Closed vedanthkadam555 closed 9 months ago
But in that tutorial there is already software installed . I don't have much program knowledge I'm struggling to install it. Thanks for Response.
https://github.com/jianchang512/pyvideotrans/releases/download/v0.9982/videotrans-v0.9982.7z
On Windows, download and unzip, double-click sp.exe
After decompressing, as above, double-click spxe to complete the process Only applicable on Windows, other systems require source code installation
If Mac or Linux
pip install -r requirements.txt
, if version conflict error occurred, please executable pip install -r requirements.txt --no-deps
python sp.py
.Oh I see. Then I guess it doesn't need any programing installation. As it was showing in Chinese tutorial that. We need to install Pycharm and all that. Okay Got it. Thank You Ma'am : )
Please Please Add the Hindi language in Source file. So that I can convert that into English. Please I'm waiting for that. Thank you.
v0.9982 Supports the use of Hindi as the original and target languages
I selected Hindi language. But while the video is generating. Beside it showing captions and those captions are in Urdu Language, Not Hindi.
Even after completing the video process. The output audio is not even near to original audio. It's just saying random words which are in the original audio. Thanks
Look this is Urdu language
"The weather is nice today" Translate this period into Hindi And provide a link to the translation results of Google Translate
"The weather is nice today" Translate this period into Hindi And provide a link to the translation results of Google Translate
看起来没什么不对啊
this is my transate result
1
00:00:00,000 --> 00:00:01,280
यह उपवन में है
2
00:00:01,280 --> 00:00:04,600
सबसे पहले हमें एक इलेक्ट्रोमैग्नेटिक कमरा मिलता है
3
00:00:04,600 --> 00:00:08,440
यहाँ तीन हैं
4
00:00:08,440 --> 00:00:10,240
यहाँ दो खुरच हैं
5
00:00:10,240 --> 00:00:11,520
यहाँ एक मिट्टी है
6
00:00:11,520 --> 00:00:12,560
फिर देखें
7
00:00:14,600 --> 00:00:16,079
नीचे है
8
00:00:16,079 --> 00:00:17,080
आरपीएस
9
00:00:17,080 --> 00:00:20,480
सब देखें, वास्तव में यह सर्किट क्या करता है
10
00:00:20,480 --> 00:00:22,520
इस चक्की द्वारा
11
00:00:22,520 --> 00:00:23,800
यह क्रिया है
12
00:00:23,800 --> 00:00:24,760
उचित लंबाई तुलना
13
00:00:24,760 --> 00:00:26,120
तुम इसी तरह समझते हो तो ठीक है
14
00:00:26,120 --> 00:00:26,640
सही है ना
15
00:00:26,640 --> 00:00:27,920
सब देखो
16
00:00:27,920 --> 00:00:29,360
यह AD है
17
00:00:29,360 --> 00:00:29,960
सब देखो
this is translation to English
1
00:00:00,000 --> 00:00:01,280
it is in the garden
2
00:00:01,280 --> 00:00:04,600
First we get an electromagnetic room
3
00:00:04,600 --> 00:00:08,440
here are three
4
00:00:08,440 --> 00:00:10,240
here are two scratches
5
00:00:10,240 --> 00:00:11,520
there is a soil here
6
00:00:11,520 --> 00:00:12,560
see again
7
00:00:14,600 --> 00:00:16,079
is below
8
00:00:16,079 --> 00:00:17,080
rps
9
00:00:17,080 --> 00:00:20,480
See all, what this circuit actually does
10
00:00:20,480 --> 00:00:22,520
by this mill
11
00:00:22,520 --> 00:00:23,800
this is action
12
00:00:23,800 --> 00:00:24,760
proper length comparison
13
00:00:24,760 --> 00:00:26,120
It's okay if you understand it that way
14
00:00:26,120 --> 00:00:26,640
that's right
15
00:00:26,640 --> 00:00:27,920
See all
16
00:00:27,920 --> 00:00:29,360
This is AD
17
00:00:29,360 --> 00:00:29,960
See all
its seems no error
