I talked about the size.
the both files were captured by the same software (gb-pvr).
the size of each video as mpeg-2 (before the conversion to xvid) is about 28mb. the length of each video is 1 minute, and they are both recordings of the same tv show. after the deinterlacing and conversion to Xvid in VirtualDub (by the same settings exactly), the output Xvid files, are different in their size: one video size ia a 16mb file, and the other one is about 20.8mb. what causes that difference between the both videos? I didn't change the settings at all.
Try StreamFab Downloader and download from Netflix, Amazon, Youtube! Or Try DVDFab and copy Blu-rays! or rip iTunes movies!
+ Reply to Thread
Results 61 to 90 of 116
Thread
-
-
I already explained this. The source is different.
If you used 1pass quantizer mode, it will give the same "quality" level for each frame, so the size will vary depending on the source (e.g. complex scenes will need more bitrate to keep it the same level, therefore bigger size. For example, a simple cartoon will be much smaller because there is less detail, and it requires less bitrate to achieve the same "quality").
If you use 2pass mode you can specify a bitrate and therefore ending filesize. -
When encoding with Xvid you have two choices:
1) Pick a file size (by selecting a bitrate). You know now big the file will be but you don't know the quality.
2) Pick a quality (by selecting the target quantizer). You know what the quality will be but you don't know the file size.
Clean video, target quantizer=4: clean.avi
Noisy video, target quantizer=4: noise.avi
Noisy video, 2-pass, 1400 kbps: 2pass.avi
In the first two I specified single pass target quantizer=4. Note the extreme difference in size/bitrate of the resulting files. The clean video only required 400 kbps. The noisy video required about 8000 kbps. In the third video I used the noisy source and specified 2-pass 400 kbps, trying to match the clean video's bitrate. Xvid couldn't get the bitrate that low and the file ended up around 1400 kbps. Notice how poor the video turned out. I did a fourth test where I specified 2-pass 1400 kbps and the result was similar in size and quality to the third file.
Originally Posted by vhelp -
which source do you mean?
if you mean-the tv channel, so it's the same channel, even the same tv-series which broadcast. I did exactly the same with the both videos. I didn't change any settings.
I use the smart deinterlace.
jagabo, I didn't touch the xvid encoding settings at all.
btw, what does 'OP' mean? -
it's something about the tv broadcast?
in any case, I recorded the both videos from the same channel, all should be the same, that's why I don't understand why there's a difference. -
With constant quality encoding the file sizes will depend on the video contents. Still shots will compress better than shots with motion or shaking. Noiseless video will compress better than noisy video. Low contrast video will compress better than high contrast video. Low detail video will compress better than high detail video.
So even if two videos are recorded from the same channel constant quality encoding will result in different file sizes because the content of the videos are different. -
ok I see, thanks.
another question: is there a way to make the xvid videos smaller, without causing damage to the video? for example, now I'll get about 800mb file, for video of 45mins. I want to decrease it to around 500mb. I saw some 45mins videos, of 300mb, in good quality (those videos were recorded by ppl from my country). can I do that too? -
can I do that too?
-
The question that I didn't ask was what is the origin of your source: Analog or Digital
Analog video (cabletv and antenna) is very noisy, and usually seen or refered to as salt and pepper (SnP) for the similarities of what looks like random black and white dots. In actuallity the noise is chroma crosstalk.
Digital, as in digital cable or satellite and also HDTV are clean sources, no filter would be necessary. You can still apply filtering to these source if you wish in a later step. Most people use avisynth for this.
If your pvr is filtering, then it might be wise to see what setting it contains because you might be able to lower it and its effects--or temporal errors, are the ghosting you see during movements. Look at the man's forehead as it sways left/right. If your source is from Digital then you can turn the filter off.
As for the 300MB size videos you mentioned. Some of those could be from digital sources, and then further filtered, and IVTC-ed to 24 or 25 progressive frames, if film based, (if that's what you are basing those from) else for interlaced, may be deinterlacing, etc. etc.
-vhelp 4922 -
well, the broadcast here is analog only. my capture card is analog as well (Hauppauge WinTV PVR-150 MCE). I use a "digital converter". it's a box what I also have at the living room.
manono, I don't know what the resolution (as mpeg-2) is, but the virtualdub makes it 720x576. which another resolution can I use? what is the common and recommended resolution? -
After Smart Deinterlace try resizing to 640x480 or 576x432 with the Precise Bilinear option. You can use Lanczos3 or Precise Bicubic to get sharper results but they won't compress as well. A very small Temporal Smoother (0 or 1) might help too.
