What kind of encoding parameters does Cyberlink Espresso converter use ?

[St Devious]

Cyberlink Espresso software can use ATI's stream and Nvidia's CUDA technology to accelerate video encoding.

I just converted video with using HD 4870.

These are the settings i used in the software
http://i39.tinypic.com/14a8jza.jpg

and this is the info i got with Media Info software

Format : MPEG-TS
File size : 86.7 MiB
Duration : 2mn 7s
Overall bit rate : 5 720 Kbps
Maximum Overall bit rate : 40.0 Mbps

Video
ID : 4113 (0x1011)
Menu ID : 1 (0x1)
Format : AVC
Format/Info : Advanced Video Codec
Format profile : Main@L3.0
Format settings, CABAC : No
Format settings, ReFrames : 2 frames
Duration : 2mn 7s
Bit rate : 5 237 Kbps
Width : 720 pixels
Height : 480 pixels
Display aspect ratio : 4/3
Frame rate : 29.970 fps
Standard : PAL
Resolution : 24 bits
Colorimetry : 4:2:0
Scan type : Progressive
Bits/(Pixel*Frame) : 0.506

Where can I find more info of what kind of encoder it uses or what are the encoding parameters it uses ?

[Inspector.Gadget]

Format settings, CABAC : No
Format settings, ReFrames : 2 frames


:( Nothing worthwhile, I can tell you that.

[Dark Shikari]

I'm pretty sure there are still only two CUDA encoders, nvcuenc.dll and badaboom, and they both suck.

[St Devious]

I'm pretty sure there are still only two CUDA encoders, nvcuenc.dll and badaboom, and they both suck.

I'm investigating this as we speak but with ati stream for now.

Inspector Gadget's suggestion in the other thread helped me.

still trying to work out some quirks like different encoding framerates.

i'll have a few comparisons in the next post

[St Devious]

Hardware - Intel Q9450 quad core @ 3.2 GHz
4 GB RAM @ 800 MHz DDR2

This is the stats of the original video.

Format : MPEG-TS
File size : 536 MiB
Duration : 2mn 8s
Overall bit rate : 35.0 Mbps

Video
ID : 49 (0x31)
Menu ID : 1 (0x1)
Format : MPEG Video
Format version : Version 2
Format profile : Main@High
Format settings, Matrix : Default
Duration : 2mn 8s
Bit rate mode : Variable
Bit rate : 33.2 Mbps
Nominal bit rate : 30.0 Mbps
Width : 1 920 pixels
Height : 1 080 pixels
Display aspect ratio : 16/9
Frame rate : 25.000 fps
Colorimetry : 4:2:0
Scan type : Progressive
Bits/(Pixel*Frame) : 0.641


Video Converted with Cyberlink Espresso using ATI Stream on HD 4870

Time taken - 47s

MediaInfo
Format : MPEG-TS
File size : 90.2 MiB
Duration : 2mn 7s
Overall bit rate : 5 948 Kbps
Maximum Overall bit rate : 40.0 Mbps

Video
ID : 4113 (0x1011)
Menu ID : 1 (0x1)
Format : AVC
Format/Info : Advanced Video Codec
Format profile : Main@L3.0
Format settings, CABAC : No
Format settings, ReFrames : 2 frames
Duration : 2mn 7s
Bit rate : 5 456 Kbps
Width : 720 pixels
Height : 480 pixels
Display aspect ratio : 16/9
Frame rate : 29.970 fps
Standard : PAL
Resolution : 24 bits
Colorimetry : 4:2:0
Scan type : Progressive
Bits/(Pixel*Frame) : 0.527

PNG Screen -
http://i39.tinypic.com/e05ogz.png

Video Converted with Megui + x264 r1162 + Avisynth Billinear resize


Time taken - 60s

MediaInfo
Format : MPEG-4
Format profile : JVT
Codec ID : avc1
File size : 76.1 MiB
Duration : 2mn 7s
Overall bit rate : 5 020 Kbps
Encoded date : UTC 2009-06-02 01:47:49
Tagged date : UTC 2009-06-02 01:47:49

