Category Archives: Non classé

Transient restoration (dynamic range expansion) on heavily compressed audio, my how-to

Recently on Twitter, I posted a very short but messy how-to about transient restoration for loud or compressed recordings. I used a 1994 eurodance single as an example in my tweets.

I’ve been doing that since 2008 which matches the time where I started to feel concerned about the loudness war. At the beginning, I was using only transient shapers. But over the time, I learned to perfect my method by isolating transients better and minimize any artefact created by my process. I think it’s time for me to write a more detailed how-to about my method of restoring transients or squished portions of audio on any kind of audio material. With very clean, transparent and artefact-free results. Yes, sir!

Before beginning, here are the prerequisites in order to complete our task:

– a sound editor with VST support
– digitalfishphones’ Dominion VST plugin (http://www.digitalfishphones.com/main.php?item=2&subItem=4)
– Michael Gruhn’s Transient Shaper VST plugin (http://dioxaz.free.fr/download/midi_vst/VSTBundleWin32_FINAL_2008-01-27.zip)
– a noise-gate VST such as GGate (http://www.gvst.co.uk/ggate.htm)
– a brickwall limiter such as TLs Pocket limiter (https://www.kvraudio.com/product/tls_pocket_limiter_by_tbt), might not be necessary
– both plugins extracted to your desired and usual VST plugins folder
– your sound editor configured to use those VSTs and/or its VST list properly refreshed
– it is highly recommanded to use a sampled-based scale rather than time-based in order to have sample-based precision for our selections, otherwise you’ll have all sorts of unwanted frequency shifts because of delays induced by VST plugins

Update: I’ve made another attempt with a slightly more striking example, using Keane’s Somewhere Only We Know. Also, if you don’t want to read the following long text and don’t care about pictures, here’s a link to a shorter and more summed up how-to: https://fr.reddit.com/r/LoudnessWar/comments/5i67ux/transient_restoration_dynamic_range_expansion_on/db73rqb/

If you happen to own the album or have that song lying around, you can even try to do it yourself, as it may give you very pleasing results (you can even attempt your own improvements).

For this example I’ll be using a Touhou Project arrangement this time. It’s Beyond That Door by circle Diao ye zong. It’s not a particularly significant or bad example but still yeilds the usual bad habits of current day loudness-based mastering with your usual pushing to the “zero” (0dBFS), its limited peaks and flattened out longer portions.

So here’s what it looks like when opened with (insert your usual audio editor). Here, it’s Adobe Audition 1.5:

dyz_btd_original
As I said not really that bad but we are going to bring that to almost late 80s-style mastering.

The first thing you’re going to do is turning it down to your desired or what you think is the ideal volume. It can be anything depending on your tastes and your habits. Don’t hesitate to go as low as 10dB of volume reduction if you feel it’s needed or if your song is definitely too loud. For this one, I’ll target 6dB of volume reduction:

dyz_btd_minus6db

Now, leave that instance alone. Create a copy of it and paste it in another instance. Your’ll need it for your actual job. You might need to paste it as many instances you need depending on how much frequency ranges you plan to work on (lows, lower mids, higher mids, highs, etc.).

Let’s begin by taking care of the bass part of the song. To do that, we need to apply a band-pass filter to cut out anything that is not bass. If using Adobe Audition, then the best effect for this task is FFT filter (the one which allows you to draw a shape and set blocked frequencies at 0% and passed ones at 100%). We need to determine the sections which are actually affected by compression and limiting. So don’t hesitate to trial-and-error and see which frequency range is the best suited to isolate the visually squished parts, by comparing the waveform of your filtered result to the original one. The narrower the range the better it is (but for bass, you can easily go up to 200Hz). On this example, I chose a range of 0-180Hz for my band-pass:

dyz_lows

So we have the bass isolated. The goal is to have transients appear, which is the case here. It means there are some potentially restorable transients in that frequency range.

Before starting, copy the entire audio of this band-passed low range into clipboard (Ctrl+C or Edit->Copy). It will be useful for later.
Now, let’s grab one of our transient shapers. For the bass, I think Dominion is much better suited, even if it’s creating some bass shift afterwards. Apply settings as indicated in the following screenshot. Set input volume to -1.5dB, tick “x2” at the bottom attack sliders, set attack level to 50 and attack length to 5ms. There’s no need to touch at other sliders:

bta_dominion_settings_lows

After applying the VST, you’ll normally see your transients going up in volume while maintaining quiet parts roughly the same. That’s exactly what we want! But it’s not over. Dominion creates a delay of 101 samples (for 44100Hz audio) when operating. So we need to delete 101 samples at the beginning of the audio. Now we have to make the difference between the original band-passed audio and this transient shaped attempt. To do that, place your cursor at the very beginning (or sample 0) and paste-mix back what you copied earlier but with inverted phase. If using Adobe Audition, use Ctrl+Shift+V and this dialog appears. You just have to tick “Invert” for both channels:

pastemix_invert

We do that in order to create some sort of expansion matrix. If we stopped at this stage and pasted our band-passed shaped transients as is back to the original audio, without subtracting it, there would be a huge EQ shift in the bass!!
Then, to isolate the transients further, we need noise-gating all the quiet parts. In this example, I’ll noise-gate anything below -21dBFS in order to minimize the bass-shifting created by Dominion. You can either use a brickwall limiter set to 21dB, then reduce the audio to -21dB, then remove the sample delay produced by your VST (for TLS Pocket Limiter, you need to delete 1 sample at the beginning) and mix back an inverted copy of the limited audio to the non-limited one, using Cancel history (Ctrl+Z or Edit->Cancel) or creating more instances. Or, more simpler, use GGate, set the threshold to -21dB, attack and fade to your likings and there you go:

dyz_lows_expanded_gated
Yes, my friends, isolated bass restored transients! Now it’s time to mix them back to the original song.

But before doing so, in order to get rid of any possible artefact created by our process, let’s band-pass a second time to the same frequency range as before. Your filter settings must not get lost at this point. They should be normally still stored in memory as your last used settings if you did not mess up or if your program didn’t crash in the meantime. In case, you can write your frequency range down or store a preset in your audio editor.

