'roving' cameras, is if the volume levels are such that the camera recording could be either clipped or 'buried' in ambient 'noise', such that it would be difficult to sync up by automagic means.
 
Premiere is very happy dealing with differing audio formats. You can mix 16, 24 and 32 Bit floating point audio in the same timeline, and even crossfade them if you wish. Can't prove it, but I suspect that Premiere treats them all as 32bit floating point files - which is what I believe is it's standard bit depth for editing audio. I have not seen any sync errors caused by bit depth, only bit rate.
 
Final Cut Pro does the same and with a 32 bit floating point you can chuck pretty much anything in it, I actually use FCP as my main audio editing software as I find it very similar to the old AMS AudioFiles I used in the past.
 
j1clark@ucsd.edu said:
The 'random' thought I have about the camera guide track, and 'roving' cameras, is if the volume levels are such that the camera recording could be either clipped or 'buried' in ambient 'noise', such that it would be difficult to sync up by automagic means.
Click to expand...
But unless the entire track is clipped hard which seems very unlikely, there is enough information there to do a rough correlation of the data-stream which is all you need for that kind of accuracy. And it is rarely clipped hard right in the first few seconds where the critical reference sync is established.
 
For what it's worth, the 50 Mbps MOV format on the GH3 records audio as uncompressed LPCM format. I have no idea if this helps in maintaining sync with an external recorder, but it does help when you want to actually use the camera audio.
 
Yes, with faster chips and bigger storage media, more video codecs feature uncompressed sampling as an option if not the default.
Of course, uncompressed is vastly preferable to any compressed codec for many reasons, sync being only one.
And certainly recording everything at the same sample rate is preferable to re-sampling in post production. All modern video codecs use 48K sample rate.
 
davidp158 said:
...
For proper sync in post with minimal drift, I realize both the cameras and recorder must record at the same sample rate. The GH2 and GH3 audio files appear in Premiere Pro as 48kHz compressed audio, so I intend to record the dual sound audio at 48kHz. However, the Tascam has the option of recording in 16 or 24 bit, but the camera manuals don't state what bit depth the cameras record in.

Any suggestions would be appreciated.
Click to expand...

I don't think drift has much to do with sample rate mismatch these days. The problem is due to small variations in the digital clocks of various devices and often seems to be a matter of luck.

Fran
 
Different bit bit depths would not affect sync in any way. Different sample rates could however. Some NLEs support different/multible sample rates on a project.
 
The biggest issues from dealing with different sample rates is the conversion which is always required at some point. Wither invisible and "automatic" or manual, it takes more time and represents potential trauma to the recording. Best to avoid it if possible.

Sync problems only come from equipment clock speed errors, not from your choice of sampling rate or depth.
 
You must log in or register to reply here.
Back
Top