#and audio study methods in general work for me even when i'm doing other stuff at the same time. making it the easiest

[rigelmejo]

audio study methods

Still working on that 'lazy' study plan post, since I am just not satisfied with any chinese grammar guide summaries online enough to recommend them as a small grammar intro. If anyone knows of any good 'grammar overview summary' articles or sites for chinese grammar, please let me know. (I like AllSetLearning's Chinese Grammar Wiki but it is huge and in depth and not something I'd recommend a learner 'just read through' on month 4 of learning, and the grammar guide summary site I used as a beginner that was very easy to read through in a few hours... no longer exists)

So in the meantime. Not a grammar study tip, but a general 'lazy' option for language learners who (like me) can't focus on stuff like anki, or just don't want to. I go more in depth about using audio lessons and audio flashcards on other posts, and on the lazy study plan post i'm drafting, but the short of it is: you can listen and learn while doing your normal daily activities. That's what makes the study method so convenient. You don't have to squeeze in any extra time, or change your daily life schedule to make time for chinese, to use audio lessons and audio flashcards.

You simply find some times during the day when you'd either normally listen to audio in the background (like if you listen to music when commuting or shopping, or if you listen to podcasts when working, or if you listen to youtube while exercising or browsing social media). As usual, the more time the better as you'll make faster progress if you study 1-2 hours a day or more. But anything is better than nothing. So lets say you commute to work 30 minutes in morning and evening, there's your hour of studying audio. Or you go for a walk at lunch for 15 minutes, and browse tumblr for an hour scrolling (that's 1 hour and 15 minutes of study). It's very easy to fit 30 minutes of audio study into a day, and it's fairly easy to fit even 2-4 hours of audio study if you're so inclined. I usually do 30 minutes - 2 hours of audio study some days, since when I walk I decide if I feel like listening to a youtube essay or chinese or japanese stuff, when driving I decide which I feel like listening to, and I want to listen to something in english 2/3 of the time.

How do you use audio study material? Well, the easy way is you just press play on it, let it play in the background while you do other stuff, and that's it. If you tend to avoid studying new stuff (like me), then I recommend PRIORITIZING listening to NEW AUDIO every time, until you get into the habit of listening to NEW stuff to learn. Then you can re-listen to stuff sometimes, as review, especially when you're doing activities you have less attention on audio during. So for example: you'd listen to new audio on the commute or when walking (when you can mostly focus on what you're hearing), and then re-listen to audio as review while working or scrolling tumblr and reading english (activities where you pay more attention to other things besides audio).

What can you listen to?

There's audio lessons - which would be something like ChinesePod101 (Immersive Language Chinese in the Hoopla library app), Coffee Break Chinese, youtube videos where teachers talk in english and explain chinese as they teach it. These are good for study material, because you comprehend what you're learning due to the english explanations of every word and grammar point you hear. These are good for beginners, because you will understand everything you're listening to, and learn new words and grammar, thanks to the explanations. The drawback with audio lessons is they require the most focus.

There's learner podcasts like TeaTime Chinese and Slow Chinese, these are more often ENTIRELY in chinese. So these are better for practicing comprehension of stuff you've studied elsewhere, rather than for learning new things. You can learn new words and grammar from these, but if that is your goal then re-listen to learner podcasts a decent amount (5-20 times or more until you can't guess/figure out any more word meanings).

There's audio flashcards (which I love). These are sentence audio in english, then repeated in chinese. The order may vary, the chinese may be repeated more than once. These are good for beginners and upward, because you get a translation of every single thing you hear in chinese. You can pick up new words and grammar from audio flashcards. Audio flashcards require less focus than audio lessons, because you can learn from sentences while you pay attention and then if your attention drifts you can just focus again to the next sentence you hear and continue learning. The drawback is there are no explanations for which word specifically translates to what, some translations are not literal, and there's no explanation of why the grammar is the way it is. Audio flashcards require the listener to try and guess what means what by exposure to chinese sentences and their translations. So it's harder than audio lessons in terms of explanations, but easier than learner podcasts. Audio flashcards are the best substitute for traditional flashcards or SRS apps like anki, if you're trying to improve your vocabulary by hundreds of words ASAP. Audio flashcards are dense with new vocabulary (usually 1 new word or grammar point per sentence you can learn), so you'll learn more words than you would with an audio lesson that is paced slower with more english explanations or a learner podcast which would ideally be mostly words you know and only 20% or less new words.

