Hi every one,
I am experimenting with the HEVC encoder, HM 14. I downloaded video sequences from this link ftp://pcs:email@example.com/ and trimmed the sequence in virtual dub to the number of frames I wanted for the experiment. In virtual dub the raw video sequences only display in I420 pixel type. I am aware that I420 is not YUV but that is the only format that displayed the video. I tried to encode the processed video sequences from virtual dub in the HEVC reference software, HM 14, but the PSNR values were very poor, between 24 and 30 dB. Please what steps am I missing out in this test.
+ Reply to Thread
Results 1 to 4 of 4
In the full-processing mode, you can set VirtualDub to convert the original video sequence to the desired/required output format.
As for the HM14 application: you know that it is a "reference" encoder, NOT an *optimized* encoder, right?
Also, you didn't tell us what were the compression settings you chose for the HM14 encoder.
Hopefully someone who has ever used it will chime in.
Last edited by El Heggunte; 31st Oct 2014 at 03:46. Reason: clarity
Hi El Heggunte,
I followed your direction and adjusted the format of the processed video in virtual dub to 420 planar YCbCr (YV12). I also set the frame rate to the desired value. The encoding turned out fine. The HEVC .bin file was same quality as the .yuv files rendered in yuv player. There are few color bars at the top in both the .bin as well as the .yuv files; and the color of the .bin and the .yuv videos that resulted from the encoding is not as good as the original video sequence in virtual dub but the PSNR is good; maybe I need to chose another color format.
Thank you very much; you saved me lots of time. I am open to receive further help.
The configuration I am using for the experiments are the default settings of the random access configurations.
Hi El Heggunte,
I found out that each of those video sequences has a configuration file prepared using visual studio. I used these files and adjusted the entry for the number of frames to the number of frames I desired, then I saved it. Then I opened visual studio and set up TappEncder as startup project. Under the property tab under debugging, I used the following settings:
-c C:\...RandomAccess_main.cfg -c C:\...\..cfg -i C:\...\.yuv --SEIDecodedPictureHash=1 -b C:\...\.bin -o C:\...\.yuv
All the output files came out perfect as the original video sequence. Those configurations files can be used to trim the videos especially if you are feeding from the first frame to any number of frames.
Thanks my Friend you gave me a good startup in my work, I appreciate it alot.