@zoran Hi Zoran, I am getting the error message below when trying to import an audio files for transcription, it claims I have exceeded a monthly limit which I was not even aware of. Can you please look into this? Never encountered it before. Thanks
Iām getting the same message also. I wasnāt aware of a limit. I would appreciate some feedback so that I can manage my imports. Many thanks
Please note that each user is limited to a maximum of 600 audio minutes of transcription per month. If you see that error message, it means that you have reached the limit. Also, we are making changes here so starting from next month, you will be able to track exactly how many minutes exactly did you spent and how much you have left.
1 Like
Zoran, thanks for the update
1 Like
@zoran Thanks for the reply
1 Like
600 minutes means only 10 hours, for per month, it is really not enough. Why change this rule? And please let me know what the limit for Youtube transcription?
Zoran, will you provide us, the users, with some way to bypass the 10-hour limit on importing audio materials? I would consider the introduction of this limit a reasonable step if you offered the ability to create lessons based on a pair of an audio file and its corresponding subtitle file. Do you have plans to add such a feature?
@vadimtsushko1 If you are importing a lesson manually, adding both text and audio, this limit wonāt affect you.
On the import lesson page, you can either copy-paste text and upload audio, or you can upload both audio and text file simultaneously.
Importing a lesson this way, wonāt affect your transcription limit.
1 Like
10 hours of audio is actually quite a lot of learning content. We introduced this limit because our costs in providing this service are quite high. Our options are to restrict its use to a reasonable amount or to increase the price. In future we may provide an optional add-on for additional transcription minutes if there is demand. We do recognize that it is convenient to have this service integrated into LingQ.
1 Like
I completely understand your decision. I rely heavily on this to learn from podcasts. If there were a premium membership option at a higher price, I would be willing to pay for it. Thank you for your response.
This audio transcription brought me back to using Lingq actively, thank you for it!
10 hours per month is a bit too little for my way of studying at an intermediate level now that I have found a good study workflow for using this and know better what kind of material works for me best (not tv show transcripts, in my case, a lot of wasted minutes taken from the monthly quota).
I would like to request for one feature. Audio replay and the study/sentence mode donāt seem to be synchronized both ways. In the audio/play mode if click a sentence, on the rolling text, I can click the text page icon and study the sentence. But if in the text mode I click the play button at the bottom left corner, the audio does not recognize the text page as the starting point. So I have to try find the right timing manually.
1 Like
Thank you, zoran. I have just tried, and this scenario works for me.
1 Like
It must be a cruel joke to introduce a limit of just 10 hours ā only 10 hours! If youāre studying more than one language, itās even less.
If only everything worked flawlessly, and you could at least be sure that those 10 hours would be transcribed accurately ā but you canāt be! Not to mention the constant bugs popping up, some of which remain unresolved for months.
And of course, youāve decided to introduce the limit without any communication or announcement, as usual.
On the bright side ā Iām glad to have this āupdateā a month before my subscription renews. At least now, Iāll have time to reconsider my plans to stay with LingQ.
2 Likes
Thanks, weāll see what we can do.
Yes, for more than one language this limits one quite a bit. Especially at the intermediate/advanced level.
1 Like
10 hours per month is unfortunately too little, im doing around 3 hours daily of podcasts, I wish there was some way to extend that. thanks for the response though
1 Like
No, 10 hours per month is very limiting. Thatās about 20 min per day. Iām studying Mandarin using both traditional and simplified character sets, so I import the same audio into both ālanguages,ā effectively limiting me to 10 min per day.
Steve has mentioned his use of Lingq for Persian podcasts several times, including in a recent video ā but no mention of a monthly audio transcription limit. Iām pretty certain his podcasts are not limited to 20 min per day. I recently received an emailed newsletter that included a notice about a new icon design ā but no mention of a monthly audio transcription limit. Not cool at all.
What costs so much? What backend is used? Whisper is freely available (GitHub - ggml-org/whisper.cpp: Port of OpenAI's Whisper model in C/C++), although it does seem to take some computational resources. What exactly are we paying for? (I say āweā because Lingq is certainly paying for the audio transcription using our subscription costs.) I suppose Iāll have to go back to transcribing using Whisperā¦ At least Iāll have some more control over the processā¦
1 Like
We are using Whisper through Open AIās api. It is not inexpensive. We are looking at ways to optimize the cost. If we can do so, we will look at increasing the limit. We previously hosted Whisper ourselves but that has ongoing costs as well in terms of manpower and server costs and brought performance issues along with it.
There are converters to convert Simplified to Traditional and vice versa. That might be the easiest and fastest thing for you to do in your situation to at least get both scripts.
1 Like
This discussion led me to look into how hard it would be to transcribe my audio using Whisper running locally on my own machine. So far, it has not been that hard to set up and to run. I transcribed and loaded my first lesson to LingQ using the import audio+text method Zoran outlined a few posts above in this thread.
So far, the results have been pretty good (āgoodā == meets my personal needs). I will need to do more tests on my end to verify I can trust the output Iām generating, but based on initial tests, I think itās going to work.
I will probably transcribe all future audio on my own machine from here on out. I realize this might not be an option for everybody, but I donāt mind the extra steps. Hopefully it will lighten the transcription load and will lower the operational costs for the LingQ crew.
I just searched for this transcript limit, because it was mentioned in another thread (the one about importing failed, because the file is empty).
While I understand that services get more expensive each and every day, I have 2 problems with this:
A) itās not communicated anywhere, and the error message does not disclose that there is a limit, and that that is the problem. So a heads up would be nice!
B) I am learning several languages actively, and so the 10 hours is really not much, if you share it for several languages. I believe I am not the only one, who is not just pursuing one language. While I know this is a new feature, which makes learning more convenient (and I do appreciate that!) I would still like to be able to use it, since LingQ is all about importing meaningful input, I mean thatās what it was made for.
In the meantime:
Is there any other way to transcribe the youtube videos, which are my main learning resource for some of the languages?
Thanks