There's Spoonfed Chinese Anki audio files (which I recommend since these start out very basic and increase in difficulty while also repeating words a lot so you can review, they're shared on reddit if you search, or ask me), if you search 'chinese english sentences' on youtube or bilibili (i've done this with chinese japanese sentences on bilibili) you'll find videos like this where you hear audio english then audio chinese. Old glossika cd files are basically this structure as well, which you can find the audio files of for free online or free in libraries (I'm using the new glossika app for japanese but I'm hesitant to recommend the modern app courses as there's significant errors in japanese so I'm not sure how good/bad the chinese one is). If you're a beginner, then the audio flashcard material you pick won't matter much as you need to learn a few thousand common words first which will be in most materials you find. But if you're an upper beginner, you may wish to prioritize finding audio flashcards with MORE unique words, more sentences, or may want to transition to using learner podcasts more for new vocabulary. If you aren't running into at least one new word for every 5 sentences you hear in audio flashcards (and ideally one new word for Every sentence), then that audio flashcard is way too easy for you and you know enough words to move onto new study material.

Audiobooks and audio dramas - use these like learner podcasts, listen to ones you can comprehend the main idea of, and then re-listen until you can't guess/figure out any more new words. If you're not very good at listening comprehension (like me lol), then you may want to listen to a given audiobook/audio drama file 3-5 times before deciding if you can comprehend the main idea (and use the material). When my listening skills are rusty, or just in general since my listening skills are bad, it can take me a few times of listening to recognize words I 'already know' and then a few more times of listening for my brain to put the words i recognize together into 'comprehending' what was communicated. So if you can read better than you can listen, you may want to listen 3-5+ times to a new audio file before deciding if you can follow the main idea or if it's too hard. And if you can READ the audio drama transcript, chapter text, but cannot understand the audio file? Then it probably IS at a good level for you to listen to, you just need a lot more practice hearing and recognizing the words you can read. So re-listen.

All of these listening study methods are good for:

Adding more study time into your day, since you can do them while doing other things.

Learning new words and grammar, when you don't have the time (or don't want) to spend time dedicated mainly to focusing on your study material.

Learning new words and grammar, if you don't use flashcards or SRS like anki but want the benefit of learning lots 'faster' than you would if you only picked up words during active study time (active study time being when you ONLY are focusing on study activities: like reading chinese, watching cdramas, chatting/texting people, and looking up words)

[mothsolar]

Long post below. Tl;dr ai art isnt new and discrediting the tool due to the misuse isnt giving it enough credit. Also image generation looks cool but everyone's a coward who wants to hide the computer fingerprints.

So I'm in the middle of a digital art program for school and we talk about the use of AI a lot. This isn't exclusive to when it started getting big due to image generators getting better, this is something that has been in active conversation since I first enrolled years ago. From chat bots to image generators to audio generators, AI is always something we are made aware of. And that's what has been making the conversation about it so interesting and disappointing to me.

Don't get me wrong, I despise the art theft as much as the next person. Every time I see someone use AI to recreate art (not just use it as an image source, just flat out recreate it) it makes me so genuinely mad. It gives a new and very efficient tool for people who can't make art to the level of quality they desire the ability to make it very quickly using nothing but some quick prompts and trial and error. Sure, you need to cycle over and over and over again while messing with text prompts to get your desired results, but it's faster than spending days worth of long painting sessions to make something that elaborate.

But the gimmick of making quick paintings is about the extent I think this can go. And it is a gimmick with how it's being misused cureently. These resources for AI generated content didn't just appear out of nowhere, it's been building for years as people keep experimenting with it to make new things. It is a tool, nothing more. The problem is that it is currently being misused and the internet is blaming it on the tool rather than the ones who use it in most conversations.

Computers process the world differently than people. They see things different. Hear things different. Think things different. When a computer sees a dog, it doesn't see a creature with fur and warmth and joy, it sees a collection of pixels that, after thousands of iterations of cross-referencing other images of dogs, is defined as a dog. There isn't any tactile or social assocciation, just raw data that is *somewhat* accurate. This is WILDLY different than what artists do when studying something to draw.

If you haven't heard of the shrimp method of learning to draw things, lemme explain real quick. When you wanna draw something new, you wanna study the real thing and draw a bunch of pictures of it and get a sense for how it exists in physic space. The texture of a surface, the way the lines make edges, how does light reflect off of it. You do this to get an idea for how to abstract this object as a 2d image rather than a 3d thing. That's how a human learns to draw stuff.

How a COMPUTER does it is based on this image recognition of mathematically figuring out which pixels approximate the learning data most effectively and if you've ever seen those websites from a couple years ago that would generate dogs, you can see how this doesn't work well as all! Everything's blurry and melty! There's too many teeth! Where does the hair end and the skin begin!? Horrors beyond my comprehension?! This is how a computer sees the world and i think that is so FASCINATING!

