2 videos: same quality and length, but different size, why ?

[gabirelms]

I have 2 videos, both of them have same length, same quality (in my eyes), same codec (XVID), same resolution and same fps. But file size is different, one video is around 200mb, and the other one around 400mb

I compared them with gspot, and all the info look the same, so... why file size is different ?
Is there any program that can tell me more information, ie Xvid compression ratio or quality ?

I can upload the videos to my dropbox and share them if necesary

Please I need help
Gabriel

[DB83]

file-size = length*bitrate

What does mediainfo tell you about the bitrate ?

[Ai Haibara]

Assuming the two videos aren't the exact same material... different things happening in a scene may be more or less compressable, compared to other scenes. For example, a scene where people are just sitting around talking, may be able to be compressed to a smaller size than, say, a loud, noisy and graphic chase scene. The same may hold true for audio, to an extent.

[Bjs]

No application can pickup two pass encoding or filter settings used which might be the case if the material match's.

Ive used a few filters to stabilize motion in sandy scenes which prevent xvid from getting greedy with requant which has produced smaller file size using the same bitrate settings ... and most people cannot pickup any difference other than file size.

[Cornucopia]

...same quality (in my eyes),...

There's the rub.
Your eyes (as are most peoples') are NOT discerning enough of the fine differences in quality. Not without manipulating motion or zooming in or using statistical analysis.

Scott

[lordsmurf]

...same quality (in my eyes),...

There's the rub.
Your eyes (as are most peoples') are NOT discerning enough of the fine differences in quality. Not without manipulating motion or zooming in or using statistical analysis.
Scott

Agreed.
Most people don't know what to look for to determine quality.

[gabirelms]

Sorry that I didnt explain myself nice.
Both videos are the same material, and of course one has higher bit rate than the other one, that is why the final size is bigger.

That is my question, why one video gets higher bit rate (higher video size) than the other one.

Im taking the videos with 2 video cameras, both of them from the same manufacturer, same camera sensor, same resolution (352x288), same fps (10fps), and same codec (Xvid). And of course I record the same material, in the same place at the same time.
I got one camera 1 year after the other one, and I think it has a new/different firmware. New camera has higher bit rate than old camera.

Then, my conclusion is that new camera has a lower compresion ratio than the older one. But after recording so many videos, I dont find any difference, both videos looks the same, but one is double size than the other one.

I will record 2 short videos and upload them, I hope a pro user can tell me if there is any difference between both of them.

[DB83]

Have to ask what these cameras are.

352*288 is very poor resolution. 10 fps is catastrophic.

[Hoser Rob]

... 352*288 is very poor resolution. 10 fps is catastrophic.

Regretfully, I agree. Seems to me they'd look equally bad.

[jagabo]

Many things can effect the image quality and bitrate. One camera may be using all I frames for example. Or very short GOPs with only I an P frames. Both of those will require more bitrate to deliver the same quality. And if the lower birate camera is delivering high quality, using even more bitrate won't improve quality significantly, just increase the file size. Different quantizer matrices will require more or less bitrate. Different motion search parameters. Etc.

[budwzr]

Is that a peeper cam? Or shoe-cam?

[gabirelms]

Have to ask what these cameras are.

352*288 is very poor resolution. 10 fps is catastrophic.

They are miniDV cameras, something like this:


Quality is far from HD, but they are very ligth and easy to carry in the pocket, nice as R/C camera, etc

[budwzr]

[Unsubscribed]

[DB83]

Whatever your cameras are they are NOT MiniDV.

[Cornucopia]

...Im taking the videos with 2 video cameras, both of them from the same manufacturer, same camera sensor, same resolution (352x288), same fps (10fps), and same codec (Xvid). And of course I record the same material, in the same place at the same time.
I got one camera 1 year after the other one, and I think it has a new/different firmware. New camera has higher bit rate than old camera...

No you don't record the same thing. Even dual cameras set up with a beamsplitter don't get the EXACT same thing - there are slight differences with reflections/angle-of-incidence, etc.
So, if you have them side-by-side, they're getting different perspectives (w/ different light in the background, etc). Plus, if,as you say, they are recording at different bitrates, that's the difference. Your "new firmware" model was set at the factory to record at a higher bitrate than your original model.

