<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Cory Chu on Medium]]></title>
        <description><![CDATA[Stories by Cory Chu on Medium]]></description>
        <link>https://medium.com/@corychu?source=rss-ac44afc6d355------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*Tkm90J-WGVM5OSSSDzrLkA@2x.png</url>
            <title>Stories by Cory Chu on Medium</title>
            <link>https://medium.com/@corychu?source=rss-ac44afc6d355------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 19:47:49 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@corychu/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Sony WH-1000XM6 sound terrible with a Mac or iPhone during calls? Here’s Why]]></title>
            <link>https://corychu.medium.com/sony-wh-1000xm6-sound-terrible-with-a-mac-or-iphone-during-calls-heres-why-eda3f8df376b?source=rss-ac44afc6d355------2</link>
            <guid isPermaLink="false">https://medium.com/p/eda3f8df376b</guid>
            <category><![CDATA[apple]]></category>
            <category><![CDATA[active-noise-cancellation]]></category>
            <category><![CDATA[bluetooth]]></category>
            <category><![CDATA[sony]]></category>
            <category><![CDATA[wireless-headphone]]></category>
            <dc:creator><![CDATA[Cory Chu]]></dc:creator>
            <pubDate>Thu, 17 Jul 2025 01:30:18 GMT</pubDate>
            <atom:updated>2025-07-18T09:55:55.543Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*LfPLnYulwuQBBcJoNa0Ldw.jpeg" /><figcaption>Sony WH-1000XM6</figcaption></figure><p>I purchased <strong>Sony WH-1000XM6</strong> headphones recently and thought it would be a good fit with my MacBook Pro or iPhone. Unfortunately, it doesn’t, especially for calls like Zoom meetings, FaceTime, Teams, and Google Meet. The sound quality is awful.</p><p>The root cause is that Sony and Apple do not work together. The technology that enables high-quality audio for the microphone and simultaneous listening through Bluetooth on Mac is only available on AirPods (and possibly Beats), not on Sony wireless headphones.</p><p>The most obvious way to see this is when you talk to Siri via your headphones’ microphone. If you use AirPods, it’ll be alright. However, if you use the WH-1000XM6, when Siri pops up, you can immediately notice that the sound quality drops instantly to something like a walkie-talkie. That is caused by a “Bluetooth audio profile switching.”</p><h3>Profiles of Bluetooth</h3><p>So, basically, there are two Bluetooth profiles relevant to our discussion:</p><ul><li><strong>A2DP</strong> (Advanced Audio Distribution Profile): high-quality stereo audio, but <strong>no microphone.</strong></li><li><strong>HFP</strong> (Hands-Free Profile): supports microphones, but reduces audio quality.</li></ul><p>When playing music, Mac will use A2DP. On the other hand, when we try to use the mic to make a call, Mac will switch to the HFP, and that’s the reason why it sounds bad.</p><h3>A quick test using Audio MIDI Setup</h3><p>There is a built-in app on macOS called “Audio MIDI Setup.”</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/400/0*CQVFulj9ERLcSMXB.png" /><figcaption>Audio MIDI Setup</figcaption></figure><p>We can check what bit-depth and sample rates our input/output devices are using in this app. Especially, we can see how they switch to different setups when starting or ending a call. You can click on different devices and see the information as shown in this screenshot:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/636/1*gHe-_tHiVCqrq7F1xXBh5A@2x.png" /><figcaption>Example of WH-1000XMS Mic during call</figcaption></figure><p>I’ve tried to compare AirPods Pro and WH-1000XM6 during normal music playback or a call. The result goes like this:</p><h4><strong>During normal playback</strong></h4><ul><li><strong>AirPods Pro</strong><br>Output: 2ch 32-bit Float 48kHz</li><li><strong>WH-1000XM6</strong><br>Output: 2ch 32-bit Float 44.1kHz</li></ul><h4>When using a microphone, such as a Zoom call</h4><ul><li><strong>AirPods Pro</strong><br>Mic-In: 1ch 32-bit Float 24kHz<br>Output: 1ch 32-bit Float 24kHz</li><li><strong>WH-1000XM6</strong><br>Mic-In: 1ch 16-bit Integer 16kHz<br>Output: 1ch 16-bit Integer 16kHz</li></ul><p>So, as you can tell, during a call, the bit-depth and sample rate of the WH-1000XM6 audio just dropped to 16-bit / 16kHz, and that’s why they sound awful.