Video
ID : 1
Format : AVC
Format/Info : Advanced Video Codec
Format profile : Baseline@L3.0
Format settings, CABAC : No
Format settings, ReFrames : 2 frames
Codec ID : avc1
Codec ID/Info : Advanced Video Coding
Duration : 2mn 7s
Bit rate mode : Variable
Bit rate : 5 019 Kbps
Nominal bit rate : 5 237 Kbps
Maximum bit rate : 17.9 Mbps
Width : 720 pixels
Height : 480 pixels
Display aspect ratio : 16/9
Frame rate mode : Constant
Frame rate : 25.000 fps
Standard : NTSC
Resolution : 24 bits
Colorimetry : 4:2:0
Scan type : Progressive
Bits/(Pixel*Frame) : 0.581
Stream size : 76.1 MiB (100%)
Writing library : x264 core 67 r1162M f7bfcfa
Encoding settings : cabac=0 / ref=2 / deblock=0:0:0 / analyse=0x1:0 / me=dia / subme=1 / psy_rd=0.0:0.0 / mixed_ref=0 / me_range=16 / chroma_me=0 / trellis=0 / 8x8dct=0 / cqm=0 / deadzone=21,11 / chroma_qp_offset=0 / threads=6 / nr=0 / decimate=1 / mbaff=0 / bframes=0 / keyint=250 / keyint_min=25 / scenecut=40 / rc=abr / bitrate=5237 / ratetol=0.1 / qcomp=0.60 / qpmin=10 / qpmax=51 / qpstep=4 / ip_ratio=1.40 / aq=0
Encoded date : UTC 2009-06-02 01:47:49
Tagged date : UTC 2009-06-02 01:48:53

I tried to use to lowest settings for the fastest time possible. I noticed that my CPU is not used 100% due to the most CPU intensive settings not being used.

PNG Screen
http://i43.tinypic.com/259ljis.png

The GPU encode has sharper "Neuhaser" sign, while the x264 encode has better grain on the wall.

please let me know if I am doing anything wrong.

[Dark Shikari]

How about you do a GPU encode and post the stream so I can do some reverse-engineering?

[St Devious]

How about you do a GPU encode and post the stream so I can do some reverse-engineering?

sure, but no personal info is embedded in the stream after the encode, right ?

[Dark Shikari]

sure, but no personal info is embedded in the stream after the encode, right ?How would I know? I would assume no.

Also, could you post the clip of the source that you use to encode the sample you post, so I can encode it with x264 as well to compare?

[St Devious]

How would I know? I would assume no.

Also, could you post the clip of the source that you use to encode the sample you post, so I can encode it with x264 as well to compare?

sure,

here is the GPU encode

http://www.mediafire.com/?tdn2zejjzmz

here is the original clip

http://www.w6rz.net/1080p25.zip

[CruNcher]

it's still the same nvcuvenc.dll for the Nvidia Encoding part that all the other 3rd Party Nvidia Cuda Apps use nothing new better then Cyberlinks own Encoder but still inferior to x264 in many terms :), especially when it goes low bitrate encoding or Edge Situations. Nice to see some ATI Encoder Results finaly, though you shouldn't compare ATI's Main Profile vs x264 Baseline that makes no real sense :) also subme 1 is much to low at least use subme 2 (ok the goal was to reach the same time not the same or better efficiency for a little more time, could you please post your whole encoding chain avisynth script ?, MSE uses i think the same as badaboom the GPU for Decoding at least for Nvidia so that could be a slight time difference + downscaling via the GPU :) )

this whole GPU chain can by now pretty good simulated via Avisynth useing all the availabe GPU filter :)

Your chain should look like this Mpeg-2 Decoding (GPU)->Resizing (GPU)->Encoding (CPU) <- not sure if this Avisynth chain works completely yet for ATI based cards i guess Decoding is a problem (no GPU Decoder for ATI yet) and resizing im not sure if the Chinese stuff works on ATI cards :)

The Result isn't progressive the same problem people reported in the H.264 forum with MSE and Nvidia it's Interlaced (even if it says Progressive), somehow the Bitstream information don't match with the container one and Mediainfo seems to prefer the Container information.