Now the waveform changed a very little bit. It’s time now to mix it back (paste-mixing) to the original song. Copy that bass restored transient audio instance into clipboard (Select all then copy). Go back the original audio instance (which has been turned down to your desired volume, here it was a 6dB reduction). Put the cursor at the beginning at the audio (sample 0). Now, paste-mix what you’ve just copied at full volume with original phase (not inverted, don’t forget to untick “Invert” for both channels if using Adobe Audition) and enjoy the result!! All bass transients came back to life again! Listen to the song and see it satisfies you. If not, try different volume mixing settings (you might try to tame things down by setting mixing volume to something like 70% in order to prevent transient shaping gone out of control and possible clipping, or increase to something like 150 or 200% it it’s not enough for you). For this song I let it mix at 100% volume.

So we took care of bass. All bass transients seemed to go up without touching at the quietest parts.

dyz_with_lows_expanded

This is starting to look really good! But we notice some parts still look flattened out, notably louder vocal or instrument parts like violins and pianos. These are definitely not transients. However, we’re going to use a similar method. Let’s go back to the bass instance and cancel everything we did or work on another instance you created or copied from your untouched original (with volume reduction applied) previously.

We have to determine on which frequency range the squishing is occurring. Depending on the song, you may or not need to take care of the low or high midrange. Let’s do some trial and error and begin by the low midrange. If you get something close to the squished parts you observed or the original waveform, then you’re on the right track. On the contrary, if your band-pass don’t reveal any transient or louder parts, then try another one.

So for this step, we are going to band-pass anything between 180Hz and 650Hz approximately. You can see how the loud parts actually match some of the remaining squished parts (I concede it’s not obvious on this particular example), such as loud vocal parts:

dyz_lowmids

Before attacking, don’t forget to copy this band-passed low midrange audio to the clipboard, just like we did for the previous frequency range.
Now repeat the same process we did above by using Dominion, this time setting input volume to -2dB and attack level to something between 25 and 40 (I used 30 in this example). Again, trial and error. Choose a setting which allow transients or loud parts not to go out of control (or clip). Don’t forget to delete 101 samples at the beginning after Dominion is applied. Put the cursor at sample 0 and paste-mix what you copied earlier (band-passed low midrange before transient shaping) with inverted phase. We can see now that the quietest parts are now below -27 or -30dBFS. Noise-gate to that, band-pass a second time to the same frequency range and you’re almost done:

dyz_lowmids_expanded_gated

If there are still some quiet parts where nothing should be altered (parts matching to quieter parts where no restoration is required), just silence them out after selecting them, one after the other (Effects->Silence in Adobe Audition) to keep only your loudest parts restored. Now you’re ready to copy the result and paste-mix it back to the original audio at desired mixing volume.

Note: this frequency range generally contains some snare components in most songs. Not in this one. However, if your audio does contain audible snare transients (even kick transients in most mid-90s Dance songs) in this frequency range and the shaping happens to go out of control, making transients really irregular in volume, then you can attempt limiting them, by using TLs Pocket limiter. Basically, you limit them at the volume threshold you desire (if you don’t want your transients to go over -6dBFS, choose a +6dB limiting), then reduce volume by the same amount, then delete 1 sample at the beginning of the audio (for 44100Hz audio) in order to compensate its VST delay.

Okay, it’s now looking like something. Just watching the waveform, it’s now looking straight like something that could have been released in the late 80s. And some of you may stop now and be happy. But sometimes, you also need to do the upper midrange and high-range too as squished areas are still present. In this example, I did another transient shaping on the upper midrange, between 660Hz and 1800Hz approximatively, with the same methodology as above, still choosing Dominion as the transient shaper and using -1.5dB as input level and 28 as attack level. My isolated, shaped and gated transients and loud parts for the upper midrange look like this:

dyz_himids_expanded_gated

If you intend to perform transient and loud parts restoration on high-range, you have to be very careful as it is the most sensitive frequency range when it comes to artefacts. Actually, it’s in that frequency range that artefacts are the most audible. At this stage, there are 2 solutions. If the highs contain only short transients, then it’s best to use Michael Gruhn’s Transient Shaper instead with attack level to max (200 if I remember). However, if it still contain longer louder parts, then use Dominion, but very carefully. You can start by reducing the attack level to something like 20 or 25 and try out some longer attack lengths. You can experiment longer releases when noise-gating too. But don’t forget that if your shaping goes out of control for highs then you’ll be more likely to kill the transparency of your song.

I didn’t perform any transient shaping for the high-range as I felt it wasn’t necessary due to the nature of the song. Here’s what the final waveform looks like after paste-mixing all shaping attempts (both low-range, lower midrange and upper midrange) back on the original audio reduced by 6dB:

dyz_expanded_final

So far, this waveform now looks amazing. And the song also sounds more open, more dynamic. As I unfortunately don’t own high-end or good monitor speakers yet, I can’t hear any obvious artefact using this method. I’m pretty sure it’s actually producing very minimal artefacts if none. However, if you happen to have high-end equipment and are able to pick some artefact of any sort, please let me know.

I posted 30s samples on Hydrogenaudio forums if you want to compare (please forgive my relative newbness there, I never liked forums — I know I should have avoided that, but it seems I’m too lazy to read rules thoroughly):
https://hydrogenaud.io/index.php/topic,113240.msg932367.html#msg932367

Respectively, you have a 30s sample (with selected parts cross-faded) of the song used in this example, after my transient and loud part restoration attempt, as done in this how-to. Then, the same song untouched with 6dB reduction. The 3rd example is an earlier attempt I did, a long time prior to writing this, using similar settings and method with slight differences. I personally prefer this sample, as I was a bit more tasteful in restoring bass transients.

Feel free to comment if you have questions and remarks about my method, or if some parts look obscure or hard to understand to you, or if there are any mistakes or oversight. I hope I have been useful in helping you fighting the loudness war.