</p><p>It’s also worth mentioning that for AirPods Pro, the sampling rate dropped from 48 kHz to 24 kHz during calls. The change is clearly noticeable, but at least it’s not terrible to hear.</p><h3>Workaround on Mac</h3><p>There is a small tip to sidestep the issue. During a call, you can select the output device to your headphones, but use the Mac’s built-in mic as the input device. This can be done by holding the [option] key and clicking on the “headphone or speaker” icon in the menu bar. That will enable you to select the input and output devices separately:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/299/1*4X6LIHzPKmi859biUjMFow@2x.png" /><figcaption>[option] + click on headphone or speaker icon in the menu bar</figcaption></figure><p>Just choose Mac’s Microphone for the Input, and you are good to go 💪💪💪</p><h3>Updates</h3><p>People on Reddit: <a href="https://www.reddit.com/r/SonyHeadphones/comments/1m1w0mm/sony_wh1000xm6_sound_terrible_with_a_mac_or/">https://www.reddit.com/r/SonyHeadphones/comments/1m1w0mm/sony_wh1000xm6_sound_terrible_with_a_mac_or/</a><br>suggests that USB Bluetooth dongles like <a href="https://www.flairmesh.com/Dongle/FMA120.html">FlooGoo FMA120</a> might be a solution, given that they said they support LE Audio with LC3. But I haven’t verified this personally.</p><p>Another post on Reddit did a deep test on the FMA120 + XM6 setup:<br><a href="https://www.reddit.com/r/SonyHeadphones/comments/1kw4hqc/wh1000xm6_via_lc3_a_limited_experience_for_gaming/">https://www.reddit.com/r/SonyHeadphones/comments/1kw4hqc/wh1000xm6_via_lc3_a_limited_experience_for_gaming/</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=eda3f8df376b" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Better than Notion? A fast-evolving visual note-taking app-Heptabase]]></title>
            <link>https://corychu.medium.com/a-fast-evolving-visual-note-taking-app-heptabase-175acd3b3462?source=rss-ac44afc6d355------2</link>
            <guid isPermaLink="false">https://medium.com/p/175acd3b3462</guid>
            <category><![CDATA[obsidian]]></category>
            <category><![CDATA[learning]]></category>
            <category><![CDATA[notion]]></category>
            <category><![CDATA[notetaking]]></category>
            <category><![CDATA[heptabase]]></category>
            <dc:creator><![CDATA[Cory Chu]]></dc:creator>
            <pubDate>Sun, 22 Dec 2024 20:35:54 GMT</pubDate>
            <atom:updated>2025-06-21T05:16:33.116Z</atom:updated>
            <content:encoded><![CDATA[<p>Heptabase is the first and the only note-taking app I’m paying for, and let me tell you why:</p><ol><li>Basically, Heptabase is an app that allows you to put something like Pages in Notion (which are known as Note-Cards in Heptabase) onto Whiteboards.</li><li>In addition to Note-Cards, it allows you to drop PDF files onto the whiteboard as a type of Card. Furthermore, the highlights on PDF can be dragged onto the whiteboard as well, making it a perfect tool for organizing and studying information stored in PDFs.</li><li>It has a very good community on Discord that really listen to all the user’s feedback! Also, their in-app support is always straight to the point and solves my problem like an engineer instead of a customer service.</li><li>They state that even if you terminate the subscription at some point, you can still access all of your notes in the app. The only difference is that you may not edit or create new content.</li></ol><p>Disclosure: Links to Heptabase in this post are affiliate links.