Youtube analysis of fully automated video production methods
Recently, I participated in a Malaysian Youtuber course training with deep emotion. The whole teaching experience is not bad. The other party mentioned in the class that many people in mainland China use technology to operate Youtube, and said that it is not recommended for beginners to operate. I talked with him in private about the specific operation method, but he still didn’t want to disclose too much, he also said that he started teaching in order to let more people create high-quality media content, and I also operated Youtube myself. I understand that he mentioned in class that this operation is not recommended. Some time ago, Li pumpkin, a project observer I followed in Zhihu, also shared his article on using Python to operate Youtube in batches. The whole content is very detailed, unfortunately, when I planned to write this article to look for some available materials yesterday, I found on his Zhihu that he had deleted his column and all the operational project information he had written., unfortunately, these signs also drove me to write this article.
Fortunately, I found his article through Google cache 《 How to use PYTHON to build a fully automatic and profitable YOUTUBE video publishing system with a monthly income of over 10,000 (passive revenue) “, I will attach the detailed information of this article at the bottom of this article. If you are interested, you can directly jump to the end to read.
As a matter of fact, I have learned about some channels Youtube automatic operation since the end of 2016, but at that time I focused on Adsense and did not study them too carefully, that is to say, the automatic operation of Youtube has appeared at least before 2016. In this article, I will explain the recently discovered methods about Youtube to make money automatically, I hope to give you some inspiration on Youtube operational thinking. You need Learn to think by yourself and don’t blindly follow suit , the road of making money online has never been short of every imitator of us. If we want to be outstanding, we have to go Learning others’ thinking advantages , this is the foundation of making money successfully.
Youtube the advantages, conditions, and risks of fully automated video production operations?
The most important advantage of automatic video production is You do not need to manually edit videos. , give the editing and splicing to the program to complete, Can produce a large number of videos simultaneously , and The video does not need to be shot or dubbed by a real person. , it’s OK Use API operations to upload videos in batches For example, this topic describes how to use Youtube API to upload videos in batches.
As we all know, Youtube need to pass the audit before they can make profits from the operation, and the standard to reach the audit is 4K hourly views and 1K subscribers At the same time, if this condition is met, it is neither difficult nor simple. After all, after meeting this condition, Youtube need to check whether your content is suitable for profit.
There are many solutions about the number of broadcasts and subscriptions online. If you are in a hurry to try, you can try to search for keywords. youtube subscribers free hit click points , some platforms may have the risk of sealing accounts, so I will not recommend them to you. Those who like gold coins can also directly purchase accounts or channels that have already made profits through some channels.
1. Own its own profit Channel
2. Familiar with at least one script programming language and database language. Python,Nodejs, and Java are recommended here.
3. Suitable niche, suitable niche can be understood as Youtube category here, not all categories are suitable for automatic processing.
Automated video processing is suitable for simple scenarios, such as image splicing and simple video splicing. I have listed the types of videos that can be automatically produced. You can continue to look at them.
As a media person, we should know that people’s copyright awareness has been much better than that of many years, while audio and video are almost in the minefield of copyright. If you are not careful about your audio, the video may infringe upon others’ copyright, it brings troubles to you. You need to think it over before making videos and try to avoid using copyrighted audio and video. Another point must be mentioned here is that Youtube is a relatively free and open platform, so you may have seen some videos that violate national laws and regulations, especially some military and political channels, I hope every media person who belongs to Chinese Youtube can understand that there is no home without a country. We have the blood of Chinese children, and every Chinese child should Support the unity and integrity of our country, do not spread rumors, do not spread rumors, and abide by laws and regulations .
What types of channels are suitable for Youtube fully automated video production?
1. Feelings of heart/quotations of chicken soup
2. Current politics/weapon military
3. Music mixing DJ/MP3
4. There may be some types that I haven’t found yet
You probably have already made a clue above, or you have already thought of a domestic writing platform, today’s headlines, Baijia, you must have seen the type of articles on the above 1 and 2 on these two platforms, many articles on the headline and Baijia are generated in batches, which is not very valuable, but there will always be people reading them. Even the titles and covers are attractive, and it is easy to be popular. Since the text content has such “junk” content, can it be made into audio and video? This is what I want to share here. It is nothing more than that these contents are messed up. Or transcribe audio and video into text content, or make audio and video from text content.
