Ok, XQ stays at native res. IQ will be confined to 1280x720.
Expected final results for self-encoded runs has been updated.
Expected final results for self-encoded runs has been updated.
SDA Forum
Anri 3.0 is out. Here's the x264 version it's using. Command line:
Code:
:2pass "%anri_dir%x264.exe" --profile high --level 4.1 --pass 1 --bitrate %3 --stats "%2.stats" --thread-input --b-adapt 2 --direct auto --qpmin %5 --vbv-bufsize 24000 --vbv-maxrate 24000 --no-mixed-refs --output NUL %1 "%anri_dir%x264.exe" --profile high --level 4.1 --pass 2 --bitrate %3 --stats "%2.stats" --thread-input --b-adapt 2 --direct auto --qpmin %5 --vbv-bufsize 24000 --vbv-maxrate 24000 --me umh --weightp 2 --trellis 2 --output "%2_video.mp4" %1 :ipod "%anri_dir%x264.exe" --profile baseline --level 1.3 --pass 1 --bitrate %3 --stats "%2.stats" --thread-input --no-cabac --qpmin %5 --vbv-bufsize 768 --vbv-maxrate 768 --no-mixed-refs --output NUL %1 "%anri_dir%x264.exe" --profile baseline --level 1.3 --pass 2 --bitrate %3 --stats "%2.stats" --thread-input --no-cabac --ref 1 --qpmin %5 --vbv-bufsize 768 --vbv-maxrate 768 --me umh --partitions p8x8,b8x8,i4x4 --weightp 2 --no-8x8dct --trellis 0 --no-mixed-refs --output "%2_video.mp4" %1
Quote from ballofsnow:
Anri 3.0 is out. Here's the x264 version it's using. Command line:
Code:
:2pass "%anri_dir%x264.exe" --profile high --level 4.1 --pass 1 --bitrate %3 --stats "%2.stats" --thread-input --b-adapt 2 --direct auto --qpmin %5 --vbv-bufsize 24000 --vbv-maxrate 24000 --no-mixed-refs --output NUL %1 "%anri_dir%x264.exe" --profile high --level 4.1 --pass 2 --bitrate %3 --stats "%2.stats" --thread-input --b-adapt 2 --direct auto --qpmin %5 --vbv-bufsize 24000 --vbv-maxrate 24000 --me umh --weightp 2 --trellis 2 --output "%2_video.mp4" %1 :ipod "%anri_dir%x264.exe" --profile baseline --level 1.3 --pass 1 --bitrate %3 --stats "%2.stats" --thread-input --no-cabac --qpmin %5 --vbv-bufsize 768 --vbv-maxrate 768 --no-mixed-refs --output NUL %1 "%anri_dir%x264.exe" --profile baseline --level 1.3 --pass 2 --bitrate %3 --stats "%2.stats" --thread-input --no-cabac --ref 1 --qpmin %5 --vbv-bufsize 768 --vbv-maxrate 768 --me umh --partitions p8x8,b8x8,i4x4 --weightp 2 --no-8x8dct --trellis 0 --no-mixed-refs --output "%2_video.mp4" %1
Why is trellis disabled for ipod? It doesn't affect decoding in any way, disabling it just lowers the quality (and increases encoding speed - a bad tradeoff for small resolutions IMO). And are you sure the ipod requires --ref 1? The lowest I've seen any device support was 2, and that was for 640x480... 3 (default) really shouldn't be a problem. Also I doubt --no-mixed-refs is necessary.
The qpmins are too high IMO and aren't the recommended way to do what you want to do (it's better to read the "final ratefactor" from the first pass, and if it's smaller than, say, 18 or so, run the second pass with --crf 18 instead of --bitrate whatever). This is because adaptive quantization might want to use lower quantizers than the average in flat areas (it's easy to see 12-14 for crf 19).
Also, --thread-input is redundant (automatically enabled when multithreading), and --profile baseline implies --no-cabac, --no-8x8dct, --partitions (not i8x8) and --weightp 0 (this last one is being overridden for you).
JBM, thanks for the input. I will review the settings for the next version.
bmn, I'll change the max to 1366x768. According to wikipedia it's uncommon to have native res of 1280x720.
bmn, I'll change the max to 1366x768. According to wikipedia it's uncommon to have native res of 1280x720.
Fucking Weeaboo
Quote from ballofsnow:
JBM, thanks for the input. I will review the settings for the next version.
bmn, I'll change the max to 1366x768. According to wikipedia it's uncommon to have native res of 1280x720.
bmn, I'll change the max to 1366x768. According to wikipedia it's uncommon to have native res of 1280x720.
