He’s a real person with a pretty active anime career fwiw:
I’d guess there’s some sort of processing/post-processing contributing, but maybe he’s just impressively precise. I’ve also come across audiobooks where I wondered about that, but I assumed it was just me not being used that formal narration
Time spent on Aomi from 7月27日 to 8月02日 : 231min (3h51min)
Time spent on Audacity from 7月27日 to 8月02日 : 0min
I think I may be seeing some progress because I struggle less on clearing difficult Aomi cards, the next time I review them.
However, it’s a bit discouraging to notice that even after spending a long time repeating a card and listening to the recording multiple times, I’m still unable to produce a good pronunciation when I read the same text (even with pitch annotations) a day or two later—before hearing the target audio again.
That said, just hearing the audio once usually brings everything back on track. It goes to show how much easier it is to mimic natural speech than to generate accurate intonation based solely on rules and facts.
On a more physiological standpoint, I’ve been wondering how much chest resonance is acceptable for proper Japanese. If I really focus on projecting my speech into my mouth with minimal chest resonance, in some way, it kind of sounds more Japanese-like but this high-pitched voice doesn’t … feel like mine.
The thing is that I have to remind myself that I have to sound unnatural, at least at first, in order to inch closer to a Japanese accent. If I were to try to speak “naturally”, that would imply starting from the position most natural for French, given that I am naturally prone to adopt French vocal placement and habits. So if it feels weird and unusual, that might actually be a good sign because the opposite would mean falling back to old speaking habits unsuited for the totally different language that is Japanese.
I don’t know if this is also true in French, but at least in English I tend to sit in the lowest comfortable range of my voice when speaking non-emphatically. A realization I made the last time I did any pitch accent study was that I couldn’t actually do that in Japanese because if I started at my normal range I couldn’t make the pitch drops that words wanted without going uncomfortably low into my vocal range. I almost wonder if by necessity you have to have a higher baseline voice in Japanese because it’s important to make relative pitch drops as you speak. Anyway, that theory has helped me come to terms with how high my Japanese speaking voice is relative to my English one is.
That’s an interesting insight, I have never thought about it this way!
I have never formally checked but I feel like my baseline pitch gets lower when I speak in English compared to French. I certainly feel more chest vibrations when speaking in English.
I came up with a dummy sentence with a bunch of English loan words to get a feel for baseline pitch in the three languages:
In terms of pitch height, it feels like English < French < Japanese.
Feel free to criticize my English accent while we’re on the topic of accent reduction in foreign languages 
In the spirit of giving feedback because it’s really hard to get when you’re a good enough speaker that natives think it’s fine. 
I don’t have a lot to say. I can tell that you probably aren’t a native speaker, but I’m doubtful I could place your native language if I didn’t know it. You’re definitely in the, “oh your accent sounds interesting” region where I don’t think you’d need to do any accent reduction if you didn’t want to.
That said, for the most nitpicky critique I can come up with, that first vowel in “this” is not quite right. That’s probably the most noticable thing out of everything. It’s really really slight, but it is off.
Also, in my accent at least, I do not hit consonants that section of “took a taxi” nearly that clearly. It’s more like “tookadaxi”. Even if I pretend I’m going to clearly enunciate like you did in that sentence, I’m still going to say “tooka”, and give that final k a lot less weight. You might want to get some of the British English speakers to weigh in on that though, bc your accent sounds like it has maybe a bit of that flavor to it.
Your ee and a (particularly in supermarket) are too long & accented (you can even see they’re where the audio spikes in the clip) - which at least to my Northeast US/New York ear sounds weird. (To be clear, I don’t have a stereotypical NYC accent - most of us don’t - but am from the area).
Otherwise things feel a bit too clearly enunciated, and not slurred enough. In my accent: “this weeken’I took a taxit’th’ supermarket” is about as well as I could approximate it… In taxit’th’, the ’ = like the u in unnecessary (short and unaccentend).
Interesting, I didn’t notice that at all. I do agree that in general it has something of a British flavor to it - in which case all my comments might be irrelevant
It is funny you say that because British English represents a really tiny proportion of my English input, which is otherwise overwhelmingly American. As a matter of fact, I recall sometimes struggling to understand the actors’ speech when I was watching Black Mirror and The Peaky Blinders. Maybe the UK Received Pronunciation I was taught at school had a bigger influence than I imagined.