And the main method of revealing today is The text in the popular article is synthesized into audio MP3 by voice, then the pictures and MP3 are added with subtitles to synthesize videos, and then the videos are uploaded to Youtube in batches. , why do we choose the tubing instead of other platforms? First of all, the video generated in this way is very fast and has a large number of operations. Manual editing and uploading are meaningless. After all, it is too slow to win by quantity. Currently, among many platforms in China, only Toutiao APIs can be uploaded and released. However, it is said that Toutiao has high requirements for qualification application. Other video platforms such as iQiyi need to simulate uploading, which makes it more difficult, if you are interested, you can use the simulation Upload scheme of Li pumpkin at the bottom of this article.
Does anyone watch videos produced in full automation Youtube?
I can tell you this question as long as your title and cover are attractive enough, you will not lack viewers, here, I will show you the background screenshot of one of my discarded headlines:
At that time, this video was derived from the idea that I had contacted the automatic video channel on the Youtube at the end of 2016 as I mentioned earlier. I made a simple statistics on its video according to the Youtube channel at that time, analyze the popular videos, check the video content, and then use PR to manually create the video. In fact, this video can be automatically operated, since I uploaded this video to today’s headlines, I have recommended a total of 450,000 times to observe 11,000 times, which means the playback rate has reached 24%. On average, 24 people watch every 100 recommended users, but if you watch it, you will know how low-end this video is: https://www.ixigua.com/6461904426717151757/
Therefore, if it is a full-automatic production achievement, you don’t have to worry too much about whether the video is popular or not. It depends on quantity to win. However, this article mainly focuses on revealing the secret, everyone of us should know that audio and video will have great copyright risks, which is very similar to our farm website. So is there really such a case in the Youtube? Yes, keep looking down.
What are the Youtube cases of fully automated video channel production?
Youtube I have found some cases of automatic production of video channels, some of which may not be fully automated, but videos in these cases can be automatically generated in batches through code. Let’s take a look at these channels first!
English Channel: https://www.youtube.com/c/BreakingNewsToday24h/featured , it is a bit similar to news and current events. It does not make any sense for a video with thousands of broadcasts, but if you have hundreds of videos posted every day, you will know how horrible the total traffic is. According to the observed video channel, one video has reached 50,000 broadcasts in two days, and the whole channel has reached 220,000 broadcasts in the past month, that is to say, there are nearly 8000 broadcasts per day on average.
Due to various factors, the income of a video on a Youtube is between 120 dollars per thousand plays. According to the lowest calculation, we earn 7 dollars per day for nearly 8000 plays per day, 30 days a month, 2,000 dollars a month, which is calculated according to the lowest standard. Even if the income is only $500 a month, if more channels are operating synchronously, the accumulated income will be considerable.
Youtube money-making skills sharing how can I earn $2000 in the first month of Youtube?
Other cases of automated Youtube channels:
- https://www.youtube.com/channel/UCvYes_0afVM6GeTTFzELW4Q
- https://www.youtube.com/channel/UCU2t_Mq9F97eWHJfgBom_wg
- https://www.youtube.com/channel/UCWaPjZJ8ZTgZ0AwxoKQ-8Mg
- https://www.youtube.com/channel/UCiqT2W41juVjd1lPSfxgCsg
- https://www.youtube.com/channel/UC0zGScLEqQ_ggw23iIBIZoA
Youtube the procedure for producing videos automatically?
Video content production
This part is mainly obtained from major news and media platforms, such as domestic headlines today, Baijia, WeChat public accounts, etc. In the early stage, a large number of popular article materials can be sorted by crawlers for backup, if you feel that it is difficult to obtain statistics from various platforms, here are some tools to view hot articles:
- Watermelon data: http://www.xiguaji.com/ More than one million articles are updated every day, covering various industries and vertical fields, hot content and high-quality original articles, helping you quickly find or create high-quality content.
