Glossika method

[etm001]

 

 

The main problem I see with their GSR files is that there isn't really any time to produce them before the next sentence comes along. The files are still useful for listening, but they probably ought to have a cue file or similar to add longer spaces between so that you can try to produce the sentence before getting the next one.

 

This was raised on the Glossika FaceBook page. Here's the response:

 

The issue specifically with GSR is this: the time required to listen to all the reps is around 20 minutes. Every second that we add to GSR would increase the whole package by 5 hours. If you wanted 4 seconds behind each recording, it would be 20 hours longer. The failure rate on listening to it daily would increase due to the length. We already have problems getting the large files to clients via download, and this would only exacerbate the problem. We can barely make a profit when people are attempting to download more than once. So we'd have to double the cost of the product. These are all things that people don't want to see happen and consider them inconveniences. Like Silat Scg Robinson noted above, our GMS files already have built-in pauses so if that's what you're looking for, use those instead. The GSR spaced repetition is supposed to build that little voice in your head so that you're able to recall from memory at any time and any place, so even if you're not speaking along with GSR it should still give you speaking ability when you're not listening to it.

 

 

 

[hedwards]

Sounds like they need to actually study the problem. If they're doubling the amount of the download, then they're doing it wrong. Silent portions of an MP3 take up practically no space. An hour long blank MP3 takes up probably less than half a megabyte. The cue sheets necessary to remove the extra pauses are likewise going to be probably less than a megabyte for all the levels combined.

 

This also reminds me of my childhood in post-segregation Seattle where black students were held to a lower standard from the beginning out of concern that they might only get 10 years of quality schooling rather than 12. I'm not suggesting that there's racism going on here, but focusing on getting people to finish by drastically lowering the standards is rarely a good idea. The GSR files seem to work well for listening but they aren't going to do anything positive for speaking as there's no opportunity for the student to produce and verify the production. Educational programs should be shortening the hear, reproduce, test cycle whenever possible. Cutting out the reproduction drastically reduces the efficacy of the program for speaking.

 

I happen to have the GMS files as well as I bought the entire package, but without a production component, you're just not going to learn to speak very efficiently. So people that only have the GSR files are going to be out of luck. Pimsleur, Michel Thomas and Language Audio Books all have audio programs for learning Mandarin and they all have space for people to attempt to produce the language before being given the answer because it's essential to the process. Also, the empty spaces in an MP3 take up basically no space. You can always remove those empty gaps using cue sheets.

 

And if bandwidth costs are such a large portion of their bills, they should consider doing what the Humble Store does and provide a Bittorrent option. I'm not sure how exactly they do it, but they use private trackers and I don't believe they have any issues with piracy. And as an additional benefit, if the files ever go bad, torrent does a good job of repairing the files without making the user redownload the entire file.

[OneEye]

I think you're missing the point of the program. It is not to be able respond to English prompts with the correct answer (which is backwards, IMO, and quite failure-prone), but to listen to and mimic correct, natural speech many, many times until you internalize the patterns and vocabulary. You have more than enough time to repeat each sentence in the GSR file. I do it slightly after the Chinese speaker. That means I'll overlap with the next English sentence a bit, but it takes very, very little brainpower for me to hear and process a sentence in English, even if I'm saying something else in Chinese. And with the amount of repetition in the GSR files (18 times over 5 days if you only listen to one file per day), you will very quickly become accustomed to the sentences and should have no trouble doing it.

 

It's important to remember that this is not Pimsleur, Michel Thomas, or any other course. The files are designed with a specific methodology in mind, and it is not the methodology that those programs use. You're trying to use them differently than they're intended to be used, which is fine, as I said earlier, but it's to be expected that your results will differ.

 

I've met Mike Campbell in person a few times. He speaks an ungodly number of languages to varying degrees of proficiency, and this is the method he uses. He used it to study several Taiwanese aboriginal languages from scratch simultaneously over a four month period a while back, then took the government proficiency exam in one of them and got a 94%. After four months of study, and while studying several others at the same time. Now, I think Mike has a particular talent for languages, and the fact that he knows so many certainly makes learning new ones easier than it would be for others, but I think this shows that the method works. I personally just allow it to work, without any alterations, and it works quite nicely. I experimented on and off with it using other sentence books before his product came out, but I did the same thing. Listen and repeat. Pay attention. Repeat, repeat, repeat. Listen. Mimic. It works.