Memories of Phantasm 7, Hisoutensoku Episode – my review

After that “silly symphony” attempt we got with previous episode with Shiki, I really wondered what they could put out next. How does this series still cope after all those episodes? Is this one really worth? Is it better than 6? What is right and wrong with this series? That’s what I’ll be trying to answer throughout this review.


Let’s forget about “silly symphony” and attempt something else

I really wondered what could come after that Shiki episode. To my surprise, Manpuku Jinja decided to go a complete different route by embracing Cirno’s story from fighting game Hisoutensoku. This is the first time the circle uses a fighting game as a basis where most viewers were expecting another take on a danmaku one. The artistic approach behind this will seem obvious to a part of you reading this. Also, this might also be the opportunity to get rid of the grimace feast we saw on episode 6.

Yes, it’s obvious they chose an easy route again by emphasising on a popular and cute character. It’s not a bad idea per se. It’s even pretty appreciable as it may help building something around the ice fairy and give her some more developed personality. And, you may disagree but I think they succeeded. Sure we still have the stupid 9-ball and “you’re and idiot” gags but when you know Manpuku Jinja’s animated series well, you know what to expect. I said many times they try to cater the widest possible audience. So this shouldn’t surprise you after 7 episodes, and what shouldn’t surprise you either after all that time is the lack of something so ubiquitous in commercial animation that some of you take a bit too much for granted (I’ll come to that at the end). And seriously, it is not that bad. At a time, we have to stop the bad faith. Just because you see something you don’t like, that doesn’t make the episode a no brainer. “But look, there’s no fighting scene, Dioxaz, no sakuga moment in that episode, it’s the worst!” some may be tempted to reply. Alright, let’s come to that.


Characterization, writing, and storytelling

Let’s talk about aspects 4chaners and all other internet posters never talk about because they’re too busy in bashing with all their jadedness.

After 7 episodes, I’m relieved to notice again how endearing their characters still remain. And to be honest, I was pleasantly surprised with what they did with Cirno and Daiyousei (litt. “great fairy”). I guess it has to do with their portrayals, their attitude and the lines they deliver (something you can only gauge with a translation when you don’t speak the language). Those are the strongest points in Manpuku Jinja’s handling of the series and this doesn’t change here. Also, I don’t have to mention the quality of the music, sound design and backgrounds which are top notch as always. Also, the mood that emerges from some scenes is really nice and the episode still flows nicely without never making you asleep.

However, that episode reveals a weakness that’s been there since the beginning but only started to get obvious from here for me. The storytelling is almost non existent. I believe this is partly why this OVA gets an undeserved bad rep among some viewers. Sincerely, there have been many MMD drama attempts that were much more touching than this whole series. Staying true to the games stories is one thing but what this series lack is building something new and ground-breaking around them. But I guess they don’t have the time and have to be concise. Instead, the circle prefers choosing having a hundred lines of dialogue instead of taking more time in order to create something deeper. You all noticed how insanely fast those episodes are produced. And indeed, this leads us with some nice whoopsies, just like I will show here:

g13_4894 g13_4858 g13_9449
Observe Alice’s eyes and Utsuho’s mouth. What the hell did the character designer in those shots smoke when he made them? This is not something you would expect from an OVA of that class. I hope the circle is not reproducing the same operating model of big commercial animation studios here, as it’s us, viewers, who will be in trouble this time, not just that poor ice fairy


Not a battle scene, rather a confrontation

What made a welcome return in that episode is a battle scene… erm, no, not really as it doesn’t last very long. It’s more accurate to name what happens between Cirno and Utsuho “a confrontation”.
If you read my reaction on Twitter, I admit I’ve been a bit too harsh. Even during that simple confrontation, you can sense how our ice fairy is in trouble. This shows a personality aspect of Cirno that is not commonly seen in fanworks (especially animated ones), her determination and courage. So, does it mean it sucks? No, your beloved sakuga is here and well present, just not as obvious and epic as in episode 4 for instance (remember that clever frame recycling from one of the openings in order to build a part of the battle scene with Sakuya). That fight scene might have looked rushed to you but it’s not as bad as what we saw in episode 1 of Maikaze’s OVA for instance. But I frankly admit I really wish that confrontation was more epic and a bit longer.

g13_10186 g13_10644 g13_10728 g13_10825
No, this isn’t a battle as it is too short but yes, the sakuga is still there as Cirno tries to dodge and block those threatening fireballs. Another amusing detail is the wallpaper-like PCB pattern that wraps what I think is the Underground Geyser Center


Controversies of Phantasm

Don’t get me wrong, producing a series like this one is already a ground-breaking effort on its own, I acknowledge it. But between producing short openings or PVs and producing long episodes, there’s an immense gap that the circle seems to have a bit underestimated. Some viewers quite rightly noted that the openings always end up unanimous while the episodes themselves always leave mixed impressions.

There’s a simple reason behind that. Making a short OP doesn’t have the same set of constraints that a full episode have. In order to make an episode, there’s a whole new set of things to consider like giving a tone, writing a story, decide what you will tell (what you will include or prioritize and what you’ll have to exclude or rule out), making stronger and decisive artistic decisions. With an episode, you have more things to show. They could have made sure the episodes stayed in tone to the openings but that would have required stronger storytelling elements, one of their weakest points, as I wrote earlier. They instead thought it was better to rely on high quality art and animation to compensate for that caveat. And definitely, the choices they will make in order to compensate for that caveat just can’t be unanimous. You just can’t produce what is literally an animated postcard and hope everyone will enjoy it. Hence the choice of including fanservice, memes, filler scenes and making sure our characters look like they always have something interesting to say in order to justify that “let’s watch it for the art” orientation.

Also, I see some viewers complaining how it doesn’t even follow the original stories accurately, throwing in the circle’s own interpretation and choosing what they want to show instead of what should be essential. This is unfortunately a point I can’t verify as I’m a sucker with the games. However, this is an area which must be explored with good stepping back and no jadedness. If you have some interesting thoughts, you can drop some into the comments.