- Clear Big data: http://www.gsdata.cn/ Ranking of Weibo, WeChat, headlines, KOL and other data. Follow the steps of large size.
- Popular article Pavilion: https://www.yibaowen.com/ A promo platform, which is designed to empower the creation of self-media, focuses on the data mining and analysis of promo data integration of various popular articles, videos, Atlas and other materials on various self-media platforms, at the same time, it provides timely capture and analysis of explosive articles.
- New list: https://www.newrank.cn/ Data platform for new media public account ranking, advertising promotion, vertical marketing and other fields.
Content speech synthesis
We recommend that you use Baidu speech synthesis. In addition, xunfei and bing also provide speech synthesis technology. For details, you can go to their official website to see the API document and directly extract the fragments of the article you sorted out in the previous step, you can directly use the speech synthesis interface provided by the above manufacturer to generate MP3 documents.
The following example uses Baidu speech synthesis API:
This section does not include the developer registration procedures for Baidu AI open platform.
Registered address of Baidu developer open platform: https://passport.baidu.com/v2 /? Reg
If you can’t handle this, this article may not be suitable for you.
The following example uses Python:
# coding=utf-8
import sys
import json
# 保证兼容python2以及python3
IS_PY3 = sys.version_info.major == 3
if IS_PY3:
from urllib.request import urlopen
from urllib.request import Request
from urllib.error import URLError
from urllib.parse import urlencode
from urllib.parse import quote_plus
else:
import urllib2
from urllib import quote_plus
from urllib2 import urlopen
from urllib2 import Request
from urllib2 import URLError
from urllib import urlencode
# 替换你的 API_KEY
API_KEY = 'nu9r2plGFi3s1ugayDPSM6Mk'
# 替换你的 SECRET_KEY
SECRET_KEY = 'G62YGnq84eKTqu0mBgvdpmC6gNBzHdai'
# 大姚的订单信息内容文本
TEXT = "三分钟前，由北京市顺义区二经路与二纬路交汇处北侧，北京首都国际机场T3航站楼 去往 东城区北三环东路36号喜来登大酒店(北京金隅店)"
TTS_URL = 'http://tsn.baidu.com/text2audio'
""" TOKEN start """
TOKEN_URL = 'http://openapi.baidu.com/oauth/2.0/token'
"""
获取token
"""
def fetch_token():
params = {'grant_type': 'client_credentials',
'client_id': API_KEY,
'client_secret': SECRET_KEY}
post_data = urlencode(params)
if (IS_PY3):
post_data = post_data.encode('utf-8')
req = Request(TOKEN_URL, post_data)
try:
f = urlopen(req, timeout=5)
result_str = f.read()
except URLError as err:
print('token http response http code : ' + str(err.code))
result_str = err.read()
if (IS_PY3):
result_str = result_str.decode()
result = json.loads(result_str)
if ('access_token' in result.keys() and 'scope' in result.keys()):
if not 'audio_tts_post' in result['scope'].split(' '):
print ('please ensure has check the tts ability')
exit()
return result['access_token']
else:
print ('please overwrite the correct API_KEY and SECRET_KEY')
exit()
""" TOKEN end """
if __name__ == '__main__':
token = fetch_token()
tex = quote_plus(TEXT) # 此处TEXT需要两次urlencode
params = {'tok': token, 'tex': tex, 'cuid': "quickstart",
'lan': 'zh', 'ctp': 1} # lan ctp 固定参数
data = urlencode(params)
req = Request(TTS_URL, data.encode('utf-8'))
has_error = False
try:
f = urlopen(req)
result_str = f.read()
headers = dict((name.lower(), value) for name, value in f.headers.items())
has_error = ('content-type' not in headers.keys() or headers['content-type'].find('audio/') < 0)
except URLError as err:
print('http response http code : ' + str(err.code))
result_str = err.read()
has_error = True
save_file = "error.txt" if has_error else u'大姚的订单信息.mp3'
with open(save_file, 'wb') as of:
of.write(result_str)
if has_error:
if (IS_PY3):
result_str = str(result_str, 'utf-8')
print("tts api error:" + result_str)
print("file saved as : " + save_file)
In fact, it is a process of sending requests and receiving results, which can be completed in other languages. The above code is from Baidu official. If you want to use other languages, please refer to the official website documentation of Baidu developers: https://ai.baidu.com/ai-doc/SPEECH/7k4nv3ngr
Audio and image production videos
The synthesized video mainly uses FFMPEG( Fast Forward Mpeg) as an open-source computer program that can be used to record, convert digital audio and video, and convert them into streams. FFMPEG is a very powerful tool, which is commonly used in audio and video encoding, decoding, editing, and other functions such as video splicing, video slicing, video snapshot, and video watermarking. For the specific parameters of FFMPEG, please refer to: Ruan Yifeng’s blog FFmpeg video processing tutorial .