</p><h3>Whiteboard</h3><p>A whiteboard is the space where you may throw things onto it. It can be as organized as what has been shown on the <a href="https://get.heptabase.com/cory-blog-tp">Heptabase official website</a>. But, to be honest, I’m a messy person, and most of the time, my whiteboard is kind of a mess at the early stage of acquiring information. For instance, in this whiteboard, I’m just throwing some information about DIY circuits like some catalogs of oscilloscopes that I’d like to compare and purchase. It is such a whiteboard, just like a desk, that I can put things onto and gradually organize them:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*92ZfToA5llrrU1J-jkk2sQ@2x.png" /><figcaption>Whiteboard</figcaption></figure><h3>Cards</h3><p>The cards are the blocks you saw on the above screenshot of the whiteboard. You may create various types of cards in Heptabase. Especially, the Note-type Cards work like Pages in Notion. If you have ever used Notion, you’ll know what I mean. Here is a quick demonstration of Note Cards:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*1SRe4E1xnw9JP8GMnaK0iw@2x.gif" /><figcaption>Note Cards Demonstration Animation</figcaption></figure><p>You may also see a few cards like PDF documents. Those are another type of Cards: “PDF-Cards.” These PDF cards allow you to make highlights and drag those highlights onto the whiteboard, which also has a backlink so that you can jump to the highlighted place in the original PDF, as demonstrated in this video:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ZmOtWoIvdgBZH6K2Ix5A0g@2x.gif" /><figcaption>PDF Cards Hightlighitng Demonstration Animation</figcaption></figure><h3>Tags</h3><p>Tag is another extremely useful feature. After adding a Tag to a Card, you can add various properties associated with this tag and create a table view or a Kanban view. For example, I can have a tag called “IC-Datasheet” and have properties associated with this tag like “ModelNumber,” “URL,” and “Status,” …… After that, I can create a Kanban view showing “Status” in various values (Pending, Reviewing, Purchased, ……) as columns. The table view and Kanban view look like this:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*LKerT1y7lG360TWSQNpYFA@2x.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*vcnlgPdYnJSQeTsLjTJwnw@2x.png" /></figure><h3>Other Functions</h3><p>There are other functions like Journal, Mindmap, <a href="https://chromewebstore.google.com/detail/heptabase-web-clipper/mlnkfjffdpccejmkcdhhbjfmdfokmkjh">Web Clipper Chrome Extension</a>, and Readwise-Integration…. Also, recently, they introduced a feature that allows you to import a <a href="https://www.youtube.com/watch?v=iNaU33cjTSM">Podcast</a> or <a href="https://www.youtube.com/watch?v=nk7BNIuZAP8">YouTube</a> as a type of Card, and you can <a href="https://www.youtube.com/watch?v=vSKfCnElaWg">generate transcripts</a> with AI. There are also mobile versions available on <a href="https://apps.apple.com/tw/app/heptabase/id6445801508?l=en">iOS</a> or <a href="https://play.google.com/store/apps/details?id=com.heptabase.heptabase">Android</a>.</p><h3>Afterword</h3><p>Currently, I feel every single function in Heptabase is like 80% polished. But the integration of all of them into a single app makes it such a unique tool and makes it just good enough to solve the problem I’m facing — Learning a complex subject and organizing all kinds of fragmented information. It’s not perfect, but it&#39;s good enough to convince me to keep paying its subscription fee since February 2023 till now.</p><p>I would suggest just <strong>downloading it for a free 7-day trial</strong> or even subscribing for a month, and you can tell if it is the right tool for you!</p><p>If you like my article, you may use this link:<br><a href="https://join.heptabase.com/?invite-acc-id=af0799d9-62c0-4bba-8d58-2f5e308df2d5">https://join.heptabase.com?invite-acc-id=af0799d9-62c0-4bba-8d58-2f5e308df2d5</a> <br>to receive a <strong>$5 discount</strong> when signing up for a new Heptabase account.</p><p>Finally, if you are interested in learning more about the long-term roadmap or the design rationale behind Heptabase, I highly recommend an article written by a co-founder of Heptabase — Alan Chan — <a href="https://medium.com/heptabase/my-vision-the-roadmap-d2fbf92fdf97">My Vision: The Roadmap</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=175acd3b3462" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to parse XML file exported from Apple iOS Health App and make a sleep schedule plot using…]]></title>
            <link>https://blog.gwlab.page/how-to-parse-xml-file-exported-from-apple-ios-health-app-and-make-a-sleep-schedule-plot-using-60c652697c81?source=rss-ac44afc6d355------2</link>
            <guid isPermaLink="false">https://medium.com/p/60c652697c81</guid>
            <category><![CDATA[pandas]]></category>