But… but!! Where are the voices?! I can’t hear anything! Don’t tell me you didn’t notice, it’s the most striking thing in the series! (also, a word on fandubs)

Reading complaints about being no voices in Youtube comments (both from Western and Japanese viewers) always bring me a smile, especially after 7 episodes.

It’s not the most striking if you’re used to watching fanworks, especially Touhou ones*. It’s not when you consider it’s a fanwork, too. And it’s not when you start thinking and considering what having good voice acting actually involves.

*There’s Nico Nico Douga for that, and also Touhousubs’ Youtube channel if you want a careful selection of them, with English subtitles, all reprinted with actual permission from authors

The constraints I listed above remind a bit of the ones you encounter when you want to go full voice for your own work, like recruiting your cast, testing it, giving acting directions, waiting for the lines to be delivered, double-checking the lines are properly delivered, eventually checking if some lines have to be redone, adjusting the tone of your animated feature to the strength and weaknesses of your VAs, adapting your story so the low number of VAs you have at your disposal isn’t detrimental, eventually adjusting the audio quality of your voice recordings (eq-ing, dynamic compression, volume, panning, effects, etc.) and make sure there’s homogeneity between all of them, etc. Ironically, that’s a whole set of things they completely decided to rule out, either as taking the easy way out, or as an assumed artistic choice, as it gives way more liberty this way (no need to alter your story because of the voices and you’re free to choose the tone you want). I speculate they were too afraid of screwing up their episodes so they ended up with written dialogue only. And this will be the final part of this review before concluding.

Again, if you’re familiar with Touhou fanworks, you probably noticed how easy it is to screw everything up with subpar voice acting. Just look at the Cookie☆ voice dramas. Just look at anything that features voice*. A hint of cheese will always be there no matter how good the voice acting is. I guess this has to do with how Touhou Project stimulates our imagination like no other franchise has done it before, thanks to how ZUN built this wonderful universe, set of characters and background behind them. So, in order to accommodate that, one of the smarter choices is to rule out voice, as it is the hardest resource to get right and as it immediately kills that imagination stimulating factor I mentioned above for this specific franchise (a bit unrelated of an example but sometimes there’s nothing wrong with preferring a well-known TV series to be read as a manga instead of being watched, as even low-grade and trivial full animation, rushed audio work, generic soundtrack and stereotyped voice acting can add their own share of cheese on their own… but it’s a whole other debate).

*I’m also including those so-called Yukkuri voice and Voiceroids here, as an improper and abusive use of them can lead to very cringe-worthy and questionable results, especially on non-parodic content (I’m not talking about people who make a cutting-edge and innovative use of them, rather the habit of using them as cheap trendy and attractive tricks because some people mistakenly think it’s nicer to have voice than nothing)

So, in the end, it’s not a bad thing at all that this OVA series has written dialogue only and no voice. I sincerely hope it will remain that way. Sure it kind of looks dodgy and clumsy. Sure there’s a little too much dialogue to make us forget that clumsiness and sure watching the endless lipflaps moving in the air without producing sound will feel unnatural to most of you, because of the strong influence commercial animation had on us for more than 90 years now. But keep in mind they’re just drawings. Think of it as an animated book and strongly convince yourself this is a fictional and imaginary world, while keeping in mind you’re not watching something meant to be aired on regular TV. If you succeed, you might find the characters cuter that way and like it better. You might even think it has its own charm and how it can be set apart. Also remember how people did when talking movies weren’t still a thing.

Now, if you just can’t because you can’t and because you think I’m a moron to be able to waste my time in making all this up, there are fandubs. A vast majority of you will only watch it like this anyway. Currently, the best fandubs are the ones made and directed by Chinese amateur VA shourei小N. They have variety (many VAs are featured, including NENE and vivian198808), excellent audio quality and the quality of the performance is rather good. vivian198808 also make her own ones alone, but I think her voice is too soft to be versatile enough (note: she also makes full voice versions of some well known MMD dramas). Other fandubs to note are the ones made by ヘイ眠気覚まし. But again lack of versatility and audio quality in earlier episodes are subpar. Another notable mention is 鶏子だよぉ (http://trkdy.com/). Because of how her Marisa sounds, she can almost be designated as the true successor to Pika Aizawa… except that can’t be because she’s obviously not as versatile (her voice is medium to high pitched and not soft enough, so deep and soft voices aren’t her strengths).

g13_4701
Call me weird but I love these awkward moments when watching the original episode with no voice. They just feel special because they look so awkward (just like this particular one).


The pikadub! the pikadub, dumbass!

And this leads us to another point you think I didn’t notice or I ignored on purpose. No, I’m coming to it. During the 3 first episodes of Memories of Phantasm, there was a series of fandubs that were unanimous with almost everybody. They were made by solely one person, amateur VA Pika Aizawa.

What was so great about them is how they sounded very close to TV animation standards, making the series look and sound like a potential flagship TV animated series. She was energetic, full of self-confidence with an unmatched versatility, allowing her to perform all kinds of characters. She by herself single-handedly eliminated the need to see elsewhere and even come back to the original unvoiced episodes for most. Except, she wasn’t so perfect. She was acclaimed because her acting matched what almost every ordinary anime enthusiast from the West expects when watching anime. That’s why those fandub versions had so much success. As a side note, it also brought some undesired and more clueless viewers to the party.

Now, coming back to the actual performance itself, it did have some weaknesses indeed. As versatile she was, sometimes her timbre sounded a bit too similar from a character to another. She was also prone to free-wheeling and exaggerating a lot, as it is very in fashion in the anime industry. Western viewers typically love this but that doesn’t make her acting the de facto example. The free-wheeling can get irritating after a while and you wish she was more subtle and tempered. The best example was her rendition of Marisa in the Bluray version of episode 1 of this series. Because, yes, she did eventually get called by the circle to perform on one of their release (as a bonus audio track however), unlike what some 4channers like to claim on /jp. It was just too much to me, to a point it even felt a bit jarring. Her first version (the one which made her famous on Nico Nico Douga and Youtube) just sounded right because it was more tempered.

