MPEG File size problem

[aladdin1001]

Hi,
I've got Canopus Edius and ProCoder. I use my ATI All-in-Wonder to capture short MPEG video which are around 250MB in size. i capture the mpeg file as follows
constant 8000kbps for video and 48khz mpeg-1 224kbps Stereo audio

I open the files in Edius and change a few things but when i export I specify the same settings as above to minimize quality degredation and at the same resolution which is 720x576 25fps

The video file that is output is double the original roughly 550MB... Can anyone explain why this is happening? and point out what I could do to fix this?

thanks in advance

[manono]

Hi-

Perhaps it's converting the audio to PCM WAV? You can check in GSpot.

[aladdin1001]

No i've checked the final audio size is around 13MB. I know because it exports two mpv and mpa i think, i'm not sure about the exact extensions but they total 550MB.

[redwudz]

MPV is a video file, MPA is the audio file. You have two video files? Are they the same? That would explain the size doubling.

I would check them all in Gspot 2.60 as mentioned, though, and see what their specifications are.

[aladdin1001]

There is only one video file and one audio file. Sorry for the confusion i couldn't remember whether it was mpv or mpa but the audio file is around 13MB and the video is 437MB.

[manono]

Then the bitrate among the 2 videos is different. You can check by running both through the preview in DGIndex. It may take a few minutes, depending on the number of frames, but it's absolutely accurate. For a real rough idea of the relative bitrates, try BitRate Viewer. I believe newer versions of GSpot can also give the MPEG-2 bitrates.

I guess one or the other of the encoders didn't give you the full 8000 for one reason or another. A bitrate calculator can give you the size you're supposed to get based on length and bitrate, and tell you which one messed up. I'd guess the ATI card didn't give you what you wanted.

...but the audio file is around 13MB and the video is 437MB.

Which doesn't really add up to the 550 MB you said, but still a great deal larger than the original 250 MB (if that's what it was)

[jagabo]

Maybe the source is 8000 kbps VBR with a max of 8000 but an average around 4000. Where as the reencoded file is 8000 kbps CBR or VBR with an average of 8000. And use, GSpot will tell you the average bitrate. So will VirtualDubMPEG2 and VirtualDubMod.

[aladdin1001]

Fantastic. Thank you very much indeed for all your help. It turned out that for some reason ATI didn't record it in 8mbps as specified in the DVD profile. I double checked it and the settings are variable bitrate at 8mbps and a target bit rate of 6mbps which was recognised by an older version of gspot that i had as 8mbps. however, when previewed in DGIndex the average didn't go above 4.5mbps and when i updated gspot to v2.6 it scanned the file and found it avaraged 4113kbps. I'll set procoder to a max of 8mbps and a target of 6mbps and see what it comes out with. As for the file size, i'm not exactly sure about the size because i actually deleted the file after trying to create it twice. for some reason i'd thought it was a mistake with my settings and deleted the file in frustration so it's only an estimate but i remember thinking it was double the size and the audio being 13mb.

Once again, thank you very much for all your help. Once final thing though, does the 4.5 average bitrate mean that it didn't require more than that to encode the file based on the video being recorded or is the ATI TV program playing up? because I intend to burn the edited clips onto DVD eventually

[manono]

Hi-

Once final thing though, does the 4.5 average bitrate mean that it didn't require more than that to encode the file based on the video being recorded

That's a little hard to tell based on what we know so far, although it is possible. I would run the original file through the free version of BitRate Viewer and check the Q-Level. If it's low (maybe 2 or under), then the quality is pretty much maxed out. 2 to maybe 5 would be nice. Over 7 or so (depending on several factors), it could have used a higher bitrate. In any event, there's not much point in reencoding at a higher bitrate, as you're not improving on the original quality, no matter how many bits you throw at it.

[omega6666]

Over 7 or so (depending on several factors), it could have used a higher bitrate.

I was wondering; would you agree that a Q-level over 7 in a complex scene (f.i. explosions) is less problematic than a Q-level over 7 in a static scene?

I'm saying this because I had a problem encoding LOTR ROTK extd, which gave annoying mosquito noise during static scenes, but not (visibly) in the more complex scenes, even though the quality level was practically constant at around 8. It had an average bitrate of 3650, and I used this matrix, which is a variation of the Fox Home Entertainment-matrix used in the original;

08 08 09 11 13 13 14 17
08 08 11 12 13 14 17 18
09 11 13 13 14 17 17 16
11 11 13 13 13 17 18 20
11 13 13 13 16 17 20 24
13 13 13 16 17 20 24 29
13 12 13 17 19 23 28 34
12 13 17 19 23 28 34 41

16 16 16 16 16 16 16 16
16 16 16 16 16 16 16 16
16 16 16 16 16 16 16 16
16 16 16 16 16 16 16 16
16 16 16 16 16 16 16 16
16 16 16 16 16 16 16 16
16 16 16 16 16 16 16 16
16 16 16 16 16 16 16 16

Only when I changed the Qual_Prec downwards and the VBR_Bias upwards (in DVD-RB), was I able to remove this problem, which resulted in a Q-level of around 7 during the (few) static scenes and around 8,5 during the (more occuring) complex scenes.

[manono]

Hi-

I was wondering; would you agree that a Q-level over 7 in a complex scene (f.i. explosions) is less problematic than a Q-level over 7 in a static scene?

Although I try and do my encoding for true VBR, meaning that both static and complex scenes have about the same average quant, in general, yes, under the theory that the eye doesn't really notice the lower quality of the more complex scenes. A long and difficult to encode movie like that one sometimes requires radical measures to have it look decent. Just my opinion.

I haven't done encoded that film, but with other difficult to encode films, I've encoded the end credits separately, using a very low bitrate matrix, to give them a much smaller size than they would have otherwise. I then joined them to the main movie during authoring. With long credits (my last one was 9 minutes, yours much longer, if you kept them all), this can free up many bits for the main movie.