Confusing record length on 16GB cards

W

WpgCameraMan

Well-known member
I bought two new Patriot 16BG Class 6 cards yesterday, decided to give them a once-through on the 150.. recording PH 1080/60i, set to colour bars & tone, display started at 99 mins space left.

I did a few start and stops along the way, but it was always the same thing, B&T.

Card A now has 193 minutes on it (confirmed in playback mode, and via media player) .. that is 3h15m.

Card B has 237 minutes (3h57m), again, confirmed.

I was under the impression that a 16BG card would hold 99 minutes, but I've never filled a card and added the clips up before, as I shoot in short bursts, could have 200 clips on a card.. So this made me wonder, is 99 a fixed time, or a worst-case ?

And the difference in times is also puzzling. Card A has 11 clips, card B has 5 clips, yet there is a 42 minute difference in length while recording the same image.

Both cards report the exact same size in bytes after a format.

Is this normal, maybe I've been getting +99 mins in the past and never noticed?
 
YES.. you were right. Should be 99 minutes at full resolution. I bet you have it on one of the lower quality settings. I get about 50 minutes for the 8 gigs and 99/98 for the 16 gigers...
 
Did you record just color bars & tone the entire time? If that's the case my guess is that the compressing algorithm is doing it's job, by noticing that your picture isn't changing much (at all, actually) from frame to frame and therefore isn't bothering to explicitly re-record sections that don't change, just every so often for keyframes.

That's why, in theory at least, you'll get slightly better record times shooting tripod-based interviews (with very little changing in any given frame) vs action sports/handheld with these fancy AVC compression algorithms.

That said, in my experience the 99 minutes is pretty dead-on accurate. I've noticed very little fluctuation from that.
 
I thought the compression would be the culprit here. And of course, bars don't change. But the variation in record lengths is a little puzzling (unless card A with twice as many files somehow ate up space elsewhere in the PRIVATE folder).

Was running off AC for this test as well.

I should aim this at something complex in the morning and see if I can get less than 99 mins.
 
Ran another test today. I recorded outside my kitchen window to a tarp that was flapping in the wind, I got 99 minutes, exactly as expected.

Next test, recorded colour bars, one clip, no stops.. 4:00:27 .. 4 hours.. 99 minutes versus 240 minutes.

Real-world shooting would never allow for this sort of compression. Still neat though.
 
Yes, 99 min. is usually dead on when shooting normal scenes. But it is funny to see the result with color bars.
 
I would expect better results, after all, there is nothing changing in the scene, NOTHING. In theory, one frame should be enough. Apparently, the minimum possible bitrate is not very low.
 
I can't think of anything I've got that shows bitrate of a file... Another test I'll try is recording with the lens cap on, iris closed, so there's still an 'analog' signal being processed.
 
Another test to try is a locked down shot of a well-lit scene on a tripod where nothing changes. I bet you may get more than 99 minutes.
 
Cranky said:
I would expect better results, after all, there is nothing changing in the scene, NOTHING. In theory, one frame should be enough. Apparently, the minimum possible bitrate is not very low.
Click to expand...

There's audio, which is at a fixed bitrate, and the data stream still needs some info... and I don't know that there's such a thing as "infinite GOP" - you probably still need at least a full reference frame in every GOP...
 
You must log in or register to reply here.
Back
Top