Hi!,I want to know wether the reference software HM-1.0 ( https://hevc.hhi.fraunhofer.de/svn/svn_TMuCSoftware/tags/HM-1.0 ) is as the same as 0.9-hm ( https://hevc.hhi.fraunhofer.de/svn/svn_TMuCSoftware/branches/0.9-hm/). Because I find their README-changes.txt files are the same.

Of course, they are different.

Don’t take README-changes.txt as a serious thing. They don’t change README-changes.txt often.

Could you tell something about the 3 different types of config?
In HM, there are Intra, Random access and Low delay config files.
If I only want to test the new encoding structure’s gain, comparing with H.264, which can be best chosen as a candidate config?
Thank you very much.

Why hevc choose C++, i guess a lot of guys here hate C++.

Thanks, Dr. Liu.
BTW: The software manual is available in JCT-VC D404

Hello
My name is Mohammed ,
Can any one please send me a help to use the software?
best regards

In HM, do they use LCEC or just VLC?
Also, I cannot find the detail document about LCEC.
TENTM suggested this one?

Dr. Liu,
I am new in this area. Could you tell me how to start using this softare?
I need to compare some new features with h.264 ones.
Thank you.

could you tell how to config for grayscale mode?

I typed this command:
tappencoder -c basketballpass.cfg -q 32 -g 8 -f 9 -s 64 -h 4

I received this error:
HM software: Encoder Version [2.0][Windows][VS 1600][32 bit]
Error: QuadtreeTULog2MaxSize must be 5 or smaller.

How can I correct it?

I want to ask that if any one knows how to run HM-1.0 code in Ubuntu?
I am facing problem regarding how ti include .cfg file in ubuntu. Like in windows it is done in visual studio by using properties. how it can be done in ubuntu.

You should change the value of QuadtreeTULog2MaxSize in basketballpass.cfg.

Hi Jdong
I changed QuadtreeTULog2MaxSize from 6u to 5u and it was corrected.

I typed this command:
tappencoder -c basketballpass.cfg

I received this error:
Assertion failed: uiNumberOfBits <= 32, file ….\source\lib\tlibcommon\tcombitstream.cpp, line 73

How can I correct it?

In HM-2.1 version I received this error:
Assertion failed: uiNumberOfBits <= 32, file ….\source\lib\tlibcommon\tcombitstream.cpp, line 74

Hi everybody. please tell me that all the input files mentioned in the configuration files are not present in the directory. even there is no directory in which in the input .YUV files must present. please tell me what to do with this issue. although i have downloaded some other YUV file from internet and replaced the name of the input file in the configuration file but, for it i have to make so many changes because, the source and height would be different and also i have to change a lot of things.

regards,
ASHFAQ AHMED

Dear Ashfaq Ahmed

ftp://hvc:US88Hula@ftp.tnt.uni-hannover.de/testsequences

I get the same error as mma238 posted earlier :
Assertion failed: (uiNumberOfBits <= 32), function write, file …../source/Lib/TLibCommon/TComBitStream.cpp, line 74.
Can anyone please advise?

To mma238 and Tamer Sherif:
Please use the following command to execute the encoder:
TAppEncoder -c encoder_lowdelay.cfg -c basketballpass.cfg

Please note that there are two kinds of config files, one for general configuration, the other for configuration per sequence.

For general configuration, there are several files for different applications, such as encoder_intra.cfg/encoder_intra_loco.cfg/encoder_lowdelay.cfg/encoder_lowdelay_loco.cfg/encoder_randomaccess.cfg/encoder_randomaccess_loco.cfg, which are located in HM-x.x\cfg\.

For configuration file per sequence, they are located in HM-x.x\cfg\per-sequence\.

Dear Yu Liu
Thank you very much
It created these two files:
str.bin & rec.yuv
What is rec.yuv and why is it larger than original file?

str.bin is the encoded bitstream file, and rec.yuv is the reconstructed video sequence. The increasing of rec.yuv may be caused by 10 bits output, instead of 8 bits.

Hi Yu Liu,

when I run the latest version of HEVC which is HM 2.2, the rec.yuv file is a video of which all frames turns to be green , do u know why ?

Hi Yu Liu,

I want to make some complement, do you know how to see the decoded result of rec.yuv file, I use YUVviewerPlus with 4:2:0 mode, but it turned out to be all green frames…..

How can I convert an image to YUV 4:2:0 format?

In order to see the rec.yuv from encoder in 8 bits 4:2:0 mode, it is better to change the InternalBitDepth = 10 in encoder_xxx.cfg to InternalBitDepth = 8. However, the change may cause the decreasing of coding performance because of the disable of IBDI.

If you want to enable the IBDI, you can keep InternalBitDepth = 10. In this case, the rec.yuv from encoder is 10 bits video. However, you can use the decoder to generate the 8 bits 4:2:0 reconstructed video:
TAppDecoder.exe -b str.bin -o dec.yuv -d 8

Oh, thank you, my problem is solved by following your instruction !