And finally, the worst happened. After dubbing 3 episodes of Memories of Phantasm, dubbing episode 1 of amateur MMD series Angels of Paradise (as Marisa) and releasing a Voice drama CD, she definitely quit the scene and wiped everything she did off Nico Nico Douga, probably as a result of that excessive and unwanted fame. People were here just for her fandubs of Memories of Phantasm and nothing else. Everybody only cared about the “pikadub” without even thinking about it. It was the ubiquity for many viewers. Now she left, this whole set of viewers are completely lost now. I hope this will help them in being a little more curious and open-minded, and realize how having someone like Pika Aizawa was something exceptional. Now the future of Memories of Phantasm is without her, that’s how it is (and I repeat, there’s nothing wrong in watching it unvoiced).


What to think now?

That section about voices and fandubs was probably too long. I’m sorry about that. You may notice how I finally more reviewed the series as a whole rather than episode 7 itself.

So, how good is the episode in the end? To be honest, like I said, we have to stop the bad faith. As much as the grimace feast we had previously was kind of disturbing, the return to something we’re more used to was more than welcome. Well-developed characters, nice moods, nice dialogue and even slice of life scenes which don’t act as soporific, as they would with Maikaze’s series. We also had the welcome return of some sort of battle (even if it short). We for once have the feeling that something is happening again (episodes 5 and 6 were vague in that regard). So, for a Memories of Phantasm episode, it is a good one. Sure it can’t be as exciting as previous ones because the discovery and novelty effects are long gone and far behind us. But there’s something else that makes the episode lack excitement, after 7 of these.

Seriously, I really wish the circle work better into the storytelling department as it is once again the true Achilles’ heel of this series. Look what Kyoto Fantasy Troupe did with their episode. They should take this as an example or at least a reference. I don’t want them to abruptly change their plans and being more story-oriented all of a sudden, as that would be even more detrimental if they screw up. No, what I’d really like is that they consider one of those elements, in order to add a slightly deeper tone and minimize memes and fanservice progressively (The Sealed Esoteric History had none of these).

Also, the other thing that bothers me and other viewers is how they give more and more the image of not knowing exactly where they go. Just look at the teaser at the end. The usual post-teaser segment (concluded by a nice “Next phantasm…”) always seen in earlier episodes is not even present here. Not very inspiring.

g13_16719 g13_16779 g13_16828
Respectively “The preview for Episode 8”, “Everybody is waiting patiently for it” and “That’s it!”*
Umm… yeah, that’s not exactly what I’d expect from a teaser…

*credits to Kuilfrayt for the translation. The English subbed version of the original unvoiced episode can be watched here (an alternate English subbed version by Melancholic Celeste can be watched here)

To give an image of professionalism, you have to know where to go. Showing indecisiveness is definitely not a good thing and will upset your viewers. So, to people at Manpuku Jinja, next time try at least to pick-up a scenario or game that will allow a slightly deeper tone and make you work on your storytelling a bit more. I don’t care if it must take more time than usual. Episode after episode, your audience widens which means bigger expectations and you have to take that in consideration, too.

So, umm… after being drown in my own thoughts, what if you’re not familiar with the series? The answer is simple. You can try them at their Youtube or Nico Nico Douga channel (or some of the fansubbed and fandubbed reprints). For me, both first fourth episodes are worth as they all have their share of interesting segments (but also some signs of laziness that may disturb you). Episodes 5 and 6 are special in their own kind as they gave the impression of trying to find their feet and never finding them. They’re more like experiments. Not the ones I’d recommend to purchase. And there’s this one, which tried something and quite succeeded. If you like Cirno, take it. If you’re too jaded and annoyed by their story shortcuts and choices, I can’t do anything for you.

And for those familiar with the series like me, this one is definitely recommended because of the good time you’ll have by watching it, even with a confrontation rather than an epic battle. If episodes 5 and 6 disappointed you, this one may do the job.

I sincerely apologize for my lack of conciseness. But I think writing thoughts as they come can be interesting too, especially from the perspective of someone with non-existent journalistic background. If this set the record straight about all crap that can often be read on forums and message boards and give you some comfort about what to think about this series, well that’s good for me. Because that’s what I wanted.

Whatever the route they pick up (hoping they never screw up), I wish the circle good luck for the future, as making decisive choices and decisions is definitely something not easy, as there are some other type of constraints that we, viewers, often forget or ignore the existence because not involved in the process or not really aware what really happens during the making of an ambitious animated feature.

g13_5364
Someone on the /jp section of 4chan cleverly noticed how this particular segment shared similarities with the original Cookie☆ voice drama* with the same exact place, the same exact BGM used as a base and how Reimu drinks her cup of tea (she’s even shown eating a cookie!). Again, an intentional reference made by the circle. In the end, having those references is definitely not a bad thing when subtly introduced, like here.

*can be watched with English subtitles here: part1part2part3

I watched The Sealed Esoteric History in full voice! Is it worth?

Some of you may remember how skeptical I was about this full voice version after watching that teaser posted by Kyoto Fantasy Troupe (the circle behind that OVA) on… April 1st. Well, I have some good news for you. Watching this in full voice was a really pleasing experience, to my surprise. Since I already reviewed the episode entirely in its unvoiced form, I will not be as long this time, if this is of any comfort to you. So, here are my impressions after watching the awaited Sealed Esoteric History in full voice.

Reminder of purchasing options still available as of December 24th 2016:

You will indeed have to use a proxy service if you want to order from these stores as they don’t offer overseas shipping.