Chinese
1
00:00:00,000 --> 00:00:01,280
这发的公园里
2
00:00:01,280 --> 00:00:04,600
首先我们拿到的是一款电磁房
3
00:00:04,600 --> 00:00:08,440
这上面有这里有三个
4
00:00:08,440 --> 00:00:10,240
这里有两个孔
5
00:00:10,240 --> 00:00:11,520
这里有一个土
6
00:00:11,520 --> 00:00:12,560
再看一下
7
00:00:14,600 --> 00:00:16,079
下面是
8
00:00:16,079 --> 00:00:17,080
RPS
9
00:00:17,080 --> 00:00:20,480
大家看一下其实这个电路它就是通过什么
10
00:00:20,480 --> 00:00:22,520
通过这个线圈
11
00:00:22,520 --> 00:00:23,800
动作就是
12
00:00:23,800 --> 00:00:24,760
常态变长比
13
00:00:24,760 --> 00:00:26,120
你就这样认为就可以了
14
00:00:26,120 --> 00:00:26,640
对不对
15
00:00:26,640 --> 00:00:27,920
大家看一下
16
00:00:27,920 --> 00:00:29,360
这个是AD
17
00:00:29,360 --> 00:00:29,960
大家看一下
Hidin
1
00:00:00,000 --> 00:00:02,760
यह उपवन में है
2
00:00:02,760 --> 00:00:07,944
सबसे पहले हमें एक इलेक्ट्रोमैग्नेटिक कमरा मिलता है
3
00:00:07,944 --> 00:00:11,784
यहाँ तीन हैं
4
00:00:11,784 --> 00:00:14,736
यहाँ दो खुरच हैं
5
00:00:14,736 --> 00:00:17,736
यहाँ एक मिट्टी है
6
00:00:17,736 --> 00:00:19,632
फिर देखें
7
00:00:21,672 --> 00:00:23,928
नीचे है
8
00:00:23,928 --> 00:00:25,704
आरपीएस
9
00:00:25,704 --> 00:00:30,672
सब देखें, वास्तव में यह सर्किट क्या करता है
10
00:00:30,672 --> 00:00:32,904
इस चक्की द्वारा
11
00:00:32,904 --> 00:00:34,968
यह क्रिया है
12
00:00:34,968 --> 00:00:37,416
उचित लंबाई तुलना
13
00:00:37,416 --> 00:00:41,568
तुम इसी तरह समझते हो तो ठीक है
14
00:00:41,568 --> 00:00:43,536
सही है ना
15
00:00:43,536 --> 00:00:45,456
सब देखो
16
00:00:45,456 --> 00:00:48,072
यह AD है
17
00:00:48,072 --> 00:00:49,992
सब देखो
https://github.com/jianchang512/pyvideotrans/assets/3378335/ce372564-08a1-4ffe-9fbb-e5872f49ecae
I didn't find any errors, according to Google translation, this is indeed Hindi
Can you translate this Hindi video into English.
I selected source file as Hindi. But it is showing Urdu Language here.
It is translating Urdu to English.
Instead it should translate Hindi to English
These two languages only have different writing styles, but their pronunciations are similar and difficult to accurately recognize. Please try using the large-v3 model, as the effect should be better
You're right. Pronunciation is Similar.
And also using ElevenLabs in TTS is unlimited? Because yesterday I have put a Script in TTS by selecting evenlabs. And the script was 2 hours length. I kept it while night. But I didn't received the output.
The software itself has no restrictions, and if there are restrictions, it should also be an issue with the API. If it does not meet expectations, please check if there is any error message in the logs directory and search for the word "error"
But it is working with Edgetts. It gave 2 hours audio file in 10 Minutes.
I have found this in logs folder.
I don't quite understand. Are you translating videos or synthesizing text into speech? If you choose to use elevenlabs TTS, then the speaker should not be edgeTTS You can delete the log information and try again. If there are still issues, you can record a video of the operation
Ma'am I'm using Text to Speech synthesis.
Is it a free account or a paid account? If it is free, there are restrictions
My elevenlabs account is free account.
Chatgpt Translation also not working
No error message visible, please check for errors. There may be reasons for errors when not in use
It seems that you filled in the wrong chatGPT API address
And for Chatgpt also we need Paid Subscription? To use that?