[hedwards]

OneEye, this gets back to the website. The website and the program itself aren't very well designed in terms of communicating the intended use of the materials. If you've read the recommendations for the GSR materials, they don't make it clear that they're talking about shadowing. Nor do they really bother to make it clear what the objective or actual methodology involved there is. And I don't really see much that talks about how quickly one should be seeing results or how to know if it's working.

 

Which is fine, if you happen to know Mike and know what he's getting at. But as a self study program, that's clearly not acceptable. Considering that they spent 1 tiny paragraph on the GSR program, I think that it's not unreasonable to suggest expanding it to something like 2 or 3 to cover the intended use. As it is, it sounds very much like they're expecting people to pause the recording between sentences as there's not sufficient time to utter the sentence, even if you've already mastered it without getting into the next sentence. Especially not in the early stages.

 

The GMS method is quite a bit better in that regard. It has its own issues, but the instructions and recommendations appear to be much easier to understand and much clearer. I mostly wish that they would provide some sort of indicator about where the various sentences are found in the MP3s, although I do realize that this isn't very easy.

 

As far as brain capacity goes, I don't agree at all about that. You're probably somebody that was gifted with a larger than usual amount of working memory, but I can't listen to something and repeat something else in my native language. how on earth do you expect me to believe that I'll be able to do that in a foreign language if I can't do that in my native tongue without making a load of mistakes?

 

That's just my 2 cents on the matter.

[imron]

Every skill can be trained, maybe not to the highest level so that you're the best in the world, but definitely to an acceptable level. If you can't do something yet, you can get that way through correct regular practise.

Mimicking, listening and repeating are all trainable skills.

[navaburo]

I also initially found the GSR files way too fast. And I do get the impression that they would be painful and inefficient to use on their own without access to the transcripts or GMS.

 

However, I have found that by using the GMS files on the first day, I can use the GSR thereafter without an issue. I have been first going through the GMS "A" files with the ebook, writing the sentences down, and checking in the transcript if I miss something. Then on subsequent days, I'm using the GSR files.

 

With GSR, I try to speak the Chinese at the same time as the Chinese speaker. If I don't recall the sentence, then I just listen. It feels really good when I can finally produce fast enough to match the rhythm of the native speaker. When I finally nailed the sentence: "zhe luguan bingbu zenme hao, weishenme zheme gui?" it was a bit of a rush.

 

It's interesting to hear the different experiences with the program from users here. I think with constructive criticism we can help refine this method for a much wider audience. (Though you still need a combination of obsession and dedication with language learning to succeed at it IMO, regardless of the method used.)

[hedwards]

I think part of it has to do with how ones brain is wired and how one is used to functioning. I may wind up having to take back what I said, I'm doing about 4 tracks every day, and I'm starting to see some suggestion that things are sinking in. The listening is definitely improving, which I don't think was ever in doubt, the speaking seems to be catching on a little bit as well.

 

I'm not, however, listening to what I'm producing while I'm listening to the track. But, I'm not really sure how important it is to be consciously listening.

[sangajtam]

I bought it but why so big image type pdf?

I would like normal text version and i would like to have it on android phone with pleco.

[etm001]

 

 

However, I have found that by using the GMS files on the first day, I can use the GSR thereafter without an issue. I have been first going through the GMS "A" files with the ebook, writing the sentences down, and checking in the transcript if I miss something. Then on subsequent days, I'm using the GSR files.

 

This is similar to my approach as well. I listen to the GMS file just once to see if there are any words with which I'm unfamiliar (almost none), then record myself speaking each sentence. After that, I listen to the GSR file and will speak the Chinese in time with the native speaker. It's a bit of a verbal workout. It works for me because the vocabulary is familiar. If I was dealing with a lot of unfamiliar vocabulary, it would probably focus on the GMS file longer before attempting to mimic the GSR audio.

 

 

 

With GSR, I try to speak the Chinese at the same time as the Chinese speaker. If I don't recall the sentence, then I just listen. It feels really good when I can finally produce fast enough to match the rhythm of the native speaker. When I finally nailed the sentence: "zhe luguan bingbu zenme hao, weishenme zheme gui?" it was a bit of a rush.

 

Exactly. Around the ~5:00 minute mark is when new sentences are introduced within the GSR file. If I find myself tripping up on a new sentence, I'll just listen to it the next 2-3 times before speaking it in time with the native speaker.  

