PTE Created Video File Sizes

[Dirk1503]

Ich komprimiere grundsätzlich die von PTE AV Studio erstellten mp4-Dateien noch einmal mit einem guten kostenlosen Konvertierungsprogramm, was auf Komprimierung ohne größere Qualitätsverluste spezialisiert ist (Zum Beispiel "Xmedia Recode, Clever FFmpeg, Handbrake oder Avidemux). Somit kann ich die mp4-Dateienauf etwa 20 Prozent der ursprünglichen Dateigröße reduzieren, ohne einen großen Qualitätsverlust zu bemerken.

____

I always compress the mp4 files created by PTE AV Studio again with a good free conversion program that specializes in compression without major quality loss (for example "Xmedia Recode, Clever FFmpeg, Handbrake or Avidemux). This allows me to reduce the mp4 files to about 20 percent of the original file size without noticing a major loss in quality. (Google Translate)

[Picsel]

Following previous interesting discussions concerning which parameter could have an impact on file size I have made several experiments on my side with an abstract

(duration 84s) of the project I was working on.

 

This abstract contents 8 slides, each of them includes only one photo with a definition of 5472x3648 pixels (3/2 format). The first two with zoom effect and vertical pan

and the 8 slides with dissolve transitions synchronized with audio.

The audio file size is 2,27Mo (mp3 constant bit rate 256kb/s with sampling at 48kHz) as part of the « file size » encompassing audio and video files.

I thought that results of these experiments could be of some interest for other members.

The results analysis raised several technical questions concerning file size ( see points 1 and 2) and concerning time encoding with Motion Blur (see point 4)

So you will find here attached two documents : configuration and results – results analysis

 

1) Video format impact on file size

Comparison between encoding at 1920x1080 format and 1920x1280 format : there are no noticeable differences although the number of pixels to be encoded is 18%

higher within 1920x1280 format that is quite strange for me ! Is there a technical explanation or error ?

 

2) FPS impact on encoding time and file size.

As we could expect with LLC mode, encoding time and file size are roughly twice (x1,8) higher at 50 than at 25 fps.

But using quality mode = 100 there are curiously no noticeable differences on resulting files size ! Is there a technical explanation ?

So it could be interesting of making choice of quality mode with 50fps or 60fps rather than 25 and 30fps (depending on national standards) in order to improve the visual

perception.

 

3) Choice of quality mode has an important impact on encoding time and file size

As already explained by Igor, with quality mode the coding rate is variable and depends on the image analysis which reduces drastically the file size.

So, as recommended, I often use quality mode. It gives quite good results even for most cases with quality 50 and optimizes the quality/file-size ratio when needed.

 

4) Motion Blur impact on encoding time and file size

As it was already mentionned in prenious posts by Denis and Igor, motion blur has no noticeable impact on file size which is even reduced a little bit.

But I was very surprised by its huge impact on encoding time (more than 10 times !) which leads me to make a test by replacing the 8 slides by only one black slide with the

same 84s duration. As a result : near 18minutes for encoding a black slide on my PC ! (I know it is abnormal to use Motion Blur for encoding a black slide but the result is

completely unexpected !)

I must have made an error somewhere or missed something with « motion blur » ! I know that my PC configuration is a bit old now but could that explain ?

When is it mandatory to use Motion Blur ?

Daniel

 

241001-configuration-results.pdf 241001-results-analysis.pdf

[Igor]

Hi Daniel,

Thanks for sharing the results of your interesting investigation.

1. If you only changed video output picture size (1920 x 1080 vs 1920 x 1280) it didn't change original show aspect ratio and thus it added black strips when original aspect ration was different with output size. So the effective picture content (sum of pixels) didn't change. So you need to change the project aspect ratio and output picture size together and reencode video again.

2. Interesting result. I think it may depends on the content - animated images or video content. Also probably Quality = 100 can be excess value for animated static images. I'll repeat your exploring later.

3. Yes, the contant quality compression is a great mode and it produces excellent results in terms of a quality and file size. I always recommend to use this mode, especially for static images or animated images.

4. Motion Blur. As far as I know, almost all apps that create this effect (including PTE AV Studio) render intermediate frames, which significantly slows down the video encoding time. For example, if you set FPS = 50, PTE generates 1000 frames per second (50 x 20) and blending these extra 20 frames into one final frame. PTE doesn't perform analysis for the content.
Motion Blur option effectively applies only to animated PTE objects, not for embedded video clips.

[Picsel]

Hi Igor,

thanks for your explanations !

1. Obviously you are right, I should have modified the aspect ratio to 3/2  and doing so the result with LCC give a 1920x1280 file size 1,26 times greater than 1920x1080 file size  which is more in line with the expected result.241003-results-v2.pdf

4. So, don't you think it should be better to place the motion blur option within Objects properties  (O&A) or Slide options rather than within encoding video parameters?

Daniel

pj : results with aspect ratio 3/2