The scripts:
Code: Select all
AVIsource("C:\Documents and Settings\Jeff\Desktop\1\Source1.avi")
ConvertToYV12(interlaced=true)
TomsMoComp(1,5,1)
bicubicresize (512,384)
deen ("a3d",3,3,5,4)
fastlinedarken()
awarpsharp()
Tweak(cont=1.0,sat=1.5,bright=0, hue=0)
ConvertToRGB32()
Code: Select all
AVIsource("C:\Documents and Settings\Jeff\Desktop\1\Overture.avi")
ConvertToYV12()
crop (0,60,-0,-60)
bicubicresize (512,384)
deen ("a3d",3,3,5,4)
fastlinedarken()
awarpsharp()
ConvertToRGB32()
The resulting Huffys
Made with Fast Recompress in VirtualDubMod
Frame size, fps: 512x384, 29.970 fps
Length: 134121 frames (1:14:35.17)
Size: 28,980,011,008 bytes
FourCC code: FourCC: [HYFU] Huffman Lossless
Decompressor: Huffyuv v2.1.1 CCESP Patch v0.2.5
Number of key frames: 134121
Min/Max/Avg/total key frame size: 151348/216056/290860 (28298534k)
No delta frames
Data rate: 51802 kbps (0.01% overhead)
No audio stream
-------------
Frame size, fps: 512x384, 23.975 fps
Length: 120476 frames (1:23:44.98)
Size: 27,384,602,624 bytes
FourCC code: FourCC: [HYFU] Huffman Lossless
Decompressor: Huffyuv v2.1.1 CCESP Patch v0.2.5
Number of key frames: 120476
Min/Max/Avg/total key frame size: 147568/219260/439180 (25796526k)
No delta frames
Data rate: 42055 kbps (0.01% overhead)
Audio Stream
Sampling rate: 48000Hz
Channels: 2 (Stereo)
Sample Precision: 16 bit
Audio tag: Tag 0x0001, not decodable. Microsoft PCM Format
Compression: PCM (Uncompressed)
Layout: 50 chunks (1630|0.50s preload)
Length: 241198848 samples (1:23:44.97)
Min/avg/max/total frame size: 8008/9768/96000 (477k)
Data rate: 1536 kbps (0.25% overhead)
Now the story about the audio stream...
I got rid of it entirely in one source, but I wanted a small 5 second or so audio clip from the second source so I left it in. It was suggested to me that I should extract the audio from the source, compress it as per EADFAG, import it into my project seperately, and use KillAudio() in my script, so I attempted to do so. And got a 14mb or so mp3, but unfortunately I started editing my timeline before I knew anything about AVIsynth and was editing the mjpeg that had the sound file. When I tried swapping in the Huffy with the KillAudio() filter, Premiere complained that the filetypes of the mjpeg and the Huffy don't match and wouldn't let me import the Huffy.
Would it be advisable to use the audio clip straight from the source? Or keep the seperate mp3 I created. But unless there is a way of getting around this I don't think I can kill the sound.
Put it all together and what have you got?
Frame size, fps: 512x384, 29.970 fps
Length: 10956 frames (6:05.56)
Size: 2,407,497,728 bytes
FourCC code: FourCC: [HYFU] Huffman Lossless
Decompressor: Huffyuv v2.1.1 CCESP Patch v0.2.5
Number of key frames: 10956
Min/Max/Avg/total key frame size: 147460/213824/362368 (2287758k)
No delta frames
Data rate: 51267 kbps (0.01% overhead)
Audio Stream
Sampling rate: 44100Hz
Channels: 2 (Stereo)
Sample Precision: 16 bit
Audio tag: Tag 0x0001, not decodable. Microsoft PCM Format
Compression: PCM (Uncompressed)
Layout: 1 chunks (0|0.00s preload)
Length: 16121416 samples (6:05.56)
Min/avg/max/total frame size: 170576/170576/170576 (167k)
Data rate: 1411 kbps (0.01% overhead)
First thing I want to say is that though looking at the two sources, the framerates are different, they were edited in Premiere as such and there is absolutely no timing issues with my video (anymore, thanks to those who helped me in my previous thread.)
Up to this point though, I think everything is normal.