 

And as for: 這旅館並不怎麼好，為什麼這麼貴? Ha! I know what you mean. The native speaker speaks the sentence at a fast clip (as she does in all the GSR files), so it takes a few tries to say it smoothly.

 

加油！

[Crush]

Does anyone have any more thoughts on/experiences with the program now that you've had it a few more weeks? Has anyone regularly done the lessons? I saw that Mike recently posted some videos about using the program to learn Thai and it seems like a rough ride for someone who is a beginner and doesn't know IPA very well and has no one to really explain things and correct their pronunciation errors. I'm not so concerned with Chinese as i think i've got a solid base and i think my pronunciation is not too bad. One thing i noticed though is that they don't seem to use the course as the instructions posted earlier in the thread say. In fact it seems kinda haphazard, one day we do these files, the next day we do these, and it doesn't seem like there's any real order to it. They don't write anything down, in fact it seems like they don't even pay much attention to the written portion of the course. There's no recording and analyzing your pronunciation, etc. It's just listening, shadowing, and following along in the book.

 

And to be honest i am really bummed that rather than have the nice 2MB text files that you can copy and paste from they've switched over to massive non-OCR'd files which you can't copy and paste from. I guess it's an anti-piracy move or something (which i'm not too crazy about, either), it's an extra bummer that it comes at the cost of convenience to the learner.

 

Despite the flaws, i decided to take the plunge for a couple reasons, namely to activate the passive knowledge i've got of the Chinese words i've studied, to build a stock of useful phrases to fall back on (i forget who, maybe etm001? But their example about "我来取餐" describes a lot of my interactions in Chinese pretty well), and, lastly, just to support Mike. I've never followed their stuff very closely but they do seem pretty active in the language-learning "community". Now just have to wait for them to send me the download link.

[imron]

And to be honest i am really bummed that rather than have the nice 2MB text files that you can copy and paste from they've switched over to massive non-OCR'd files which you can't copy and paste from

I really don't understand the logic in this.  It's not going to deter a pirate (who would probably just upload the whole thing as a torrent anyway), and is a serious inconvenience to the learner, even those who aren't copying and pasting, because it also means you cannot search within the document for anything.  Considering the PDFs don't have any table of contents, it makes it even more cumbersome to navigate to a specific point in the file.

[Crush]

I hadn't even thought about that, that seems like a much more useful thing to have. I figure i'll just add my own bookmarks as i go through the book but it would be nice to be able to search it and have a nice table of contents.

[wibr]

I am using the Chinese GSR files, starting with the first book. I try to do one file every day, since at the moment I have more important things to take care of. With more free time I might increase the number of files to two or three. My pronunciation is reasonably good and I already know many characters but not so many words. So I can only speak very little Chinese.

 

I found that the glossika files fit my needs quite well. What I really like about this approach is that I can integrate it quite well into my daily life, similar to flashcards. I try to find methods with a low "activation threshold", meaning that I don't have to do much to get started learning Chinese. Just start the file and go, I immediately get repetitions and new content.

 

However, I have the following issues:

• I am also VERY unhappy about the "text as curves" rubbish in the pdf, it has so many drawbacks and so little benefits (for them, absolutely no benefits for us)
• It's difficult to look up new words and almost impossible to look up words which were introduced in a previous file but you don't know which one... and personally I feel that something essential is missing if I have no idea about the characters in a word. I want to make sure I get the pronunciation right, sometimes it's quite fast so instead of dai4 I hear da4, so I need to look up many sentences to make sure about those subtleties
• Sometimes the repetitions in the GSR files can get annoying, e.g. very simple sentences like "No, I'm not" repeated so often that I almost stopped the whole file...

Due to these issues I am experimenting with OCRing all the files. I try to do as many automated checks and corrections as possible, but in the end I guess I would have to check the result manually. It's quite frustrating to work around those arbitrary limitations...

 

It has been mentioned in this thread several times that basically all those problems could be fixed with an app. Even a very simple one would be much better than the current solution. They wouldn't have to distribute huge audio packs, you could see the text everytime the audio plays, right on your mobile, they would get some copyright protection and could add more functionality over time.

 

Right now they seem to be focused on releasing more content, they linked to this article on their facebook page saying that they want to expand into endangered languages. While I admire the mission to support those languages, from a business perspective I can see more pressing issues...

 

Overall, I have mixed feelings about this but in the end the positive aspects still prevail, at least for me

[Touchstone57]

For someone who knows what they want I think this product works really work. As an intermediate user who has has used the mass sentence method before from my own materials I found this fairly easy to jump into. But I do think some beginners may be put off from using this as comprehension may be quite low in the beginner and you will hardly know what you are saying. 