[deekey777]

I've converted the Big Bug Bunny sample ( -> http://www.fixstars.com/en/products/codecsys/sample.html ) to 1280x720 (PS3). and MediaInfo says:
General
ID : 0
Complete name : E:\Neuer Ordner (3)\Big_Buck_Bunny.m2ts
Format : MPEG-TS
File size : 34.2 MiB
Duration : 27s 441ms
Overall bit rate : 10.4 Mbps
Maximum Overall bit rate : 40.0 Mbps

Video
ID : 4113 (0x1011)
Menu ID : 1 (0x1)
Format : AVC
Format/Info : Advanced Video Codec
Format profile : High@L4.0
Format settings, CABAC : Yes
Format settings, ReFrames : 2 frames
Duration : 27s 440ms
Bit rate : 9 777 Kbps
Width : 1 280 pixels
Height : 720 pixels
Display aspect ratio : 16/9
Frame rate : 25.000 fps
Resolution : 24 bits
Colorimetry : 4:2:0
Scan type : Progressive
Bits/(Pixel*Frame) : 0.424

Audio
ID : 4352 (0x1100)
Menu ID : 1 (0x1)
Format : AC-3
Format/Info : Audio Coding 3
Duration : 27s 441ms
Bit rate mode : Constant
Bit rate : 256 Kbps
Channel(s) : 2 channels
Channel positions : L R
Sampling rate : 48.0 KHz
Video delay : 8ms

I'll upload it later.

http://rapidshare.com/files/240372316/Big_Buck_Bunny.m2ts.html
MD5: 477FEAA2DDA9E8B3C2E16548BBE06164

[St Devious]

Nice to see some ATI Encoder Results finaly, though you shouldn't compare ATI's Main Profile vs x264 Baseline that makes no real sense :) also subme 1 is much to low at least use subme 2 (ok the goal was to reach the same time not the same or better efficiency for a little more time,


I intended to compare main in both, but i missed that. i'll reconvert and update the OP.

could you please post your whole encoding chain avisynth script ?, MSE uses i think the same as badaboom the GPU for Decoding at least for Nvidia so that could be a slight time difference + downscaling via the GPU )




I don't know how to do that, you will have to help me there

[Dark Shikari]

sure,

here is the GPU encode

http://www.mediafire.com/?tdn2zejjzmz

here is the original clip

http://www.w6rz.net/1080p25.zipAnalysis:

1. WTF? It's encoded as hard telecined, but without interlaced coding? Is this a joke? The encoder has "frame_mbs_only_flag" set to zero... and all of the frames are progressively coded.

2. No B-frames?

3. No inter partitions!

4. No CABAC?!

5. No adaptive quantization (but there's slight variance per-row, so it might be per-row ratecontrol).

6. No psy optimizations that bias in favor of inter blocks (e.g. psy-RD or similar).,

7. No weighted prediction.

8. No matter what the QP of the scene, the encoder writes the "pic_init_qp_minus_26" as zero, probably because they're lazy bastards.

9. No scenecut detection.

10. I-frame QP offset is probably 2.

11. The encoder won't go below QP 20 in a P-frame, and correspondingly 18 in an I-frame.

12. The encoder appears to do CBR, or at least something close to it.

13. The encoder's motion search seems to get caught in the common problem where its motion vectors tend to "follow" sharp, straight edges.

Conclusion: this is a joke encoder.