ChatGPT free account is basically unavailable, only allowing 3 requests per minute as API
I will buy paid GPT account. Only if it works well. If you have that. Can you please try the video link which I sent you earlier.
large-v3 model
Hindi
1
00:00:00,000 --> 00:00:30,000
तो फाइनली बिडो कूमा ने अपना पंच सैटन पर चड़ी थिया, और सही में ये पंच लिटरली काफी जादा सेटिस्फाइंग था, बस एक आदा और चड़ जाता ना, तो पूरा कूमा का दर्द मेरे अंदर से भी निकल जाता, बट हम इतने में भी खुश हैं, क्योंकि
2
00:00:30,000 --> 00:01:00,000
जैसे की कूमा में सेल्फ डिस्टक्शन मेकनिज्म डाला ही नहीं गया, लेकिन जो मेकनिज्म डाला गया है कूमा में, उसके बापजूद भी कूमा जिन्दा कैसे है, एंड पनी तक कैसे पहुँचा है, जिनको हम इस वीडियो में जानेंगे, तो चेक्टर 1104 की सुरुवा
3
00:01:00,000 --> 00:01:30,000
दिया था, हे वेगा पंग, आखिर ये क्या सीन है, लेकिन विचारे वेगा पंग के भी, कूमा को देखकर खोब शुरुवा होते हैं, एंड इतने में ही कूमा अपना एक ही पंच सैटन पर बरसा देता है, जिससे उसकी रेल बन जाती है, और ये रेल जाके टकराती है वहा
4
00:01:30,000 --> 00:02:00,000
मूव कर पा रहे होते हैं, जिससे कई बनी को बचाने भागते हैं, एंड डॉक्टर वेगा पंग तसली से बैट कर, सैटन द्वारा पुझे गए सिनेरियोस के वारे में, कि आखिर कूमा जिन्दा कैसे है और यहां तक कैसे पूचा है, अपनी खटेवी कोपली से सोचने ल
5
00:02:00,000 --> 00:02:30,000
अक्टिवेट किया था, जिसकी वज़े से वो किसी भी ओर्डर को रिसीप करने में इंकैपेबल बन जाता है, लेकिन फिर भी डॉक्टर वेगा पंग की खोपडी, कूमा के यहां आने का रीजन नहीं खोच पाती, लेकिन अचाने की उसे कूमा के ब्रेड लाइन से लिटे
6
00:02:30,000 --> 00:03:00,000
वो एजिली अपनी पावर से रीजनेट कर लेता है, और सैटन की इस पावर के बारे में भी अपने को कोई आइडिया नहीं था, और वो अगेन वही अटेक्ट यूज़ करता है, जिसे उसने लास्टली बनी पर यूज़ किया था, लेकिन इस पॉइंट अफ टाइम पर कू
7
00:03:00,000 --> 00:03:30,000
जो कि सैटन के बोड़ी के आरपार चला जाता है, लेकिन जैसे कि अपने वो पता है कि सैटन अप रीजन नेट भी कर सकता है, तो इस अटेक्ट का कोई फायदा नहीं होता, लेकिन वहाँ कि जारू फ्रैंक के पीछे आकर अपनी किक से उसे डेल कर देता है, और इस पॉइ
8
00:03:30,000 --> 00:03:38,920
और तुम भी मेरे तरह कुमा के इस पंच का वेट कर रहे होगे, तो वीडियो का एक लाइक करके जाना, और चैनल पर अगर नए हो तो चैनल को सब्सक्राइब मारो, मिलता है तुमसे एक और चैप्टर के एक्स्प्रिणेशन में
English
1
00:00:00,000 --> 00:00:30,000
So finally Bido Kuma landed his punch on Satan, and in fact this punch was
literally very satisfying, if only one more punch had been landed, then all the
pain of Kuma would have gone out of me too, but we are happy even with this. ,
Because
2
00:00:30,000 --> 00:01:00,000
Like, self-destruction mechanism has not been put in Cooma, but despite the
mechanism which has been put in Cooma, how is Cooma alive, and how did it reach
the end, which we will know in this video, then the beginning of Sector 1104.
3
00:01:00,000 --> 00:01:30,000
Had given, Hey Vega Pang, what a scene is this, but even the thoughts of Vega
Pang, after seeing Kuma, they start getting restless, and just then Kuma rains
his single punch on Satan, due to which he becomes a rail, And this train hits
there
4
00:01:30,000 --> 00:02:00,000
are able to move, causing many to run to save the bunny, and Doctor Vega Pang
calmly bats, talking about the scenarios suggested by Satan, as to how Kouma is
alive and even asking, with his sour copy. lets think
5
00:02:00,000 --> 00:02:30,000
Was activated, due to which he becomes incapable of receiving any orders, but
still Doctor Vega Pangs skull is unable to find the reason for coming to Koomas
place, but suddenly he is taken to Koomas bread line.