Even when a human attempts to steal art, they trace the lines they see and do similar kinds of studying. When a computer does it it appoximates based on image training and that's how you end up with these portraits of girls looking to the right with melty hands with 7 fingers that looks similar to the other portraits of girls looking to the right with melty hands and 7 fingers. The prompts and image data are all too similar and everyone's using the same program! Even tracers have a tracing style that's unique to them!

Computers create images in such a strange and interesting way that I adore picking apart and seeing all these images being generated to mimick human painting is just such a waste! You're trying to erase the one thing about AI that makes it so unique! This of course without drawing attention to the noncconsentual art theft required to even do such a thing. I can't claim to know when visual ai art began but I know it's not new because I was playing with tools like deep dream in 2019 for a digital art class. The reason this got big is because people finally learned how to hide the fingerprint of a computer well enough for it to win an art show and now it's seen as "good enough to replace the artists withholding their art from us."

I don't think AI can replace people. At least not commercially (which is where most of this stems from as art is seen as a comodity to the people making these claims rather than a form of expression and exploration but I'm not getting into that.) In the cimmercial scene, relying on a computer and random generation to create your assets is incredibly time consuming and inefficient. An artist provides updates and in-progress work that can be changed easily before being finalized. A computer processes the images until it's finished to give to you and if you don't like them, you need to start over again and wait. Doing this for a long term project would take ages and get expensive very quickly. It will not take off commercially.

But what about fine arts? AI has been in galleries since the first chatbot was developed and has only gone on from there. There are digital artists that have been ai generating music, images, conversations, movies, and so on since they got their hands on computers. This isn't new, it's just another tool.

Now I know this entire post has been talking about AI in a glamorous light, and that's because I still like it. I fucking love how computers process information amd seeing the debug tools and all the little calculations they do on the backend. This is an expression of that for image recognition. HOWEVER, like i have been saying throughout this entire post, THE ART THEFT CAUSE BY LEARNING ALGORITHMS AS WELL AS THE PEOPLE USING THIS TOOL FOR BLATANT THEFT IS BAD! I am so SO tired of seeing how horribly its being used and all the discussion being shaped by that. The fact that people put specific artists and art repositories in their prompts disgusts me to no end and i could go on about the backlash and other artist media and my mixed feelings on the ai ban, but my thumbs hurt and its 3. Just let it be known that in the right hands, ai art can be really cool.

[rigelmejo]

3/11/2025 Chinese Listening Experiment Update: 136 Hours

If I could go back and do something different when studying these years? I probably would've told past me, "please listen ONLY to stuff a few times." Every time I finished reading a graded reader, just tell myself "now listen to that without reading it a few times." More often when I read a webnovel chapter, I would've played audio as I read along, and then listened to the audiobook on it's own once or twice.

I made SO much progress a couple years back when I was experimenting with the Listening Reading Method, and I only did like 30 hours of it! But the method required me to listen way more than I was listening, and just listening that small amount was enough to make a difference compared to what I was doing prior. But if you've tried the Listening Reading Method, parts of it are very intensive, and it was draining to focus, so I gave up doing it. But the part where I read chinese while following along to the audio, and relistened to the audiobook chapters 5 times, those parts I could have kept doing...

I am going through TuTu DaWang as an audiobook right now lol. It was the first 'easy' novel for kids I read on Heavenly Path's Rec List, after struggling with way harder stuff and Xiao Wangzi as the easiest thing (but still harder than tutu dawang). I remember when I read it I looked up any words I was bad at recognizing, to fill in any gaps in basic words I knew how to read. I had been reading harder stuff, with more unique words used, and it had helped expand my vocabulary a lot, but I was also leaning on some less common vocabulary to guess some more common words, so truly 'few unique word' stuff where you NEED to know every word to understand what's going on was a bit harder for me... because I had no surrounding 'less common words' to use to figure out what's going on.

So I figured, now that I'm working on listening, I could use tutu dawang again the same way. To fill in gaps of any basic words I can't recognize in listening. A fun fact, tutu dawang is the first audiobook I've used so far with NO sound effects and no sound track, making it more important I actually know the words being said. I do think the narrator I found helps though, she yells when the characters yell, makes crying sounds when the character cries, does different voices for the characters. It reminds me of being small and the librarian reading a class books for an hour. It's just like that honestly lol, she even says 'hello everyone, last week we talked about, this week what will happen? jiejie will tell you the story."

On the same site I found tutu dawang I found dalin he xiaolin since that's another kid's book I'd read, and xiao wangzi. I figure I can listen to those next, since I theoretically know how to read the words in them (I managed to extensively read them a couple years ago). Then the site also has more kids audiobooks, and general ones, so I might try a couple more kids ones then browse their selection of other general novel audiobooks.

But yeah. All I can think is "wow I could've saved myself so many hours of struggle (at least in one great nonstop heap lol), if I had just listened to AUDIO only versions too of the things I read over the years.