First impressions


The announcement made by the circle on April 1st listed only professional and high-grade voice actors (“seiyuu” if you prefer) probably as a result of a poll conducted some months earlier. Here is the complete cast, sourced from the circle’s blog, and translated with help of some other various sources:

  • Maribel Hearn (CV: Kana Hanazawa)
  • Renko Usami (CV: Eri Kitamura)
  • Kaguya Houraisan (CV: Mamiko Noto)
  • Eirin Yagokoro (CV: Sumi Shimamoto)
  • Watatsuki no Yorihime (CV: Ayako Kawasumi)
  • Watatsuki no Toyohime (CV: Tomatsu Haruka)
  • Tsukuyomi (CV: Shibata Hidekatsu)
  • Prince Tsuku (CV: Inoue Kazuhiko)
  • Prince Sasarae (CV: Koyama Rikiya)
  • Yakumo Yukari (CV: Inoue Kikuko)
  • Ibuki Suika (CV: Shimizu Ai)
  • Saigyouji Yuyuko (CV: Kayano Ai)
  • Hakurei Reimu (CV: Imai Asami)
  • Kirisame Marisa (CV: Asumi Kana)

Some names might inspire you something or bring some memories. This indeed gave high expectations, except that the little teaser they posted on Youtube and Nico Nico Douga didn’t sound very inspiring. All voices sounded pretty generic and there was very little room for improvisation. This gave me mixed impressions at first. Then, some time after the product was released, I stumbled across an unauthorized reprint of the episode in full voice on Youtube (it has been deleted since, as a result of the reprinter closing his Youtube account). This is how I watched it.

Update! The circle finally uploaded a low-resolution of the episode on Youtube and Nico Nico Douga on August 12th 2016. It looks like they finally learnt their lesson and are doing the same thing as Manpuku Jinja now (remember how they were against that idea at the time they released their unvoiced version).

Quality of the performance and acting


Indeed, what I suspected was confirmed while I was watching it. However, it turned out better than I expected. When put in context, with moving pictures and in more various situations than isolated voice clips, everything sounded more in tone. The performances weren’t exceptional but were enough on their own. Since it’s a professional cast, no misstep is to note, everything is pretty much spot on. Sure there were very little to no room for improvisation which led the cast to deliver their lines rather plainly. There’s no key or really striking moment that would have allowed the voice acting to truly shine unfortunately… except perhaps during that battle between Suika and Prince Sasarae (one of my favorite moments, even in the unvoiced version). You might also add the segments with Renko and Merry, which sound rather delicious when voiced (but both segments already had their own charm even with no voices anyway). Another amusing but predictable feature is the addition of background voices in the mix, mainly during the street walking scene and… you guessed it, during the battle involving Suika and Prince Sasarae (for background characters and fighters). This was to be expected indeed, but it’s still a nice and welcome addition.

I guess this has to be with the way the episode was written, and it didn’t really give possibilities for the voice acting to shine, even with help of high-grade and professional voice actors. I liked almost all the voices, with a crush for Suika. Merry and Renko weren’t bad at all. I also liked the performance of the men cast. The only voices I didn’t like were Eirin, Marisa and Reimu. Eirin’s voice felt out of place for me, especially compared to what we heard in the full voice version of Orbital Rabbit* (that I, ironically, prefer unvoiced*), the iconic MMD drama from Harumao. And for Reimu and Marisa, they also sound completely out of place and way too high pitched to my tastes. Fortunately, they only appear a few minutes at the end. However, if they were to appear more often on possible later episodes, I will effectively start to cringe, especially if the circle decides to go full voice only.

*both can be watched with English subtitles here and here respectively (two parts for each, only first parts are shown here), thanks to Touhousubs

What does the voice really add to the episode?


You noticed how I enjoy Touhou animated fanworks without voice much better. This gives that unique fanmade cachet, leaves room for your imagination and gives a really special mood when done properly. Touhou is one of the rare franchises that allows this. This clashes with what we see in amateur animated works in the west, as they are almost always fully voiced no matter what’s inside and not always with good voice acting to begin with. I’ll not develop on how insisting on always having full voice actually decreases your creative power and gives you many constraints (see my rants about Yomi’s MMD series, Angels of Paradise, on Twitter if you’re curious). I’ll save that for later.

So, when the opportunity is given to watch a Touhou fanwork in full voice, there’s some excitement because it is always welcome (a part of the audience do want full voice). There’s also some concerns as it is quickly easy to mess up, especially in Touhou fanworks.

In this particular case, it simply made the episode flow much better and watch more naturally. It gave the characters more presence (some will say more “life”, which I don’t really like because I’m convinced there are other ways to bring life to your characters), which is to be expected in such cases, as always. And because the unvoiced one felt a bit dodgy at times, the addition of voice was more than welcome. Clearly, it felt more like watching a real TV OVA, which is exactly what some people are for, and exactly what I’m not for. But I’m always open for other possibilities. Also, full voice gives an unbeatable advantage when showing it to friends if you have them as guests at home (full voice is more indicated in this case… but you might want to try without if you want to act anti-conformist) or to an inexperienced audience which might not react positively when seeing endless lipflaps without voice, and on which you want to introduce to Touhou. In the end, as much as I prefer my “2hus” mute, full voice is also more than welcome, as it can bring some fresh air and make those “2hus” more real. In clear, voice only added very little in my opinion, but that “very little” can actually make a huge difference for most of you, especially those not familiar or having trouble with fanworks featuring subs or speech bubbles only but no voice.

What version to pick up in the end? Is that full voice version really worth?


I said earlier that this version adds rather plain voice acting which actually turns out pretty good when watching it in real time, on the characters. If you read what I wrote before, I think you pretty much drew your own conclusions.

If you’re like me and already got the unvoiced version and liked it (and possibly preferred it), you won’t miss much by not buying this one. You can try watching it if you come across a reprint, just out of curiosity and decide, afterwards, if it is worth giving your money again (giving JPY3,500 twice does hurt hard indeed).

Now, if you didn’t buy the unvoiced version and already watched it reprinted and felt it was too dodgy, then this voiced version is more than indicated. I think you won’t regret your purchase, as voice will bring exactly what you expect.

Then, if you’re pretty new to Touhou and its animated fanworks and didn’t decide yet what side you’re on, then it’s a bit tricky. I’d rather recommend to purchase this voiced version, as it is harder to find in illegally reprinted form (I mean uploaded on Youtube or tor***ts or whatever other sharing means). You could then try out the unvoiced one later, as it is easier to find, and decide for yourself if you think the lack of voice does indeed bring some undiscovered character or if it simply feels too dodgy, awkward and unnatural for you.