6
00:02:30,000 --> 00:03:00,000
He easily regenerates with his power, which Satan had no idea about, and he
again uses the same attack that he used last time, but at this point of time.
coo
7
00:03:00,000 --> 00:03:30,000
Which goes through Satans body, but as he knows that Satan can also net in the
up region, then this attack is of no use, but there Jaru comes behind Frank and
defeats him with his kick. does, and this poi
8
00:03:30,000 --> 00:03:40,800
And like me, you too must be waiting for this punch of Kuma, then do like the
video, and if you are new to the channel then hit subscribe to the channel, see
you in the explanation of another chapter.
mp4
https://drive.google.com/file/d/1ZAtiiARUfaFTzyaTz_1rE76Qp9PfQLfP/view?usp=sharing
The Openai Whisper model has always had this problem, as it cannot accurately recognize languages with consistent pronunciation but different writing styles. For example, Chinese has two writing styles: Simplified and Traditional, but with consistent pronunciation. When recognizing Chinese, the original intention was to recognize Simplified Chinese, but always recognized as Traditional Chinese.
Oh Nice. Large v3 does work well. Now there is not much issue with Translation.
But if you observe the video the first Sentence till "Because" is stopping at 18 Seconds. And there is a Pause till 30 Seconds.
I also tried to use voice acceleration to fix that. But still it is giving that Pause. What should I do ma'am.
Try using the whole mode in conjunction with voice acceleration or video deceleration. "countdown_sec" in videotrans/set.ini, Set to 0 to skip pause. Restart software
Ma'am Software Runs without Internet too?
Ma'am Look I have selected "whole" And Enable "Voice Acceleration" Also Edited "Countdown sec=0"
And When I restart and run the project. Then it gave me output with the pause.
countdown_sec set 1
print screen
Ma'am I set Countdown Sec to 1
Still I am getting Pause
?record video
Ma'am can we please Connect on Telegram?
My User ID : t.me/Maximum51
As Like this it is taking too long.
I don’t see any problem. If the execution has been completed, of course it will be displayed like this. Then select again and continue clicking to continue execution. You record a video of the entire process starting from opening the software, selecting the video, and then executing it.
countdown_sec set 1
print screen
You about about something "Print screen" in this message. what should I do with that?
I just did Countdown_sec to 1
After changing to 1, it should be paused for 1 second before automatically continuing. I tested it and there was no problem. If there is a problem on your end, please change it to 1 and record the screen
Ok I have put on recording. Till then can you please check this video.
Take this video. It is in Original Hindi Language. And need to Target in English. Please give me this video without pause. Let's see how it works for you.
https://youtu.be/a5PyzSDIwLE?si=1XQEC42q3HWv7pWj
Original: Hindi
Target: English
Thank you
Here I recorded the whole process.
https://drive.google.com/file/d/1APx3HgofU76tahBbtEPc4dPYzKVyN3tp/view?usp=sharing
There will definitely be pauses because automatic recognition always has errors. Intentionally reserving pauses for manual adjustment. If you don't want the 60 second pause, changing the parameter "countdown_sec" to 1 will change it to pause for 1 second before continuing. If you want to completely avoid this prompt, it is not allowed
Speech recognition is done perfectly with Large V3 model. But output speech and the video is not aligned with audio.
Because when we convert Hindi to English. Then English Speech will take less time to complete.
So there is a Long pause after 14 seconds to 30 Seconds. But if i enable the voice acceleration it should make the speech faster or slower that it should Align with the video.
its normal
The pronunciation duration in different languages after translation is different. For example, a sentence in Chinese is 3s, which may be 5s when translated into English, leading to inconsistent duration and video.
2 solutions:
Forced dubbing to speed up and play, in order to shorten the dubbing duration and align with the video
Forced video to play slowly, in order to prolong the video duration and align with the dubbing.
Can you please provide the Installation Tutorial of the Software. Actually I Found One but. It's in Chinese Language. Can you please give that in English. Thanks.