            <category><![CDATA[python]]></category>
            <category><![CDATA[matplotlib]]></category>
            <category><![CDATA[apple-watch]]></category>
            <category><![CDATA[apple]]></category>
            <dc:creator><![CDATA[Cory Chu]]></dc:creator>
            <pubDate>Sat, 16 Jul 2022 08:49:43 GMT</pubDate>
            <atom:updated>2022-07-17T20:28:22.013Z</atom:updated>
            <content:encoded><![CDATA[<h3>How to parse XML file exported from Apple iOS Health App and make a sleep schedule plot using Python, Pandas, and Matplotlib</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6lagsAJTrItTQzDVXodKkQ@2x.png" /><figcaption>Sleep Schedule plot by matplotlib</figcaption></figure><h3>Source Code</h3><p>The complete IPython Notebook to make the plot is available at my GitHub repo: <a href="https://github.com/c0rychu/apple-sleep">https://github.com/c0rychu/apple-sleep</a></p><h3>Export health data as an XML file</h3><p>Apple iOS Health App allows you to export all your health data. The steps are straightforward:</p><ol><li>Open your Health app and <strong>click on your profile picture</strong> in the top-right corner.</li><li>Select [<strong>Export All Health Data</strong>], which appears at the bottom of the menu.</li><li>After a few minutes, you can send the exported export.zip to your computer via, e.g., AirDrop.</li></ol><p>After unzipping the export.zip, you may find different types of files. For instance, workout routes are stored as .gpx files that are GPS tracking information. The relevant file for plotting the sleep schedule is in the export.xml file.</p><h3>Parse the XML file in Python</h3><p>You may look at the export.xml file by opening it in a text editor. The main structure of this XML file looks like this:</p><pre>&lt;HealthData locale=&quot;en_US&quot;&gt;<br> &lt;ExportDate value=&quot;2022-07-14 17:53:49 -0500&quot;/&gt;<br> &lt;Me ...xxxxxxxxx... /&gt;<br> &lt;Record ...xxxxxx... &gt;<br>  &lt; ...xxx... /&gt;<br> &lt;/Record&gt;<br> &lt;Workout ...xxx... /&gt;<br> &lt;Record ...xxxxxx... &gt;<br> &lt;Record ...xxxxxx... &gt;<br> &lt;Record ...xxxxxx... &gt;<br>&lt;/HealthData&gt;</pre><p>The root tag is &lt;HealthData&gt;, and there are different kinds of tags under it. Especially, we want to acquire the data in those &lt;Record&gt; tags belonging to SleepAnalysis, i.e., with type=”HKCategoryTypeIdentifierSleepAnalysis”, for example:</p><pre>&lt;Record <br>  type=&quot;HKCategoryTypeIdentifierSleepAnalysis&quot;<br>  sourceName=&quot;MyAppleWatch&quot;<br>  sourceVersion=&quot;2021111&quot;<br>  creationDate=&quot;2021-12-01 17:14:48 -0500&quot;<br>  startDate=&quot;2021-12-01 08:38:00 -0500&quot;<br>  endDate=&quot;2021-12-01 08:43:59 -0500&quot;<br>  value=&quot;HKCategoryValueSleepAnalysisAsleep&quot;<br>/&gt;</pre><p>To do so, first, we use the xml package to parse the XML file:</p><pre>import <a href="https://docs.python.org/3/library/xml.etree.elementtree.html">xml.etree.ElementTree</a> as ET</pre><pre>tree = ET.parse(&quot;apple_health_export/export.xml&quot;)<br>root = tree.getroot()</pre><p>You can check the tag of root, which is “HealthData” as we saw in the XML file:</p><pre>&gt;&gt;&gt; root.tag<br>&#39;HealthData&#39;</pre><p>To iterate over the child &lt;Record&gt; tags under root, and append them to a list, we can perform a list comprehension:</p><pre>records = [i.attrib for i in root.iter(&quot;Record&quot;)]</pre><p>Then, we may convert records into Pandas.DataFrame, and cast the string-type date/time data into datetime datatype:</p><pre>import pandas as pd</pre><pre>records_df = pd.DataFrame(records)</pre><pre>date_col = [&#39;creationDate&#39;, &#39;startDate&#39;, &#39;endDate&#39;]<br>records_df[date_col] = records_df[date_col].apply(pd.to_datetime)</pre><p>Now, we can select only the Sleep Analysis data in our records:</p><pre>sleeps_df = records_df.query(&quot;type == &#39;HKCategoryTypeIdentifierSleepAnalysis&#39;&quot;)</pre><h3>Cut each overnight record into two records</h3><p>We may have some sleep records starting from one day but ending on the next day. To make the plotting process easier, we can cut those overnight records into two:</p><ol><li>Separate sleep records into two Pandas.DataFrame, the no_cross and cross, where cross means overnight records.