FFMPEG has many ways to play, but it is not limited to what is mentioned in this article. You can use it to generate Gif, which is also a dynamic image generation tool on the internet. You can also use Gif images to generate video clips and so on.
FFmpeg command format: ffmpeg [全局选项] {[输入文件选项] -i 输入_url_地址} ... {[输出文件选项] 输出_url_地址} ...
The command to produce audio MP3 and image into a video is as follows: ffmpeg -loop 1 -i ./1.jpg -i ./1.mp3 -absf aac_adtstoasc -s 1920*1080 -r 8 -c:v libx264 -x264-params keyint=1:scenecut=0 -c:a aac -b:a 64k -pix_fmt yuvj420p -t 20.00 ./out.mp4
The following code provides an example on how to combine multiple images and audio files into a video:
import subprocess
import time
imglist='1812190119_47_xh_img.txt'
mp3='1812190119_47_xh.mp3'
mp4='1812190119_47_xh_q1.mp4'
start = time.perf_counter()
cmd = 'ffmpeg -threads 5 -y -f concat -safe 0 -i ' + imglist + ' -i ' + mp3
+ ' -vcodec libx264 -pix_fmt yuv420p ' + mp4
# p = subprocess.call(cmd, shell=True, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
# p=subprocess.call(cmd,shell=True, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
p = subprocess.Popen(cmd, shell=True, universal_newlines=True, stdin=subprocess.PIPE, stdout=subprocess.PIPE,stderr=subprocess.STDOUT)
# p = subprocess.call(cmd)
end = time.perf_counter()
print(end-start)
imglist content format (time, and the form of the last image)
file ./1812192159_49_xh/1812192159_49_xh_1.png duration 0.56 file ./1812192159_49_xh/1812192159_49_xh_10_1.png duration 1.2 file ./1812192159_49_xh/1812192159_49_xh_10_1.png
The above is just the general code. In practice, you may need to consider how long to change an image. Here I will directly explain the operation process of others. There are two types of MP4 videos, one is that the image loop appears in a video, for example, four images in an article. Set the time when the images are converted and changed. After the four images are used up, repeat the loop again. The average audio time is divided by images. For example, if the audio is 4 minutes, the whole video is actually 4 minutes, and 4 pictures are 4 frames. If one picture is displayed for one minute, the frame speed can be calculated.
Next is the image + audio synthesis MP4, where MP4 has no subtitles. The method parameter comments are clearly written, and many parameters can be changed by yourself.
This operator first combines images into videos and then audio. I suggest using the imglist above to directly process images and audio in one step.
Subtitle processing
Subtitle processing needs to be performed when the text is generated into MP3. subtitle files are generated when the text is generated into MP3. after subtitle files are created, they can be directly mounted to the video.
Automatically upload videos
When uploading a video, you need to generate keywords for the text of the video content. You can search for the keyword extraction method yourself. You can also extract two keywords from similar channel videos into a list.