XviD compression settings:
Any omissions of a checkbox option indicates that option is unchecked
Profile @ Level: Simple @ L5
Zones: Frame 0, Weight 1
Quantization Type: H.263
B-VOPs:
Max Consecutive BVOPs: 2
Quantizer Ratio: 1.5
Quantizer Offset: 1.0
Level:
Suggested: 720x576x30
Max frame size (macroblocks): 1620
Max processing rate (mb/s): 48600
Video Buffer Verifier:
Max buffer size (bits): 1833216
Max bitrate (kbps): 8000
Pixel Aspect Ratio: Square pixels
Full Quality First Pass
FourCC used: DX50
Motion:
Motion Search Precision: 6
VHQ Mode: 4
Use Chroma Motion
Frame Drop Ratio: 0
Maximum I-Frame interval: 300
Quantization:
Min I-Frame Quantizer: 2
Max I-Frame Quantizer: 31
Min P-Frame Quantizer: 2
Max P-Frame Quantizer: 31
Min B-Frame Quantizer: 2
Max B-Frame Quantizer: 31
Trellis Quantization
The Output
Fast recompress
I look at that and I see a file of about 12.5mb when it's 10% finished. So I think "Great, my first pass is going to be about 130mb, shouldn't be too hard to get it to 95mb or so."
Now, when the dub finished, it SAID 130 or so mb in the XviD screen, but what I got was a bit different...
The First Pass
Frame size, fps: 512x384, 29.970 fps
Length: 10956 frames (6:05.56)
Size: 185,753,600 bytes
FourCC code: FourCC: [DX50] DivX5
Decompressor: XviD MPEG-4 Codec
Number of key frames: 566
Min/Max/Avg/total key frame size: 2164/25235/63471 (13949k)
Min/Max/Avg/total delta frame size: 7/10244/63674 (103946k)
Data rate: 2642 kbps (0.22% overhead)
Audio Stream
Sampling rate: 44100Hz
Channels: 2 (Stereo)
Sample Precision: 16 bit
Audio tag: Tag 0x0001, not decodable. Microsoft PCM Format
Compression: PCM (Uncompressed)
Layout: 1 chunks (22050|0.50s preload)
Length: 16121416 samples (6:05.56)
Min/avg/max/total frame size: 88200/88200/88200 (87k)
Data rate: 1411 kbps (0.03% overhead)
Now I was starting to get a little bit anxious. Things didn't look too promising but I decided to use the XviD calculator and try for a second pass.
95 x 1028 = 97660, so that's what I entered as my target size.
The other information in the calculator was:
Subtitles (kbyes): 0
Format: AVI-OpenDML
Overhead (kbytes): 1918
6 mins, 5 secs, 29.97 FPS (NTSC)
Audo format of MP3-CBR, size of 13778 kbytes
The values it gave me was:
Size(kbytes): 83696
Average bitrate (kbps): 1878
So I entered the target filesize as 83696 and prepared my second pass. Here is how it looked.
(Again, quick recompress in VirtualDubMod)
I-Frame boost (%): 10
I frames closer than... (frames): 1
Are reduced by (%): 20
Overflow Control Strength (%): 20
Max Overflow Improvement (%): 5
Max Overflow Degradation (%): 5
High Bitrate Scenes Degredation (%): 0
High Bitrate Scenes Improvement (%): 0
The final filesize in the XviD console indicated about 83mb. But of course things aren't always as they seem.
Frame size, fps: 512x384, 29.970 fps
Length: 10956 frames (6:05.56)
Size: 150,618,112 bytes
FourCC code: FourCC: [DX50] DivX5
Decompressor: XviD MPEG-4 Codec
Number of key frames: 566
Min/Max/Avg/total key frame size: 2164/21878/63471 (12094k)
Min/Max/Avg/total delta frame size: 7/7045/52708 (71489k)
Data rate: 1873 kbps (0.31% overhead)
Audio Stream
Sampling rate: 44100Hz
Channels: 2 (Stereo)
Sample Precision: 16 bit
Audio tag: Tag 0x0001, not decodable. Microsoft PCM Format
Compression: PCM (Uncompressed)
Layout: 1 chunks (22050|0.50s preload)
Length: 16121416 samples (6:05.56)
Min/avg/max/total frame size: 88200/88200/88200 (87k)
Data rate: 1411 kbps (0.03% overhead)
That's about all there is.
I'm very sorry if this is too much info, but I'm really at my wits end here and I have no idea what to do.
I saw in the example in EADFAG that the bitrate there was 780kbps so I tried setting that as my goal, but it ended up looking like what EADFAG said it shouldn't (lots of frames with low quantizers.)
I just want to get this video over with and put it behind me so I can get a better start on my next video. I just didn't want to have to redo my entire timeline with this video after I figured out AVIsynth.