Just to note, even those tiny Acer netbooks (with 11" monitors) have a widescreen resolution of 1366x768. No joke. ^^;;
(Too bad these things aren't beefy enough to handle a HQ video, much less an IQ one.)
Hi,
I've just updated to Anri Chan 3 (I used anri chan 3 beta 4 and I had lots of failed encoding) and I noticed a big increase for video size.
After-encoding sizes for the same video in normal quality :
anri chan 3 beta 4 : 950 Mb
anri chan 3 : about 1200 Mb (don't remember exactly)
Have some changes been made in compression rate ?
I can't use zshare.net to upload my video for free anymore because of the 1Gb+ size.
Are there free upload websites with a higher size limitation ?
thx
I've just updated to Anri Chan 3 (I used anri chan 3 beta 4 and I had lots of failed encoding) and I noticed a big increase for video size.
After-encoding sizes for the same video in normal quality :
anri chan 3 beta 4 : 950 Mb
anri chan 3 : about 1200 Mb (don't remember exactly)
Have some changes been made in compression rate ?
I can't use zshare.net to upload my video for free anymore because of the 1Gb+ size.
Are there free upload websites with a higher size limitation ?
thx
Entirely possible. But the qualities still shouldn't exceed LQ/MQ/HQ's ~192/576/2176 kbps.
Separate note:
Anri needs a workaround to the problem with AviSynth/Windows/whatever not accepting all of the user's videos that need appending. One that comes to mind is encoding NMF segments. I'll have to think more about this later. Ideas are welcome.
Separate note:
Anri needs a workaround to the problem with AviSynth/Windows/whatever not accepting all of the user's videos that need appending. One that comes to mind is encoding NMF segments. I'll have to think more about this later. Ideas are welcome.
Here's some testing I did:
I took a short fraps video, made 127 copies. Ran it through Anri 3.0 as avi automatic. The verification process started and failed. Brought it down to 76 files, still failed in Anri, still failed when dragging the temp file verifysource.avs directly into virtualdub. Tried 75, passed anri verification and loaded in vdub, encoded to LQ/MQ/HQ.
Directshowsource performed worse than avisource. When dragging into virtualdub it failed at the 39th video.
Still need to figure out the logic to get around this problem.
I took a short fraps video, made 127 copies. Ran it through Anri 3.0 as avi automatic. The verification process started and failed. Brought it down to 76 files, still failed in Anri, still failed when dragging the temp file verifysource.avs directly into virtualdub. Tried 75, passed anri verification and loaded in vdub, encoded to LQ/MQ/HQ.
Directshowsource performed worse than avisource. When dragging into virtualdub it failed at the 39th video.
Still need to figure out the logic to get around this problem.
Was working on the above, but got distracted.
See http://speeddemosarchive.com/w/index.php?title=Anri-chan/Source/anri.bat&curid=1498&diff=4466&oldid=4387
I put defaults for most of the questions. If you disagree with any of them, let me know.
See http://speeddemosarchive.com/w/index.php?title=Anri-chan/Source/anri.bat&curid=1498&diff=4466&oldid=4387
I put defaults for most of the questions. If you disagree with any of them, let me know.
cool - but why not make the defaults the last used settings? can probably just save/load settings every exit/launch and then show those instead of hard coding them.
i know i do nothing on this project and coding in doze batch is just freaking fantastic fun so it's anything but an order - consider it a jocular suggestion.
on another note, how much win7 testing have we done? i've done none and i ask because of this.
wild on the failure of directshowsource()++. i guess things are pretty dire for appending hella vids. an nmf pyramid scheme is the only way out i can think of that seems more than mildly reliable.
i know i do nothing on this project and coding in doze batch is just freaking fantastic fun so it's anything but an order - consider it a jocular suggestion.
on another note, how much win7 testing have we done? i've done none and i ask because of this.
wild on the failure of directshowsource()++. i guess things are pretty dire for appending hella vids. an nmf pyramid scheme is the only way out i can think of that seems more than mildly reliable.
Quote from nate:
cool - but why not make the defaults the last used settings? can probably just save/load settings every exit/launch and then show those instead of hard coding them.
i know i do nothing on this project and coding in doze batch is just freaking fantastic fun so it's anything but an order - consider it a jocular suggestion.
i know i do nothing on this project and coding in doze batch is just freaking fantastic fun so it's anything but an order - consider it a jocular suggestion.