</li><li>Make two duplications c1 and c2 of cross.</li><li>Set the end DateTime of the records in c1 as 23:59:59 of the Starting Date of each record, and set the start DateTime of the records in c2 as 00:00:00 of the Ending Date of each record.</li><li>Concat the no_cross, c1, and c2, and sort them by the Starting DateTime.</li></ol><pre>no_cross = sleeps_df[sleeps_df[&quot;startDate&quot;].dt.day == sleeps_df[&quot;endDate&quot;].dt.day]</pre><pre>cross = sleeps_df[sleeps_df[&quot;startDate&quot;].dt.day != sleeps_df[&quot;endDate&quot;].dt.day]</pre><pre>c1 = cross.copy()<br>c2 = cross.copy()</pre><pre>c1[&quot;endDate&quot;] = c1[&quot;startDate&quot;].apply(lambda x: x.replace(hour=23, minute=59, second=59))</pre><pre>c2[&quot;startDate&quot;] = c2[&quot;endDate&quot;].apply(lambda x: x.replace(hour=0, minute=0, second=0))</pre><pre>sleeps_splitted_df = pd.concat([no_cross, c1, c2]).sort_values(&quot;startDate&quot;)</pre><h3>Plotting using Matplotlib</h3><p>Finally, we can make our plot by using matplotlib . Some points are worth mentioning:</p><ol><li>We used <a href="https://matplotlib.org/stable/api/_as_gen/matplotlib.patches.Rectangle.html">matplotlib.patches.Rectangle((</a><a href="https://matplotlib.org/stable/api/_as_gen/matplotlib.patches.Rectangle.html"><em>x, y)</em></a><a href="https://matplotlib.org/stable/api/_as_gen/matplotlib.patches.Rectangle.html">, </a><a href="https://matplotlib.org/stable/api/_as_gen/matplotlib.patches.Rectangle.html"><em>width</em></a><a href="https://matplotlib.org/stable/api/_as_gen/matplotlib.patches.Rectangle.html">, </a><a href="https://matplotlib.org/stable/api/_as_gen/matplotlib.patches.Rectangle.html"><em>height</em></a><a href="https://matplotlib.org/stable/api/_as_gen/matplotlib.patches.Rectangle.html">)</a> to plot our data.</li><li>We used <a href="https://pandas.pydata.org/docs/reference/api/pandas.DataFrame.itertuples.html">pandas.DataFrame.itertuples()</a> to iterate over records.</li><li><a href="https://matplotlib.org/stable/api/dates_api.html">Matplotlib uses a floating number representation of daytime in its axes</a>. We used <a href="https://matplotlib.org/stable/api/dates_api.html#matplotlib.dates.date2num">matplotlib.dates.date2num</a> to convert our DateTime data into its representation.</li><li>Somehow, if the input time delta is too small (e.g., our y-axes are only 24 hours), we need to divide the value by 1000 to get the correct Ticks label. That’s the reason why we have some /1000 in the code.</li><li>We found the Awake data is not particularly useful for our application. So we skipped them in our plot.</li><li>The reference to Ticks Formatter and Locator can be found in the <a href="https://matplotlib.org/stable/api/dates_api.html">documentation of </a><a href="https://matplotlib.org/stable/api/dates_api.html">matplotlib.dates</a> . The format string used in DateFormatter like %I%p are <a href="https://strftime.org/">strftime format strings and can be found here</a>.</li></ol><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/e5d993843290ae547150b2ea33ce9b9b/href">https://medium.com/media/e5d993843290ae547150b2ea33ce9b9b/href</a></iframe><h3>Again, the Full Source Code</h3><p>Again, The complete IPython Notebook to make the plot is available at my GitHub repo: <a href="https://github.com/c0rychu/apple-sleep">https://github.com/c0rychu/apple-sleep</a></p><p>Hope you’ll enjoy~</p><h3>Reference</h3><ul><li><a href="https://stackoverflow.com/questions/72989080/how-to-make-plots-in-python-like-the-sleep-schedule-plot-in-ios-health-app">https://stackoverflow.com/questions/72989080/how-to-make-plots-in-python-like-the-sleep-schedule-plot-in-ios-health-app</a></li><li><a href="https://www.python-engineer.com/posts/apple-health-data-python/">https://www.python-engineer.com/posts/apple-health-data-python/</a></li><li><a href="https://www.relevantmisc.com/r/python/2020/05/26/visualizing-baby-sleep/">https://www.relevantmisc.com/r/python/2020/05/26/visualizing-baby-sleep/</a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=60c652697c81" width="1" height="1" alt=""><hr><p><a href="https://blog.gwlab.page/how-to-parse-xml-file-exported-from-apple-ios-health-app-and-make-a-sleep-schedule-plot-using-60c652697c81">How to parse XML file exported from Apple iOS Health App and make a sleep schedule plot using…</a> was originally published in <a href="https://blog.gwlab.page/">GWLab</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Flaws/Problems of Garmin Watches you should know before buying them]]></title>