 

As I product I think it is a great idea, as you are saving yourself a lot of work by paying someone to produce a thousand useful sentences built in repetitions with audio files and native pronunciation, it really does build up conversational fluency over a period of time. 

 

However, I am still a put put off by the quality of the website. It is a bit all over the place. Logging in seems hard enough, but there doesn't seem to be any organisation to it, as each product page contains a list of every other product they have in a blank link which isn't always relevant. It makes sense to have everything clearly accessible from a products section, perhaps divided by language. There is a lot of example sentences and images thrown around in different places that aren't always relevant to the product they are selling.

 

I hope they clear this up in the future as I think the product works they just need to package things a bit better. 

[Snuggles]

I also wish they had standard mainland Mandarin.

[Crush]

I second the wish for a "standard" mainland Mandarin pack, i really wouldn't care where they took that standard from (north or south).

 

I also have concerns about using the course from a beginner level, i'm sure it works alright if you stick it through, but having no idea what people are saying and being left completely on your own to figure out grammar and pronunciation (albeit with help from IPA transcriptions which even Mike mentions don't always match the audio perfectly) seems a bit overwhelming and frustrating for most learners. I'm past that stage (probably somewhere around HSK4) and i imagine the first couple fluency levels will largely be a review, but i'm hoping they can build some spontaneity in my speech that wasn't there before. I've never done massive sentence methods but i've been using Anki and Mnemosyne off and on for the past 6 or so years, but this isn't your average SRS system and seems a bit stale in comparison (eg. practicing "No, I'm not" alongside things like "Who's the man in the photo?").

 

But anyway, i got the download link to the fluency packs today so i'm looking forward to testing it out. I scrolled through the first ebook and it doesn't seem like there's really anything new for me, but it might be a nice review before i head back to China in a couple months...

 

EDIT: I will add a couple things: the links are now clickable in the e-mail so you don't have to copy/paste them. Also, while being massive files (one of them is over 2 gigabytes), being hosted at Amazon they do download pretty quickly. If you have a slow internet connection, however, that doesn't really help any. The mp3 files are also saved at a hefty 191kbps, i'm not sure why as i doubt you'd really notice if they dropped it down to 32 or 48kbps (or at least offered the option of downloading at a lower bitrate for folks with slower connections). There are also no tags whatsoever on any of the mp3s. Luckily the names are meaningful so batch processing them won't be a huge issue. It's just small things like this that make you scratch your head sometimes. The branding at the beginning of every file is pretty annoying, too, in particular the "brought to you by Glossika" clip. I'll probably just cut all that stuff out.

[realmayo]

the "brought to you by Glossika" clip

 

Fun the first time. Increasingly infuriating thereafter.

 

being hosted at Amazon they do download pretty quickly

 

I was throttled, presumably my internet service provider? Meant 5 hours to download the biggest file, which I couldn't do in the 24 hour window. They resent the link when I asked, but it came midweek and again had no time! Will have to ask again.

[wibr]

@realmayo Yeah those introductions are annoying, it's the same story with every audio product I tried (chinesepod, CLO) and I have no idea why they do that. There should be absolutely nothing which is not specific to the current lesson.

[Crush]

I agree, at least here it's only a few seconds (maybe 12-13 seconds). With ChinesePod i remember the introduction being over a minute.

 

I got about 1MB/s, so i can't really complain. It took under an hour to download, though without a fast internet connection it would be really annoying, especially that silly 24-hour deal. A bit-torrent option with a private tracker that someone else mentioned here would be a nice alternative. It'd be nice if they at least provided an md5 hash so you know if the file you're only allowed to download once downloaded successfully or got corrupted. The whole process is a huge pain.

 

I just finished reading the introduction and noticed a couple silly errors like "words get strung together in sentences which is has been well analyzed" on page 8 and "If you ever learn new in isolation, you have to start using it immediately..." on page 20.

 

Also, the app idea is great. This seems like it would fit perfectly in that format and wouldn't be too difficult to do.

[hedwards]

@Realmayo, it shouldn't be on Amazon's end, I've got a 5mbps connection and I think the files were taking less than 4 hours each, even when I was doing more than one at a time. I wonder if you just wanted to download at the wrong time. Either that or you have a cable modem and other people are doing a lot of downloading.