Time spent on Aomi from 8月03日 to 8月10日 : 172min (2h52min)
Time spent on Audacity from 8月03日 to 8月10日 : 0min
If it weren’t for this study log, this would have most likely been the week where I give up on shadowing and chorusing exercises. The problem is that I tend to study Japanese on the go (on the train, walking outside, …), while waiting (queues, 乗り換え, doctor’s waiting room, …), or late at night. None of those situations are suitable to repeating the same phrases over and over out loud like a lunatic.
Putting those concerns aside, this week, I’ve unearthed another unconscious bad habit: I raise the pitch of て in the middle of words (e.g. つ＼い／て＼たね). Even if I try to keep it level with the surrounding morae, the pitch graph still shows a peak around て.
Now, it could be that Aomori’s automatic pitch recognition engine simply gets it wrong because it is not perfect and is giving misleading signals but I don’t have a good way to tell. It is quite frustrating when you can’t trust your own senses but you can’t fully trust the measurement tools either. It is like fighting an invisible enemy.
A related difficulty that I’ve also come across is that there are times where I find it extremely hard to tell whether the audio difference between my recording and the native recording simply boils down to the fact we don’t have the same body and voice or that there are mistakes in my prosody or my phonetics. This problem gets way worse when the reference native speaker is a woman.
Tangentially, in an attempt, to better understand how a person speaking both French and Japanese fluently would contrast the phonemes in both languages, I’ve also briefly searched on YouTube for videos of native fr-ja bilinguals but all I could find were mostly videos of 3 year old 日仏ハーフ children babbling words in French. I’ve come across a few such bilinguals in real life (older than 3yo this time) but they live far away and we are pretty much strangers to each other, so I don’t even know how I could get in touch with them to ask for a few audio clips. Although, now that I am writing this down, I am starting to wonder whether I’ll truly be able to extract anything useful from such clips. Chances are that I’ll just be like “huh, yeah that certainly sounds French”, “huh, that’s definitively Japanese-like.”, “hmm, as expected, the /a/ sounds different” but how will that help me? 
This whole pronunciation endeavor feels so fuzzy and vague. At the end of the day, producing sounds should just be a mechanical process and yet it feels so magical and hard to grasp. Am I thinking too much? Maybe I should just repeat what I hear as faithfully as I can for an absurd amount of times without wasting time on thinking or analyzing and let the law of large numbers guide me towards the true value. But then what if I end up (unconsciously) reinforcing bad habits instead?
One of the more enlightening things I’ve gotten is actually from Japanese speakers when they speak English. It helps me dial into what the sounds they are using to substitute into English. It’s a lot easier for me to accurately hit those sounds by mispronouncing English words (and yet, when I have to speak in katakana words I still struggle 
).
I haven’t really thought about paying attention to pitch in Japanese foreign speakers or English, but maybe there’s something there too? And I’m sure it can’t be tooooo hard to find people speaking French to do similar analyses if it is helpful…
If you are having difficulty gauging the differences between your own speech and native speech on your own, there is always the possibility of enlisting a native speaker for help and feedback! 
That could be a good addition to your routine that would move things forward.
I booked a one-hour lesson on iTalki with the intent of spending it on corrected reading but we ended up chatting for the whole hour. It is not the first time, why does it keep happening 
? It would feel weird and abrupt to immediately hit the books but I can’t start small talk without getting carried away for a hour. Especially if it is my first time with that teacher or a long time has passed since our last lesson.
Time spent on Aomi from 8月11日 to 8月17日 : 36min
My excuse? I have no excuses. Honestly, I thought I would have been at 0 mins this week but I did squeeze in 36 mins so that’s… something…
Time spent on Aomi from 8月18日 to 8月24日 : 134min (3h14min)
| 月 | 火 | 水 | 木 | 金 | 土 | 日 | 合計 |
|---|---|---|---|---|---|---|---|
| 46 | 27 | 47 | 13 | 42 | 17 | 0 | 134 |
I recovered this week.
