VideoHelp Forum
+ Reply to Thread
Page 2 of 2
FirstFirst 1 2
Results 31 to 48 of 48
Thread
  1. What filtering are you referring to? I mean, the code is the same, there must be some randomized calculations inside those functions which sometimes do a better job and sometimes worse, but in average they do a good job.
    Quote Quote  
  2. Originally Posted by Santuzzu View Post
    What filtering are you referring to? I mean, the code is the same, there must be some randomized calculations inside those functions which sometimes do a better job and sometimes worse, but in average they do a good job.
    If you use the same filters/settings (and same versions of the filters) and same encoder/settings you will get similar file sizes every time you encode the same source. Something was different between the 41 and 31 MB encodings.
    Quote Quote  
  3. Originally Posted by jagabo View Post
    If you use the same filters/settings (and same versions of the filters) and same encoder/settings you will get similar file sizes every time you encode the same source. Something was different between the 41 and 31 MB encodings.
    If some randomized algorithms are in the background, it usually happens that the output is not the same. If there are no randomized calculations inside these functions, then it can be only different versions of filters (more likely) or different settings (I doubt this is the case). Anyway, it can be confirmed by trying the same encoding process two/more times under the same installation and checking the file size.
    Quote Quote  
  4. Member
    Join Date
    Mar 2008
    Location
    United States
    Search Comp PM
    Randomized ? I don't think so. When using CRF the size you get is highly depended on the source. How much detail, how much noise, etc,etc.
    For example, using a soft resizer such as bilinear will give a smaller file than a resizer that accentuates detail and sharpness such as lanczos
    Quote Quote  
  5. Originally Posted by davexnet View Post
    Randomized ? I don't think so. When using CRF the size you get is highly depended on the source. How much detail, how much noise, etc,etc.
    For example, using a soft resizer such as bilinear will give a smaller file than a resizer that accentuates detail and sharpness such as lanczos
    "Highly dependent" and "Completely dependent" are different things and this actually was my question. Of course, in case of two different filters it's expected that the file size will be different, whereas the same file size in that case would be just a coincidence.
    Quote Quote  
  6. Some filters use random dither/noise patterns. But the nature of the variations make the difference in compression very small from run to run.

    Also, x264 isn't deterministic by default. There will be small differences each time you encode. The number of threads used makes a difference too. But those differences are very small until you get up to several dozen of threads.
    Last edited by jagabo; 24th May 2019 at 23:10.
    Quote Quote  
  7. Originally Posted by jagabo View Post
    Also, x264 isn't deterministic by default.
    It's vice versa: x264 default is deterministic, can be turned off via --non-deterministic switch.
    Quote Quote  
  8. Originally Posted by sneaker View Post
    Originally Posted by jagabo View Post
    Also, x264 isn't deterministic by default.
    It's vice versa: x264 default is deterministic, can be turned off via --non-deterministic switch.
    Sorry, I remembered it wrong.
    Quote Quote  
  9. @Santuzzu
    You seem to compare jagabo's 41MB file from post #12 which includes the audio with your encoded file which is video only, right?
    Quote Quote  
  10. If I remember correctly, the audio was only about 3 MB.
    Quote Quote  
  11. Originally Posted by Sharc View Post
    @Santuzzu
    You seem to compare jagabo's 41MB file from post #12 which includes the audio with your encoded file which is video only, right?
    Yes, my mistake. But as @jagabo wrote, the audio doesn't increase the overall file size significantly, it's only 3.2 MB.
    Quote Quote  
  12. Check the x264 metadata (MediaInfo shows it) in your encoding vs. mine. See if there are any differences.
    Quote Quote  
  13. On the left side is your encoding, on the right side is mine.

    Click image for larger version

Name:	avc.png
Views:	77
Size:	17.9 KB
ID:	49167
    Quote Quote  
  14. You need to compare the entire "encoding settings" string. You can mark the text then copy/paste into Notepad.

    The version of x264.exe I used is older than yours. That may be the source of the difference.
    Last edited by jagabo; 25th May 2019 at 18:26.
    Quote Quote  
  15. Originally Posted by jagabo View Post
    You need to compare the entire "encoding settings" string. You can mark the text then copy/paste into Notepad.
    Encoding settings : cabac=1 / ref=5 / deblock=1:0:0 / analyse=0x3:0x113 / me=umh / subme=8 / psy=1 / psy_rd=1.00:0.00 / mixed_ref=1 / me_range=16 / chroma_me=1 / trellis=1 / 8x8dct=1 / cqm=0 / deadzone=21,11 / fast_pskip=1 / chroma_qp_offset=-2 / threads=6 / lookahead_threads=1 / sliced_threads=0 / nr=0 / decimate=1 / interlaced=0 / bluray_compat=0 / constrained_intra=0 / bframes=3 / b_pyramid=2 / b_adapt=2 / b_bias=0 / direct=3 / weightb=1 / open_gop=0 / weightp=2 / keyint=50 / keyint_min=5 / scenecut=40 / intra_refresh=0 / rc_lookahead=50 / rc=crf / mbtree=1 / crf=22.0 / qcomp=0.60 / qpmin=0 / qpmax=69 / qpstep=4 / vbv_maxrate=30000 / vbv_bufsize=30000 / crf_max=0.0 / nal_hrd=none / ip_ratio=1.40 / aq=1:1.00

    ----------------------------------------------------------------------------------------------------------------------------------------------------------------------

    Encoding settings : cabac=1 / ref=5 / deblock=1:0:0 / analyse=0x3:0x113 / me=hex / subme=8 / psy=1 / psy_rd=1.00:0.00 / mixed_ref=1 / me_range=16 / chroma_me=1 / trellis=2 / 8x8dct=1 / cqm=0 / deadzone=21,11 / fast_pskip=1 / chroma_qp_offset=-2 / threads=6 / lookahead_threads=1 / sliced_threads=0 / nr=0 / decimate=1 / interlaced=0 / bluray_compat=0 / constrained_intra=0 / bframes=3 / b_pyramid=2 / b_adapt=1 / b_bias=0 / direct=3 / weightb=1 / open_gop=0 / weightp=2 / keyint=500 / keyint_min=50 / scenecut=40 / intra_refresh=0 / rc_lookahead=50 / rc=crf / mbtree=1 / crf=22.0 / qcomp=0.60 / qpmin=0 / qpmax=69 / qpstep=4 / ip_ratio=1.40 / aq=1:1.00
    Quote Quote  
  16. I had keyint set to 50 and keyin_min set to 5. That would make my file bigger.
    Quote Quote  
  17. I'm using MeGUI and couldn't find this parameter in advanced settings. Is there any .ini file to set this parameter, or perhaps it could be set from MeGUI or directly from command prompt?
    Quote Quote  
  18. It's called GOP size in MeGUI. But really, who cares. You get different sizes because of different filters, different encoders, and different settings. Not because the software just randomly produces different sizes each time.
    Quote Quote  



Similar Threads

Visit our sponsor! Try DVDFab and backup Blu-rays!