To finish with the last case, if you hate or loathe voice acting in Touhou, the issue doesn’t even arise. There are also chances that you don’t even like Touhou animation (well, in the form of long episodes) to begin with anyway.

So, this ends here my review of The Sealed Esoteric History in full voice. I hope it helped you visualize its quality much better and comforted you in your future purchase or not.

My old music page is online again

As I said on my twitter, you probably never saw that page anywhere, because it was very poorly advertised. Back in the day not only did I do wallpapers but I also used to compose MOD music from time to time, using Fast Tracker II. Not the best tunes ever but still nice anyway. Before I reupload some of them to my Youtube channel and make a proper presentation or review of said selected songs, here’s my old music page, in plain HTML, that I started in 2002-2003:

http://dioxaz.free.fr/dioxazmusic_old/

Beware! It is really plain and dull and is in French! Because I never bothered setting a proper page to reach an international audience (I though it wasn’t worth at the time). I originally began making MOD music in 2000 (I just graduated high school and was entering university at that time*) with rather rudimentary Fast Tracker 1.00, then quickly migrated to Fast Tracker II. Making music in both Protracker 4-channel and XM 32-channel (max) formats was really fun and I enjoyed it, my only gripe was my lack of patience and the trouble I had to timing my notes properly. I’m still proud of some of the tunes I did (SILENEM5, BSTHEM9A and DXDREAM2 come to mind), even today.

Note: I updated the page with later revisions of the song I originally posted (better quality samples or better used polyphony in most cases). I hope you’ll enjoy… or smile while listening to them. 😉

And, something I forgot. VLC plays those tunes natively, in case you wonder.

*university didn’t last long and wasn’t for me, I preferred preparing a IT 2-year degree instead

Magic Non Stop – Dioxaz’s wallpapers

What the hell is this will some of you ask? I’ll answer this is what happens when you mix 80s, Kraftwerk and Touhou together.

Obviously, it takes its inspiration from the music video for one of Kraftwerk’s hit singles, Musique Non Stop, released in 1986. The video was directed by Rebecca Allen (work originally started in 1983 but due to how slowly work on the album was progressing at the time, the video was only ready for 1986!). However, this is not a mere copycat or a screenshot of one of the video frames. What I did is recreating some of elements that struck me the most when watching the video, by memory. And I came up with this…

To my knowledge, no normally constituted MMDer would use wireframes in his works… yet I’m doing it, to great effect I think. Types I used are NimbusSanP, NimbusSanT (both URW’s take on Neue Haas Grotesk, respectively “poster”/”display” and “text” variants) and a Japanese font called NKD02 14-segment Hematitic. My only gripe is that it looks pretty busy, possibly making your icons a bit drown out if you plan to actually use it as a wallpaper. Also, I did the composition with widescreen-only in mind, so I had to improvise a bit for older squared aspect ratios.

Anyway, here’s the beast (click on thumbnails for original size):

– 2K format (2560×1600)
wallpaper_magicnonstop_2560

– FHD format (1920×1200)
wallpaper_magicnonstop_1920

– SXGA and WXGA+ formats (1440×1024)
wallpaper_magicnonstop_1440

– HD format (1366×800)
wallpaper_magicnonstop_1366

– SVGA format (1024×600)
wallpaper_magicnonstop_sd

– VGA format (720×512)
wallpaper_magicnonstop_sd

Note: there’s an odd one out in this picture, will you be able to find it?

Alice Manufacture – Dioxaz’s wallpapers

I told you I’d use an argyle pattern for this one (and self-made, in totally vector form). Again, an idea I had for a little while. Now here it is materialized.

The particularity of this wallpaper is how it gets away only very basic elements. All models (by nya) were rendered using MMD stock settings with no effects whatsoever. The processing you see is made with Photoshop. Some inexperienced viewer might think I used an MME effect where in fact it’s not the case. I also reintroduced my usual play with typography (Univers Condensed and Ultra Condensed are used here). I think it’s a tiny little more inspired than the previous one with Reimu. Anyway, you’ll be the only judge. Even if people end up not interested in them, my wallpapers remain a hobby, and I like making them. So I’ll keep on anyway.

Now here’s the beast (click on thumbnails for original size):

– 2K format (2560×1600)
wallpaper_alice_manufacture_2560

The 2K one is actually 2650×1600 in size instead of 2560×1600! I only typed a wrong number when beginning the project. It’s my mistake. Anyway, it shouldn’t make any difference when cropping.

– FHD format (1920×1200)
wallpaper_alice_manufacture_1920

– SXGA and WXGA+ formats (1440×1024)
wallpaper_alice_manufacture_1440

– HD format (1366×800)
wallpaper_alice_manufacture_1366

– SVGA and VGA format (1024×640)
wallpaper_alice_manufacture_sd

Not offering a VGA variant this time because the contents of the SVGA one fit the entire VGA resolution when cropped, without suffering too much… in case any retrogamer and old 14′ CRT PC monitor user might stumble across this page. 😛

Someone on Reddit suggested the idea of a mobile resolution, too. Definitely something to think about for later attempts (which would require producing in portrait, too).

Sad Reimu – Dioxaz’s wallpapers, the return!

Ten years after… ten years my friend! This should probably show that I’m older than you think. This is basically what happens when you download MMD, fiddle with a couple of things inside it and let inspiration flow again. I was a little greedy for this one. I had that idea of a sad Reimu for so long. In fact, I barely know how to manipulate bones in MMD at this point, so don’t panic if the pose doesn’t look very natural. I also tried to minimize the damage by carefully choosing an appropriate viewing angle. For my next attempts I’ll not be as greedy.