Hi Yu Liu,

I run into another problem. When I test anther video (newspaper004.yuv) and change some parameters of encoder_randomaccess.cfg as following:

resolution: 1024×768
length: 300 frms
QPs: 35
IntraPeriod: 16
GOPsize: 16
RateGOPSize: 16
SearchRange: 96

the resulting decoded dec.yuv file only has one frame. I do not know why? since when I use the original default encoder_randomaccess.cfg file, the problem does not happen .

p.s. when I run the decoder, the command is showed as follows
TAppDecoder.exe -b str.bin -o dec.yuv -d 8

I want to ask a question, does the website have a chinese edition?

no, only English version. However, you can use the “Translate it” at the top-right corner to translate the webpage to your desired language.

Thank you very much, Mr Liu, it’s very nice of you.
I have another problem.I have been studying the HM software and its edition is 2.0 in which the complex structure and too long variable makes me puzzled. Can you tell me how to study the HM, and is there the help document or other file can solve my problem , which, I think is others’problem.
Thank you for replying.

Can anyone help me ? I got the following error
HM software: Encoder Version [2.0][Windows][VS 1600][32 bit]
Error: QuadtreeTULog2MaxSize must be 5 or smaller
I tried to change it the max size in racehorces.cfg but I couldn’t find any input as this?

Dear Sir,

I could build the code in Windows and both encoder & decoder working fine.
However, if I try to build code unders linux using cygwin,by running command :
$ make
I get error.
So, please let me know:
1. Can we build code using cygwin?
2. Does it require any other command line parameters in make command?

Thanks & Regards,
Dr. Ganesh Bhokare

Does anyone know how to compile the software HM2.0 in ubuntu?I just type command: $make.But it doesn’t work.

Hi,

I have a question, when syntax of bitstream in reference software and in working draft will be unified? I need to save in text file all syntax elements of bitstream, but it is hard to do in present reference softwere. It was much more easy to do in reference sofwere to h.264 standard…?

Thanks for any replying.

Hello, how i can lauch it? i have downloaded HM3.2rc1 and how i execute aplication? or i must build it first in some compiler?

Why I cannot to download HM any version from repository?
I installed TortoiseSVN, made all following instruction, but it says “couldn’t connect to repository”.
Tell me please, how is it possible to download HM.

to John Ogah::
You can change it in cfg files which are in cfg folder, not in the cfg/per-sequence one.

to doris:
Is your network using proxy? If you do this, you should edit files of svn.

Compilation fails for me in Linux unless I remove “-Werror” from CPPFLAGS in build/linux/common/makefile.base

Hi ppl,
I am trying to implement hm 4.0 on ti ccs. Can somebody help me with what files to include in the project. I am getting a linker error.

Thank you in advance.

Hi Dr. Liu,

I’am using the software HM-2.0. After make I typed this command:
./TAppEncoderStatic -c ../cfg/encoder_randomaccess.cfg -fr 30 -f 10 -i Flowervase_416x240_30.yuv
and got the following error:
POC 0 ( I-SLICE, QP 32 ) 208 bits [Y 99.9900 dB U 99.9900 dB V 99.9900 dB] [ET 0 ] [L0 ] [L1 ] Segmentation fault

Could you help me, please?

Thank you in advance.

Hello!
I test the code and find out that HEVC’s compression ratio is good ,but the coding speed is too slow. I costed 1.76h to encode 300 frames which is YUV420 352*288 ,using encoder_lowdelay_loco.cfg in the project. I want use HEVC in real-time business. How to make HEVC’s encoding speed more faster, which parameter should i set？ Please help me .

Mars,
You have to wait until the HVEC codec is finalized and optimization has completed, this code is proof of concept only. For real time performance it will take additional time for a hardware acclerated encoder to be implemented.
For the foreseable future you really have to look into hardware encoded h264 for good quality at real time performance…

I can’t download the codec, where is the link?

Hi, I’ve been testing HEVC (HM4.1) and it produces expected results with most test sequences using high efficiency and random access. But for sequences with small number of frames (like one view of BookArrival with 100 frames 1024×768) it produces a similar bit rate when compared to H.264/AVC (just 4% ~ 20% less)… Specially for higher QPs (which is understandable for QP). Does anyone know why the results are so inferior for these small frame number sequences? I think something like this would be expected for low resolution images not low frame number sequences…

I want to learn about h265, please instruct on how to download the software…
It would be really helpful
thanks

Does anyone know how to use constant bit rate (or control rate like H.264) in HM4.0 encoder.
THX
duo1313@gmail.com

Current HEVC test model does not have a rate control algorithm integrated.

I have tried to convert 14 bit grey scale images(YUV 4:0:0) which are scaled up to 16 bit to a video sequence by HEVC Test Model Reference Software, in lossless mode but the output file is not lossless.(I have done it for 8 bit images without any problem). Any help would be appreciated.Thx.