            <link>https://corychu.medium.com/flaws-of-garmin-watches-you-should-know-before-buying-them-bc2aa6848ff?source=rss-ac44afc6d355------2</link>
            <guid isPermaLink="false">https://medium.com/p/bc2aa6848ff</guid>
            <category><![CDATA[smartwatch]]></category>
            <category><![CDATA[garmin]]></category>
            <category><![CDATA[iphone]]></category>
            <dc:creator><![CDATA[Cory Chu]]></dc:creator>
            <pubDate>Sat, 09 Jul 2022 03:00:18 GMT</pubDate>
            <atom:updated>2022-07-13T01:29:16.681Z</atom:updated>
            <content:encoded><![CDATA[<p>Garmin made some excellent smartwatches. But there are some issues that you should know before you decide whether to buy them. I realized them after I received my Garmin Tactix 7 - Pro Edition. Now, I just returned it</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uGoPmULPZlzncFs6p-sgUg.jpeg" /><figcaption>Garmin tactix® 7 — Pro Edition, I really like its looks, but its UX design is awful</figcaption></figure><h3>1. It cannot record your multiple sleeping periods or daytime naps a day</h3><p>Garmin watches only record one sleeping period (probably your nighttime sleep) per day. If you travel a lot, as a shift worker, have an irregular sleeping schedule, or want to take some naps during the daytime, it just won’t record any of them. Many people cannot have a regular sleep at night no matter that is because they need to work at midnight or because of some medical condition. Garmin just doesn’t let you record those fragmented sleeping.</p><p>Since 7 years ago, this problem of tracking sleep for shift workers has already <a href="https://forums.garmin.com/apps-software/mobile-apps-web/f/garmin-connect-web/89095/shift-work-and-sleep-tracking">been discussed on Garmin’s forums</a>, but it just never got resolved.</p><p>On the other hand, if you have a regular sleeping schedule, although Garmin watches can track that, it’s probably not that important to track your sleep anyway.</p><h3>2. It won’t detect your activity and ask you whether you want to start recording your running/walking/biking… automatically, if you forget to</h3><p>If you do not manually select what activity you are doing and start recording manually, Garmin watches only tag such activity as a Move IQ Event. Move IQ Events neither shows up as an activity in Garmin Connect App nor records complete information such as GPS tracking of your activity.</p><p>Even some cheap $30 smart band will detect your activity and pops up to ask whether you want to start recording your running/walking/biking…</p><p>Of course, this shouldn’t be a problem if you are doing some training on purpose. But this can be very useful to let you record your commuting walk/biking when you are very likely forgetting to press the start tracking button.</p><h3>3. You cannot choose which notification from your iPhone shows up</h3><p>Suppose you use iPhone and want to manage the notification from which app should appear in the Garmin watches. In that case, the only way is to turn the notification on/off for the iOS notification center as described on <a href="https://support.garmin.com/en-US/?faq=TLeDN92ZU0AgN4df6HakwA">Garmin’s website</a>. This can be super frustrating because we usually want to keep unimportant notifications in the iOS notification center while only sending important notifications to our smartwatches.</p><p>Again, even some shitty cheap smart band’s app allows you to do so on iOS. So, it’s really not a technical restriction, and I have no idea why Garmin does not take care of this kind of basic thing in their iOS app.</p><h3>4. It cannot select which calendar to sync from your iPhone</h3><p>I used to access my Google calendars in the iOS built-in calendar app. However, there are some unimportant calendars in my Google accounts. So, I’ve hidden them in my iOS calendar app. But the big problem with Garmin smartwatches is that they sync all calendars on iOS even if you already disable (hide) some calendars in your iOS calendar app.