I get that I am transferring SO many words I know from reading into the "recognize in listening too" part of my brain, and transferring so many words from "can recognize slowly and translate mentally" to the "can recognize quickly and don't translate" part of my brain. So of course it's going to take a WHILE. It's got to do this to thousands of words! My brain can't start focusing mainly on brand new words to pick up, until it's done sorting through all the words I partly know/partly remember.

I know it's going to take at LEAST a few hundred hours. If not 1000. Haha ;-; hahaaaaaa.

Two other notes:

I can definitely understand a decent bit without translating in my head. Because when I hear things like "their name is/this is X" I no longer can recognize if someone said jieshao/shi/jiao/demingzi to introduce them, I just recognized it and moved on, and didn't stop to pay attention to what word unless I'm purposely trying. This is an improvement, because I'm doing so much better recognizing names (after like 130 hours of listening lol), whereas I used to REALLY cling to the words right around an introduction to identify which words were names and which were objects. I had to pay special attention or I totally missed names, reading or listening. Now I'm noticing a lot of the time the names jump out to me because I can tell they are either a word I don't know in a position where I usually do know the word, or a word I know in a position I know that word isn't usually used. It usually only takes me a few paragraphs now to identify a name, if that much. Like in SCI the names pop out to me now pretty quick if they're in a sentence where names are usually DOING a verb, even if there's been no introduction sentences.

Complaining/venting about a CI approach below

I really think it sounds like bullshit all the automatic language growth method, and Dreaming Spanish method, advice that one will simply speak with good pronunciation and grammar if one waits to speak for 1000 hours. But what do I know, maybe I fucked myself up as an example because I spoke for a few hours early on.

I just... I honestly don't know if I could pronounce tones right, if I hadn't done SOME prep to study hearing the difference early on, since that allows me now to tell the difference between words even if I can't consciously tell you which tone pairs I'm hearing. I don't know if I could pronounce chi zhi shi as okay as I do (which hey may not be that good), if I hadn't heard and seen someone explain the mouth position in youtube videos. I predict my pronunciation is honestly going to be YIKES whenever I do try to speak, and at best I'll be able to hear a lot of what's wrong in my spontaneous speech, and then correct it with shadowing. I cannot imagine what a learner would go through if they tried to learn tones by hearing, without knowing in advance what the differences between what they're hearing are, or the way to make some of the sounds. I imagine at the very least, assuming they have good listening skills and can notice a lot, it's going to add a few hundred hours of confusion until they notice those differences.

But like... if my pronunciation is 'permanently damaged' by a couple of hours of speaking early on, to the degree I say tones of words wrong to an extensive degree (which would be significant enough to cause communication issues), well that's fucked. I get that ALG suggests you will have imperfect pronunciation if you speak early, but it also suggests you'll have pronunciation that's understandable and clearer than many other learners... especially if you avoid speaking until 1000 hours, even if you did do some 'early' damage of speaking a few times. Many Dreaming Spanish learners spoke tons in class before finding DS, or spoke 10 hours every 100 hours of listening because they live in a spanish speaking country and need to, and their pronunciation seems understandable... so I should 'theoretically' be able to get an understandable pronunciation if I just listen enough. But I don't know if that will ACTUALLY end up being true. I fear I'm going to need to shadow and correct mistakes...

And my grammar? When I try to make up a sentence, it's fine... if it's very simple. If it's short phrases or concise sentences. If the sentence is long, and I'm spontaneously trying to say something, I put the word order in english instead of chinese, which I know is not the right order after I've said or written it and can compare with a better example. I barely did any writing or speaking early on, maybe 2 hours total, compared to 1000 hours reading and now maybe 600+ hours of listening (with text sometimes). If 2 hours damaged my ability to speak with correct grammar, lmao, then most people doing Dreaming Spanish are fucked in terms of being able to speak with correct grammar... they often have years of spanish CLASS experience writing and speaking.

Maybe it's because spanish grammar is closer to english, and word order is closer to english, so if they make grammar mistakes it's less noticeable than with chinese? Maybe they ARE making grammar mistakes? I don't think so though... DS learners mention they do word things an 'english' way if they try to speak around 600 hours, but if they wait 1000-1200 hours they mostly speak grammar well. So it seems like prior experience is not 'damaging their results' SO much it becomes an issue...

Maybe it's because spanish is more similar to english, than thai or chinese is, so if Dreaming Spanish learners HAVE damage to their pronunciation or grammar use that's ingrained from prior experience (as in additional input doesn't correct it), it's not as noticeable?

I know there's thai learners that took ALG classes, and a lot of them seem to be fine at pronouncing words with the right tone! I need to dig and see if any had prior experience studying thai and speaking early, then waiting to speak again for 1000 hours, and seeing how their results were...