[deekey777]

DS, if you have an HD4000 in your PC, MSE uses ATi's AVIVO converter. I've replaced my HD3800 by my 8600GT, so my file is encoded by the Cuda encoder (with CABAC etc).

[CruNcher]

Huh why is it recognized as Main Profile if all the main stuff is not inside the bitstream (Cabac,B-frames) ? somehow it sounds to me more like Cyberlink is totally doing something wrong with Nvidias and ATis API here implementing everything wrong and the whole input/output logic including flagging is screwed :P

[Dark Shikari]

Huh why is it recognized as Main Profile if all the main stuff is not inside the bitstream (Cabac,B-frames) ?marketing.

[St Devious]

Analysis:

1. WTF? It's encoded as hard telecined, but without interlaced coding? Is this a joke? The encoder has "frame_mbs_only_flag" set to zero... and all of the frames are progressively coded.

2. No B-frames?

3. No inter partitions!

4. No CABAC?!

5. No adaptive quantization (but there's slight variance per-row, so it might be per-row ratecontrol).

6. No psy optimizations that bias in favor of inter blocks (e.g. psy-RD or similar).,

7. No weighted prediction.

8. No matter what the QP of the scene, the encoder writes the "pic_init_qp_minus_26" as zero, probably because they're lazy bastards.

9. No scenecut detection.

10. I-frame QP offset is probably 2.

11. The encoder won't go below QP 20 in a P-frame, and correspondingly 18 in an I-frame.

12. The encoder appears to do CBR, or at least something close to it.

13. The encoder's motion search seems to get caught in the common problem where its motion vectors tend to "follow" sharp, straight edges.

Conclusion: this is a joke encoder.

WOW bro, that's some serious analysis. wonder how you got all that info :goodpost:

[CruNcher]

This is what Nvidia provides them

http://s10.directupload.net/images/090404/anqpx9nw.png

and imho they totaly screw the logic of input/output and flagging in some cases and set the settings wrong in Nvidias Encoder and it wouldn't surprise me if they do the same wrong with ATIs Interface :P

[Dark Shikari]

WOW bro, that's some serious analysis. wonder how you got all that info :goodpost:Elecard Streameye.

[St Devious]

Elecard Streameye.

wow that's one expensive software. does the trial version provide this info ?

[Dark Shikari]

wow that's one expensive software. does the trial version provide this info ?Sure, if you're experienced enough to be able to interpret it all ;). It's worth a download either way though; the interface is easy to use and the visualizations are good.

Trial version is just length-limited AFAIK.

[St Devious]

Sure, if you're experienced enough to be able to interpret it all ;). It's worth a download either way though; the interface is easy to use and the visualizations are good.

Trial version is just length-limited AFAIK.

ya, its 21 days.

need some more info from me with the GPU encode and all ? happy to help.

[St Devious]

Dark and everybody else, Am i using the fastest settings possible for x264 ? aside from setting main prfofile and level 3.0.

If yes, wouldn't you say that at least GPU encode is faster for almost the same quality ?

[CruNcher]

ehhh do this --scenecut -1 then it would be a little more fair then now and stay with baseline as the ATI Encode is baseline too MSE just flagged it wrong why ever :P

[St Devious]

ehhh do this --scenecut -1 then it would be a little more fair then now and stay with baseline as the ATI Encode is baseline too MSE just flagged it wrong why ever :P

what does --scenecut -1 do ?

do i manually write in the commandline below? or can i select it from gui options in megui ?

[CruNcher]

disables scenecut (faster) like it is setup in MSE for the ATI Encode also you can gain more speed by setting --ref 1 instead of 2, dunno if MeGUI has option to disable scenecut

[St Devious]

disables scenecut (faster) like it is setup in MSE for the ATI Encode also you can gain more speed by setting --ref 1 instead of 2, dunno if MeGUI has option to disable scenecut

MSE uses 2 ref frames, so i wanted to use similar settings.