</p><h3>Final thoughts — Fantastic hardware with poor software/firmware</h3><p>Finally, I’d like to say, although the built quality, the battery life, the physical buttons, and even the flashlight of the Garmin Tactix7 are excellent, its general user experience is not ideal. The menu system is unnecessarily complicated. And the UI response speed is not very good. If you have ever used <a href="https://www.amazon.com/Seiko-S141-300-Memory-Stopwatch/dp/B0069Q3CFE">a professional stopwatch like Seiko S141</a>, you may be disappointed that how come a $1000 smartwatch’s UI/UX in 2022 still works like a <a href="https://en.wikipedia.org/wiki/Nokia_3310">Nokia 3310</a>. I even doubt whether they have some good UX designers in their company. I hope that Garmin will improve the UX design of their products such that I can love their products in the future again. And I can say Garmin Tactix7 is an excellent watch, just not smart at all.</p><h3>Take action</h3><p>If you also share the same feeling that we do love Garmin, but we really want Garmin to improve their products, please don’t hesitate to submit the request to Garmin: <a href="https://www.garmin.com/en-US/forms/ideas/">https://www.garmin.com/en-US/forms/ideas/</a></p><p><a href="https://www.garmin.com/en-US/forms/ideas/">Share</a></p><p>The more customers give them applicable/specific feedback, the more chance we can get great smartwatches from them in the future!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=bc2aa6848ff" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Series Solutions to an ODE, Ordinary Point, and Regular Singular Point — the Frobenius’s Method]]></title>
            <link>https://blog.gwlab.page/series-solutions-to-an-ode-ordinary-point-and-regular-singular-point-the-frobeniuss-method-b12a0adde8ca?source=rss-ac44afc6d355------2</link>
            <guid isPermaLink="false">https://medium.com/p/b12a0adde8ca</guid>
            <category><![CDATA[mathematics]]></category>
            <category><![CDATA[differential-equations]]></category>
            <category><![CDATA[solutions]]></category>
            <category><![CDATA[ode]]></category>
            <category><![CDATA[physics]]></category>
            <dc:creator><![CDATA[Cory Chu]]></dc:creator>
            <pubDate>Thu, 23 Dec 2021 18:08:30 GMT</pubDate>
            <atom:updated>2021-12-23T18:08:30.837Z</atom:updated>
            <content:encoded><![CDATA[<h3>Series Solutions to an ODE, Ordinary Point, and Regular Singular Point — the Frobenius’s Method</h3><p>This article is based on my unpublished book about Mathematics for Physics.</p><p>When you learned about the series solutions to the ordinary differential equation (ODE), you probably heard about the ordinary points and regular singular points. But in many cases, the explanation of the logic or motivation is not adequate. I hope that I can do a better job here for you.</p><p>So, what is a series solution to an ODE? Let’s consider a homogeneous linear differential equation:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*hDWZivfaqdW5PlqQUHHqxQ@2x.png" /></figure><p>, where the leading order coefﬁcient is chosen to be unity. Roughly speaking, instead of ﬁnding the functional form of the solution <em>y</em>(<em>x</em>), we would like to assume our solution can be expended about a certain point x₀ as something like a Taylor series:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/276/1*a9DaCBu_BBt_oK4XD22R0A@2x.png" /></figure><p>By doing so, we are turning our problem of ﬁnding <em>y</em>(<em>x</em>) into ﬁnding those coefﬁcients <em>aₙ</em> in our series. However, if the actual solution <em>y</em>(<em>x</em>) is not well-behaved (analytic) at the point <em>x₀</em>, we may not be able to expand <em>y</em>(<em>x</em>) about this <em>x₀</em> as a Taylor series from the ﬁrst place. In that case, if you blindly expand it, you will ﬁnd that when you try to solve <em>aₙ</em>, you cannot ﬁnd reasonable <em>aₙ</em>. So, the series solution is not going to work in this case.</p><p>Now, you may ask: “<strong>how do we know whether the actual solution <em>y</em>(<em>x</em>) is well-behaved at the point <em>x₀</em> or not without actually knowing the solution <em>y</em>(<em>x</em>)?