Can that be adjusted? Depends on the model/brand (I doubt it).

When Professionals want identical cameras, they buy 2 from the same batch at the same time and have them tweaked in a lab so the ARE as close as possible (none are EVER identical), then share saved "SceneFile" data to maintain that similarity.
But with lower-end consumer cams like you've got (sorry, not trying to denegrate), you're just lucky that your cams don't have noticeable color balance or light level differences...

Scott

[jagabo]

Also check the audio bitrate. With small frame sizes and low video bitrates the audio bitrate becomes more significant in the overall file size.

[Cornucopia]

Yep.

[gabirelms]

I really appreciate all your replies, thanks very much.

Many things can effect the image quality and bitrate. One camera may be using all I frames for example. Or very short GOPs with only I an P frames. Both of those will require more bitrate to deliver the same quality. And if the lower birate camera is delivering high quality, using even more bitrate won't improve quality significantly, just increase the file size. Different quantizer matrices will require more or less bitrate. Different motion search parameters. Etc.

jagabo, I think that can be the real difference between the 2 cameras, please do you know any software to analyse the video and get the information about the frames ?

audio bitrate is fixed and its the same for both of them, 128kbps

No you don't record the same thing. Even dual cameras set up with a beamsplitter don't get the EXACT same thing - there are slight differences with reflections/angle-of-incidence, etc.
So, if you have them side-by-side, they're getting different perspectives (w/ different light in the background, etc). Plus, if,as you say, they are recording at different bitrates, that's the difference. Your "new firmware" model was set at the factory to record at a higher bitrate than your original model.

Can that be adjusted? Depends on the model/brand (I doubt it).

Cornucopia, that was my first impression, but I recorded many videos, and in all of them one camera got higher file size than the other one.

By the way, video bitrate is not fixed, I mean, if I take a video at home, or outside, bitrate is different. But one camera will always get higher bitrate than the other one.

[jagabo]

I really appreciate all your replies, thanks very much.

Many things can effect the image quality and bitrate. One camera may be using all I frames for example. Or very short GOPs with only I an P frames. Both of those will require more bitrate to deliver the same quality. And if the lower birate camera is delivering high quality, using even more bitrate won't improve quality significantly, just increase the file size. Different quantizer matrices will require more or less bitrate. Different motion search parameters. Etc.

jagabo, I think that can be the real difference between the 2 cameras, please do you know any software to analyse the video and get the information about the frames ?

GSpot tells you the the ratio of I, P, and B frames, the average GOP length, and the max consecutive B frames. It does this by scanning the entire video, not just reading header information You can even view the I/P/B structure of the entire video with the VGS button. Be sure "enable full file scan.." is enabled.

Why don't you just post short samples from the two cameras.

[Cornucopia]

No you don't record the same thing. Even dual cameras set up with a beamsplitter don't get the EXACT same thing - there are slight differences with reflections/angle-of-incidence, etc.
So, if you have them side-by-side, they're getting different perspectives (w/ different light in the background, etc). Plus, if,as you say, they are recording at different bitrates, that's the difference. Your "new firmware" model was set at the factory to record at a higher bitrate than your original model.

Can that be adjusted? Depends on the model/brand (I doubt it).

Cornucopia, that was my first impression, but I recorded many videos, and in all of them one camera got higher file size than the other one.

By the way, video bitrate is not fixed, I mean, if I take a video at home, or outside, bitrate is different. But one camera will always get higher bitrate than the other one.

That does not contradict what I said.
Most cameras that do longGOP encoding in hardware have a fixed IBP cadence, and usually have a fixed bitrate, AFAIK. Maybe one cam is set to "Better" quality and the other is set to "Best", or some such - but behind the scenes and unavailable for you to tweak.

Your choices:

• Post mediainfo info about each of these cams and MAYBE learn someting that MIGHT be user-adjustable (and then adjust it).
• Live with the size mismatch
• Spend time & energy & quality recompressing one or the other or both
• Buy a new PAIR of (hopefully better) cameras

Scott