Good idea, adding to the to-do list. Probably won't be in 3.1 though.
Quote:
on another note, how much win7 testing have we done? i've done none and i ask because of this.
I'm running Win 7 Pro x64, also have XP mode installed for further testing.
Quote:
wild on the failure of directshowsource()++. i guess things are pretty dire for appending hella vids. an nmf pyramid scheme is the only way out i can think of that seems more than mildly reliable.
Yeah... It's going to be interesting finding a fix. Vdub/AviSynth isn't returning a meaningful error code for "oh shit too many files!!". It just says it couldn't load the decompressor for fourcc: (blank)....
I'm looking at the video verification step which is before you get into the settings and encoding. Right now it just appends everything together. The idea is that if AviSynth can load it then all is well, and if it can't, don't waste time going through the settings and attempting to encode the videos when you really had zero chance of a successful encode to begin with.
Maybe a game of guess the number would work.
Maybe a game of guess the number would work.
It varies. I know my old computer never got close to appending 75 videos like I did on my current computer.
Some thoughts on bringing in HD console support: Anri detects NTSC vs PAL based on resolution. With HD we can't use this method anymore.
Why NTSC and PAL Still Matter With HDTV
So it needs to be based on framerate instead. Do you think this would work? I mean.. PAL 25 / NTSC 29.97... but would something like PAL-60 mess things up?
Some thoughts on bringing in HD console support: Anri detects NTSC vs PAL based on resolution. With HD we can't use this method anymore.
Why NTSC and PAL Still Matter With HDTV
Quote:
when comparing HDTV to analog NTSC and PAL standards, the fundamental common foundation of both systems is the Frame Rate.
So it needs to be based on framerate instead. Do you think this would work? I mean.. PAL 25 / NTSC 29.97... but would something like PAL-60 mess things up?
I have a question about DVD ripping. I recently finished my glitched run of Tomb Raider 3 and began ripping all the segments with anri. All of them ripped fine and in less than 5 minutes each, except for one. It's strange because I have a differnet ripping program that was able to do it fine. I'm guessing there must be a scratch or something on part of the writing of the disc that I cannot see. I'll give it a few more tries though.
try looking at the disc in the sun. sometimes you can see what it is especially if you know about how far through the disc it screws up (it starts at the middle and goes to the outside). you can also easily see how much data is on the disc by how it looks. if you see something you can try rubbing it off by using a very soft wet cloth.
is there pal-60 in hd? or are we talking about doing this for all resolutions? pal-60 is 30 fps ?x480 afaik.
Quote from ballofsnow:
So it needs to be based on framerate instead. Do you think this would work? I mean.. PAL 25 / NTSC 29.97... but would something like PAL-60 mess things up?
is there pal-60 in hd? or are we talking about doing this for all resolutions? pal-60 is 30 fps ?x480 afaik.
May have found the answer to the avisource problem with too many files: use ffmpegsource instead.
Loaded up 300 avi files (same file over and over again) which used about a gig of memory. The limit seems to be system memory. Avisynth will throw an exception if your system runs out of memory.
Code may look like:
Obviously this would be a pretty fundamental change in how videos are loaded. If all goes well I'll probably release an experimental version of Anri 3.1.
Quote:
Compatibility
* AVI, MKV, MP4, FLV: Frame accurate
* WMV: Frame accurate(?) but avformat seems to pick keyframes relatively far away
* OGM: Frame accurate(?)
* VOB, MPG: Seeking seems to be off by one or two frames now and then
* M2TS, TS: Seeking seems to be off a few frames here and there
* Image files: Most formats can be opened if seekmode=-1 is set, no animation support
* AVI, MKV, MP4, FLV: Frame accurate
* WMV: Frame accurate(?) but avformat seems to pick keyframes relatively far away
* OGM: Frame accurate(?)
* VOB, MPG: Seeking seems to be off by one or two frames now and then
* M2TS, TS: Seeking seems to be off a few frames here and there
* Image files: Most formats can be opened if seekmode=-1 is set, no animation support
Loaded up 300 avi files (same file over and over again) which used about a gig of memory. The limit seems to be system memory. Avisynth will throw an exception if your system runs out of memory.
Code may look like:
Code:
Audiodub(ffvideosource("AM001.avi"), ffaudiosource("AM001.avi"))
\++\
Audiodub(ffvideosource("AM002.avi"), ffaudiosource("AM002.avi"))Obviously this would be a pretty fundamental change in how videos are loaded. If all goes well I'll probably release an experimental version of Anri 3.1.