2-Pass Encoding, different settings for both passes?

MysteryX · 15th December 2014, 18:49

For encoding AviSynth results with x264, what gives best quality is 2-pass encoding. However, processing the output of AviSynth twice would just kill it (double the processing time).

Now I'm thinking... the first pass is to get statistic information about each frame. What takes the most time is the EEDI3, and then the NNEDI3 filters.

Can I do the first pass WITHOUT EEDI3 and NNEDI3 filters with a plain Spline resize, and then use the real settings only for the 2nd pass?

Are there any side-effects to doing this?

TheFluff · 15th December 2014, 18:57

no don't do that, it can have weird and unpredictable results

also, there is really no reason to use twopass unless you absolutely have to hit an exact filesize, and you probably don't, so don't use twopass

if you still really do need twopass and want to avoid running avisynth twice, encode to a lossless intermediary file first

Reel.Deel · 15th December 2014, 18:58

How about saving to a lossless file and then encoding with x264?

Edit: TheFluff was quicker

MysteryX · 15th December 2014, 19:11

Quote:

Originally Posted by Reel.Deel

How about saving to a lossless file and then encoding with x264?

Great idea!

Is 2-pass really worth it? I was thinking I could just use CBR but then one file would go from 35MB to 60MB which is reasonable but then another file would go from 25MB to 108MB which is not. I want to keep the file size in reasonable proportion to the original file because I had hundreds of such videos to re-encode.

Doing it this way might not be that bad on performance though. Most of the time is running through the EEDI3/NEEDI3 filters.

Asmodian · 15th December 2014, 20:09

Not CBR but CRF. CBR is basically never a good idea for us home users.

2-pass means the relative quality will vary while the size will be predictable, CRF means the quality will be predictable while the size will not be. You cannot have both and only you can decide if one is worth it over the other for you.

I always use CRF myself but I am targeting a many TB raid array or similar size agnostic storage. I might as well not do the encode if it comes out with bad quality due to the bit-rate being too low. If a video is too big turn up CRF for everything. The more space saved from the easier to compress videos the more space for everything. If you cannot turn up CRF anymore because it starts to look bad then the large video you want to use two-pass to reduce the size of would look very bad after two-pass encoding because the bitrate would be much too low. In my mind two-pass is for filling a specific storage medium, like a DVD+R, not controlling size of an entire collection.

15th December 2014, 18:49	#1 \| Link
MysteryX Soul Architect Join Date: Apr 2014 Posts: 2,559	2-Pass Encoding, different settings for both passes? For encoding AviSynth results with x264, what gives best quality is 2-pass encoding. However, processing the output of AviSynth twice would just kill it (double the processing time). Now I'm thinking... the first pass is to get statistic information about each frame. What takes the most time is the EEDI3, and then the NNEDI3 filters. Can I do the first pass WITHOUT EEDI3 and NNEDI3 filters with a plain Spline resize, and then use the real settings only for the 2nd pass? Are there any side-effects to doing this? __________________ FrameRateConverter \| AvisynthShader \| AvsFilterNet \| Natural Grounding Player with Yin Media Encoder, 432hz Player, Powerliminals Player and Audio Video Muxer Last edited by MysteryX; 24th June 2015 at 06:25.

15th December 2014, 18:57	#2 \| Link
TheFluff Excessively jovial fellow Join Date: Jun 2004 Location: rude Posts: 1,100	no don't do that, it can have weird and unpredictable results also, there is really no reason to use twopass unless you absolutely have to hit an exact filesize, and you probably don't, so don't use twopass if you still really do need twopass and want to avoid running avisynth twice, encode to a lossless intermediary file first Last edited by TheFluff; 15th December 2014 at 19:00.

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

15th December 2014, 18:58	#3 \| Link
Reel.Deel Registered User Join Date: Mar 2012 Location: Texas Posts: 1,665	How about saving to a lossless file and then encoding with x264? Edit: TheFluff was quicker

15th December 2014, 20:09	#5 \| Link
Asmodian Registered User Join Date: Feb 2002 Location: San Jose, California Posts: 4,407	Not CBR but CRF. CBR is basically never a good idea for us home users. 2-pass means the relative quality will vary while the size will be predictable, CRF means the quality will be predictable while the size will not be. You cannot have both and only you can decide if one is worth it over the other for you. I always use CRF myself but I am targeting a many TB raid array or similar size agnostic storage. I might as well not do the encode if it comes out with bad quality due to the bit-rate being too low. If a video is too big turn up CRF for everything. The more space saved from the easier to compress videos the more space for everything. If you cannot turn up CRF anymore because it starts to look bad then the large video you want to use two-pass to reduce the size of would look very bad after two-pass encoding because the bitrate would be much too low. In my mind two-pass is for filling a specific storage medium, like a DVD+R, not controlling size of an entire collection.