AVS Forum banner
Status
Not open for further replies.
1 - 6 of 6 Posts

·
Registered
Joined
·
3,557 Posts
Discussion Starter · #1 ·
OK, I've had my E80H for about a month now, and love it. I think I've read every post in this forum, but I don't get how it determines what the best bit rate is.


If I chose XP will it try to average the 4.7GB across 60 minutes of recording time? If so, will it actually waste space on static/slow moving images because an average threshold that it needs to maintain?


For example, let's say I'm recording a 60 minute show that starts with black and white still shots, and then switches to color and action at the half-way point. Will the total program end up being 4.7GB and take up a full DVD? Is the average bit rate for the first half of the disk going to be the same as the second half? Or will the disk only be partially filled? Will the recorder calculate unused data from the first half of the disk and use it to make adjustments as it continues recording?


And how would this compare to FR mode? The recorder can't know ahead of time how much data it will need to record a show in a certain time period, it can only know the average bit rate available over the recorded length.


How would two recordings compare that were set to XP mode: one programmed for an hour of recording, the second just set to record, with no set ending time? Would the second look better because it wouldn't have any preset recording length?


Maybe I'm just thinking this out too much, but I'm really curious.


Scott
 

·
Banned
Joined
·
3,539 Posts
In XP mode, one hour of video to fill a 4.7GB DVD is using the maximum bitrate allowed for DVD (close to 10Mb/sec) constant. It can't increase the bitrate beyond this and to lower the bitrate (variable bitrate) for some sections would result in more than an hour of viceo on the DVD - thats other modes. SP and slower modes could use variable bitrate to get an average bitrate need for 2+ hours on the DVD and use higher bitrate for fast motion scenes, but thats difficult to do in realtime it works best with offline two-pass encoding schemes, thats what they do for DVD movie releases.
 

·
Banned
Joined
·
3,539 Posts
It doesn't need to be VBR at one hour / DVD speed since thats what you get with constant maximum (9.8Mbit/sec) bitrate. VBR is only needed to get an average bitrate lower than the maximum yet use up to the maximum for fast motion scenes. If you can use constant maximum bitrate for the entire video - why would you ever want to vary it?
 

·
Registered
Joined
·
3,700 Posts
Quote:
It doesn't need to be VBR at one hour / DVD speed since thats what you get with constant maximum (9.8Mbit/sec) bitrate. VBR is only needed to get an average bitrate lower than the maximum yet use up to the maximum for fast motion scenes. If you can use constant maximum bitrate for the entire video - why would you ever want to vary it?
The fact is if you use the status feature on the DMR recorders to view the bitrate in real time or during playback, it varies anywhere from 4 to 10+ Mbps. The reason Panasonic does this is anyone's guess, but it is probably due to the fact that Panasonic uses a specific, efficient VBR MPEG2 encoding algorithm that is common across all recording modes with only the avg bitrate and resolution setpoints changed depending on the recording mode. To implement fixed bitrate for XP would require a different algorithm. The fact is that even with XP mode and a high avg bitrate, Pansonic has enough "headroom" left on the disk to support LPCM audio recording.


Vic
 

·
Registered
Joined
·
2,354 Posts
It does the math. Assuming CBR (for simplicity and for this example), and knowing that a DVD-R will hold 2 hours at SP, it picks the bitrate that will fill the disc after two hours of recording. It's that simple.


FR? It still just does the math. You want to record 102 minutes on a disc? It again picks the bitrate that will fill the disc, this time after recording for 102 minutes.


Now, VBR. First of all, VBR is not extensive, although one would think so by watching the wildly swinging meter on a Panny. It is difficult for consumer-level decoders to track VBR-encoded files accurately on playback, which is why Panasonic PVR's limit their VBR ceiling to about a 2% swing. And at some point it aggravates lipsync. DVDR's certainly have the same limitations.


VBR by itself makes sense if you are compressing an AIFF into an MP3, (where lipsync isn't an issue). When less bits are needed, less are used, making a smaller file. But a smaller file doesn't benefit you in DVDR, because that only means that the disc will not fill up quite as much as it might under CBR, and having an extra un-filled minute or two is useless.


The only way VBR can benefit you is if there is a buffer, that can allow more bits during scenes that need them and restrict bits during scenes that don't, sort of robbing Peter to pay Paul, and thereby give you an average bit rate that is equivalent to CBR but with higher quality. If low-impact frames are encoded and as they go by the extra bits build up in a buffer, these can be used to increase the momentary burst bit rate during high-impact frames without causing a rise in the average bit rate, reducing bit-starving in the high-impact frames to give an aggregate increase in quality. Of course if enough high-impact frames occur in a row, the benefit is lost for some frames.


So, VBR is helpful, but limited on purpose, and not dramatically effective. If it were extensively used without buffering, it would be difficult for the DVDR to predict the average bit rate and discs might also have extra useless space left, or even run out early. Since VBR combined with buffering is very likely the approach used in consumer DVDR's, the average bit rate (which is the really important figure regarding record time) under VBR will normally turn out to be very similar to CBR, so those issues aren't very likely a problem.
 
1 - 6 of 6 Posts
Status
Not open for further replies.
Top