-
I did exactly the same, but after I changed the numbers above to 640x480, the numbers below (which I marked here in red), were changed automatically to different numbers than the numbers in your picture.. why does it happen?
-
I was using a different resolution source file when I made that image. Don't worry about the Relative % numbers.
-
The relative % numbers changed because the relative % of the new resolution was changed from the original, by you. The numbers show the resolution change as a percentage from the original. 640 is 88% of the original res.
You keep asking how long certain processes will take. You have NOT filled out your computer specs. You do understand that these processes happen on the PC, and the spped of the PC will affect how long it takes?
I would suggest the OP start doing some serious reading as there seem to be too many things he needs to know. I would pick some Guides that seem to apply and just start reading. After 5 to 10 or so, things will start coming together. It's not a matter of where to Start, it's where to Stop. -
so I just need to change the resolution every time to 640x480, without regarding the Relative numbers?
Nelson37, my pc is not new, Pentium 4 3.0GHz, (one core), 2gb rem memory. -
Originally Posted by Yonidan
-
if I always record from the same tv channel, may the aspect ratio change suddenly?
-
Originally Posted by Yonidan
-
Not in the UK - stations change dynamically depending on the programme. All new programmes are 16:9. Only old archive shows are 4:3. Adverts are always 16:9, so a commercial channel showing a 4:3 programme with 16:9 adverts will change aspect ratio several times per hour.
We don't have that much HD yet.
Cheers,
David. -
He appears to be capturing from a standard definition source. It will always be 4:3. If widescreen material is broadcast on that station it will be letterboxed into a 4:3 frame. Or pan-and-scanned in a 4:3 frame. All of them should be resized to 640x480 or some other 4:3 frame size. Of course, the pillarboxed widescreen video could be cropped down to eliminate the black bars.
HDTV station here (USA) always broadcast the same format. A 1920x1080i station will upscale SD material and broadcast it pillarboxed in a 1920x1080i frame, or stretched to fill the entire 1920x1080i frame. I suspect the same is true in the UK. TVs don't like switching format, they usually take a second or two to adjust. Some channels do stretch non-16:9 widescreen videos to fill the screen. That is, a 2.4:1 movie would be stretched vertically to fill the 16:9 (1.78:1) screen. In that case you would want to use a 2:4:1 frame size rather than a 16:9 frame size. -
Originally Posted by jagabo
I didn't know the OP was capturing analogue broadcasts though. Those at least are always 4x3 raster (with or without letterboxing of the actual content).
Cheers,
David. -
Originally Posted by 2Bdecided
Originally Posted by 2Bdecided -
I'll dig one out for you. I might not mean much though - the UK digital terrestrial broadcast specification uses something called AFDs to signal how the content should be displayed, in addition to the standard MPEG aspect ratio flags.
This is kind of correct...
http://www.pjdaniel.org.uk/afd/index.asp
This is better, but not UK specific...
http://en.wikipedia.org/wiki/Active_Format_Descriptor
Cheers,
David. -
Here it is:
http://rapidshare.de/files/40877736/16x9_4x3.demuxed.m2v.html
16x9 > 4x3 change between the sponsor link and the actual program, from channel five, Freeview (DTT / OTA), UK.
Cheers,
David. -
Some players recognize the AR change and display both parts correctly. Others don't. I've added it to my collection of video odds and ends. Thanks for the sample.
-
thanks for all, but now I have a new problem:
I discovered that it's impossible to watch TV live and to capture simultaneously, with the TV capturing sofware I use (GB-PVR).
can anyone recommend me a pro software (except SageTV, that doesn't work in my pc), with an option of watching TV live and capturing at the same time?
Similar Threads
-
Need help from video experts, hdmi to vga to dvi
By edgarmiranda in forum Newbie / General discussionsReplies: 3Last Post: 20th Jun 2011, 22:24 -
Best way to add subtitles to youtube video? Need experts' advice!
By qi99 in forum SubtitleReplies: 31Last Post: 10th Jul 2010, 22:34 -
Attention: Video conversion experts
By Canon GL-2 Guy in forum Newbie / General discussionsReplies: 12Last Post: 1st Apr 2010, 11:57 -
Any Novell GroupWise Experts here? I have a problem
By scustalow in forum ComputerReplies: 1Last Post: 15th Jun 2008, 13:52 -
This Is A Loaded Question For You Video/Audio Experts Out There...
By pmolson in forum Newbie / General discussionsReplies: 2Last Post: 20th May 2007, 13:32