Honestly, I was lowkey hoping I’d be able to make some breakthroughs in my understanding of the mechanics of speech that I could share in this study log but in the end I don’t feel like I have much to share. Maybe it is just too early and I am being impatient. Maybe I’m not invested enough.
Sometimes I see some learners obsessively break down everything they hear to listen for pitch drops and it really makes me realize that we do not all play in the same league. Ironically, it is almost demotivating to see this level of passion in people. I am nowhere close to this level of dedication.
For me, hearing something repeatedly in different contexts and with time being able to reproduce it unconsciously is good enough. The only thing I need to know for that is, that differences in pitch exist, are important and I have to watch out for them. You already know that, so maybe it’s time to relax a bit.
Time spent on Aomi from 8月25日 to 8月31日 : 91min (1h31min)
| 月 | 火 | 水 | 木 | 金 | 土 | 日 | 合計 |
|---|---|---|---|---|---|---|---|
| 35 | 0 | 0 | 56 | 0 | 0 | 0 | 91 |
This week, I’ve got a surprisingly long streak of 90%+ score on the first take! I didn’t count but I hit at least 5 cards in a row with a good score on the first try.
Time spent on Aomi from 9月01日 to 9月07日 : 48min
| 月 | 火 | 水 | 木 | 金 | 土 | 日 | 合計 |
|---|---|---|---|---|---|---|---|
| 0 | 14 | 3 | 0 | 0 | 0 | 31 | 48 |
I think I really need to put a cap on the amount of retries I can do. Trying again and again for over 15 times in a row without a single recording hitting 85%+ accuracy is so demoralizing. I think it will be more productive to accept the loss and move on. I’ll give myself 6 tries before moving on.
In other news, during a Japanese conversation table (IRL), I got corrected by a native on my pitch for 後悔 （☓こうかい￣, ◯こ＼うかい）. It was totally unprompted and it took me a second to realize what he meant. I immediately repeated after him and continued on with what I wanted to say but after I was done talking, this seemingly small correction stayed in the back of my mind during the conversation: has this day really come? Has the ratio of properly pronounced words finally surpassed the ratio of improperly pronounced words so that picking out individual mistakes is no longer a futile endeavor?
This is the first time someone corrected my pitch accent without an explicit request to do so (as is the case in corrected reading sessions with iTalki teachers). I mean… isn’t this amazing? This was just a normal oral conversation so I wasn’t consciously paying particular attention to the way I pronounced the words, like I would in an 音読. There is a chance that I am reading too much into it and blowing it out of proportions but I’ll take it as a win.
Time spent on Aomi from 9月08日 to 9月14日 : 145min (2h 25min)
| 月 | 火 | 水 | 木 | 金 | 土 | 日 | 合計 |
|---|---|---|---|---|---|---|---|
| 0 | 29 | 30 | 12 | 0 | 54 | 20 | 145 |
Late report but I’m still going! After trying for a week, I think the limit of 6 tries before moving on really helped. I’ll keep this system in the future. Still a long way to go.
It would probably be helpful if I added intensive pitch listening sessions in my practice to train my ears to pick up on pitch drops in natural speech because I am totally not attuned to that. Although, before being able to constructively do that, I think I should do more progress in my pitch-focused Kaishi 1.5k Anki deck — so that I know what pattern to expect at least consciously. Some time ago, I started going back through all the 1.5k most common Japanese word, this time testing for pitch accent, but I have been neglecting my Anki these days.
Time spent on Aomi from 10月04日 to 10月11日 : 48min
| 土 | 日 | 月 | 火 | 水 | 木 | 金 | 合計 |
|---|---|---|---|---|---|---|---|
| 12 | 11 | 25 | 0 | 0 | 0 | 0 | 48 |
Oh wow, almost a month since last update 
If I keep up this pace, I won’t even make it to half the content by the end of my annual subscription (I am still at level 3 out of 10).
I’ve been attending a lot of in-person events with Japanese speakers these past few weeks, and have more events already lined up for the coming days, so it is not as if I’ve been totally neglecting pronunciation. It is just that it took the form of hard-to-track unscripted conversations instead of easy-to-track focused pronunciation drills.