The typeface I used is Plantin (from a *ahem* pirated Linotype typo CD) and the final result is not so bad… I mean for a come back. I know I saw more inspired days but I’ll do my best to bring them back. All in all, that trip within the working of MMD was pretty fun and it made me able to produce something again… after 10 years of what one could call “hibernation”. Now here’s the beast (click on thumbnails for original size):

– 2K format (2560×1600)
wallpaper_sad_reimu_2560

– FHD format (1920×1200)
wallpaper_sad_reimu_1920

– For older small widescreen and standard monitors (1440×1024)
wallpaper_sad_reimu_1440

– HD format (1366×854)
wallpaper_sad_reimu_1366

– SVGA format (1024×640)
wallpaper_sad_reimu_sd

– VGA format (864×540)
wallpaper_sad_reimu_sd

Ah, one major mistake on this one!! I forgot the credits!! I’m posting them here:
Touhou Project © Team Shanghai Alice
Model by ISAO, Stage by nya, Effects by 針金P (HgPointLight), そぼろ (SvSSAO and Diffusion7)
Wallpaper made by Dioxaz © 2016

I’ll probably fix that later. I’m too lazy for now.

Monochrome Sonic – Dioxaz’s wallpapers, 2016 remake edition

Oh noes! Another one! And one of my favourite pieces on top of that. This reminds me of the days when I was called the “maniac of PNG”. Again, another sign of the times as almost everybody uses it now. This format almost replaced GIF in all the situations where GIF was used before. I think we call that a victory (in fact, I think it’s due to Internet Explorer 6 successors increasing market share starting from 2007 and the release of IE7 which finally introduced full alpha transparency support with PNG, without using any DX transformation plug-in).

Another very easy case because almost everything was in vector form (I had to convert only a few things). The tricky part was more the restricted composition by itself, which didn’t allow playing much when having headroom. Hence my decision to offer one more resolution, targeted for both SXGA (1280×1024) and WXGA+ displays (1440×900).

About the wallpaper, it just looks so great it’s unbelievable. I could almost have made it now. I can’t remember what inspired me to do it at the time. All I can say is that I was doing a two-year technical degree in IT at the time (yes, I’m way older than you think).

And now, the beast (click on thumbnails for original size):

– 2K format (2560×1600)
wallpaper_sonic_04_2560

– FHD format (1920×1200)
wallpaper_sonic_04_1920

– For older small widescreen and standard monitors (1440×1024)
wallpaper_sonic_04_1440

– HD format (1366×800)
wallpaper_sonic_04_1366

– SD format (1024×600)
wallpaper_sonic_04_sd

In case you didn’t notice it, there’s a good reason why it’s monochrome. This way it can compress as efficiently as it was greyscale, but instead uses a 256-colour palette made of shades of one colour. I think nobody other than me can produce such PNGs without messing up (palette must be kept in order and compression must be made with filter 5, using Ken Silverman’s pngout for example).

Adventure DX Sonic – Dioxaz’s wallpapers, 2016 remake edition

Yay, another one! Originally made in 2004. I’ll probably do one or two more and stop there. As you can see, another relatively easy case with lots of vector graphics (title bands at the bottom) and text. The trickiest part were the repeated “SONIC” patterns at the background (both the straight and rotated one), as they were already bitmapped (Photoshop doesn’t let you fill with a vector pattern yet, to my knowledge). I obviously had to recreate those parts with a higher resolution source (thank you http://info.sonicretro.org/). Then you have the artwork itself and the logo. You may remember how it was not the best quality (getting art assets was very hard back then and you often had to scan or extract them yourself, let alone getting them in professional quality!). Now I stumbled across this wonderful resource which is Sonic Scanf. Its official art gallery (http://en.sonicscanf.org/gallery/official-art/) can be easily designated as a true spiritual successor to my now abandoned Sonic Art Archive (Sonic Retro also has worthwhile pieces of art in high resolution). Sure it still has some uncredited and unauthorized reprints from my old website but I don’t really care now. In fact, its webmaster made the correct decision IMHO. This gallery “saved my life” as it provided huge resolution PNGs of both the artwork and the SADX logo. Its webmaster earns good points for maintaining his gallery the same way I maintained mine ten years ago (it has the latest pieces available). This is an indicator of how times have changed in ten years.

About the wallpaper, this is one of those pieces that still look great when I look at it, even a decade later. The same cannot be said about countless attempts made by inexperienced fans back in the day.

Here it is for your greatest pleasure (click on thubnmails for original size):

– 2K format (2560×1600)
wallpaper_sonic_03b_1920

– FHD format (1920×1200)
wallpaper_sonic_03b_1920

– HD format (1366×800)
wallpaper_sonic_03b_1366

– SD format (1024×600)
wallpaper_sonic_03b_sd

Did this piece really stand the test of time? You are the judge. It sure has a retro vibe now anyway.

Amy at the races – Dioxaz’s wallpapers, 2016 remake edition

Prefiguring what I’m planning to do with wallpaper-making, I decided to remake an old piece I did in 2003 (13 years ago!) but in widescreen and 2K format, instead of the completely obsolete 1024×768. As I want to cover as many resolutions I can, I decided I’ll go with crop-only wallpapers, but with several sizes. This way you can pick up the closest size to your monitor resolution and crop it to fill your screen.

Probably one of the easiest to remake, as only simple elements are featured (text, which remained in vector form and a chequered pattern, easy to resize). The only tricky part is the Amy artwork, as I only have it in low resolution. To upscale it, I used nnedi3_rpow both on the RGB part of the picture and the alpha channel separately (nnedi3 doesn’t support the RGB32 colorspace). The Amy artwork is a part of Sonic Screensaver, a goody accompanying the PC port of Sonic 3 & Knuckles (Sonic 3 & Knuckles Collection).

Here’s the result (click on thubnmails for original size):

– 2K format (2560×1600)
wallpaper_amy_race_2560

– FHD format (1920×1200)
wallpaper_amy_race_1920

– HD format (1366×800)
wallpaper_amy_race_1366

– SD format (1024×600)
wallpaper_amy_race_netbook

This is how, folks, I plan to do my future wallpapers if I ever feel like doing it again. Also, I’m definitely not one of those artists who constantly spit on their old works and are too afraid to see or show them again, not feeling any hint of nostalgia. My old works are a full part of me and still reflect my own personality to this day. Besides, I still love watching them again (saying to myself “I can’t believe I did that when I was young. It looks so great even now!”)