</strong>” (Of course, the reason why we use series solution is exactly that we don’t know the actual solution <em>y</em>(<em>x</em>)) Thanks to a great mathematician, <strong>Lazarus Fuchs</strong>, he told us how to know whether the actual solution <em>y</em>(<em>x</em>) is well-behaved or not at a certain <em>x₀</em> by looking at those coefﬁcients <em>A</em>_{<em>n −1</em>}(<em>x</em>), <em>A</em>_{<em>n −2</em>}(<em>x</em>), . . . , <em>A</em>_{0}(<em>x</em>) in our homogeneous linear differential equation, i.e., the first equation in this article, without actually knowing the solution <em>y</em>(<em>x</em>)<strong>.</strong></p><h3>Series solution about an Ordinary Point</h3><p>If <em>A</em>_{<em>n −1</em>}(<em>x</em>), <em>A</em>_{<em>n −2</em>}(<em>x</em>), . . . , <em>A</em>_{0}(<em>x</em>) are all analytic in a neighborhood of <em>x₀</em> in the complex plane, we call <em>x₀</em> is an <strong>ordinary point of this homogeneous linear differential equation</strong>. According to Fuchs, all n linearly independent solutions to this differential equation are analytic in a neighborhood of <em>x₀</em> and can be expanded as Taylor series:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/276/1*a9DaCBu_BBt_oK4XD22R0A@2x.png" /></figure><p>Therefore, we can differentiate this formal solution and obtain the expression of its derivatives:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/413/1*kJic9oc91bFoRD1hfBSmsg@2x.png" /></figure><p>By substituting them back to our original problem, and requiring the equation is satisfied for each order of (<em>x</em>-<em>x₀</em>)<em>ⁿ </em>we can solve those aₙ, putting them into the Tayor series, thereby obtaining our solutions <em>y</em>(<em>x</em>).</p><h3>Series solution about a Regular Singular Point</h3><p>If <em>x₀ </em>is not an ordinary point, however, if all of</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/634/1*uW_xyN74NHk71gRqoMdynQ@2x.png" /></figure><p>are analytic in the neighborhood of <em>x₀</em> in the complex plane, we call <em>x₀</em> is a <strong>regular singular point of this homogeneous linear differential equation</strong>. Although a solution <em>y</em>(<em>x</em>) to our ODE might still be analytic at a regular singular point <em>x₀</em>, it’s not necessary analytic. If it’s not, nevertheless, it doesn’t go that crazy at point <em>x₀</em>, and that’s the reason why we call it a “regular” singular point. Its singularity must be either a pole or an algebraic or logarithmic branch point [1]. As you learned before, e.g., if <em>y</em>(<em>x</em>) has an isolated pole at <em>x₀</em>, we cannot expand <em>y</em>(<em>x</em>) about <em>x₀</em> as a Taylor series. Instead, we expand <em>y</em>(<em>x</em>) as a Laurent series about <em>x₀</em> to catch up or mimic the behavior of the pole. Here, Fuchs tells us that there is at least one solution <em>y</em>(<em>x</em>) can be written as:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/262/1*1PUAS_BFTNJDI_ddwy3Pzg@2x.png" /></figure><p>, where <em>Y</em>(<em>x</em>) is analytic at <em>x₀</em>, therefore can be further expanded about <em>x₀</em> using Taylor series:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/408/1*c9xzyItg74ah7VcSp41UCQ@2x.png" /></figure><p>The wisdom here is to use (<em>x</em>-<em>x₀</em>)<em>ˢ</em> to catch up the singular or non-analytic behavior of <em>y</em>(<em>x</em>) at the regular singular point <em>x₀</em> and the number <em>s</em> is called the <strong>indicial exponent</strong> and is to be determined when we solve the differential equation. In general, the indicial exponent <em>s</em> need not be an integer. If <em>s</em> is not an integer, the solution <em>y</em>(<em>x</em>) has an algebraic branch point at <em>x₀</em>. If <em>s</em> is a negative integer, <em>y</em>(<em>x</em>) has a pole at <em>x₀</em>. If<em> s</em> is a positive integer or zero, <em>y</em>(<em>x</em>) is analytic at <em>x₀</em> and this series is reduced to a Taylor series. The last equation, which is a series expansion with power <em>n+s,</em> is called the <strong>Frobenius series</strong>, and the method using the Frobenius series to solve a homogeneous linear differential equation is called the <strong>Frobenius’s Method</strong>.</p><h3>Reference</h3><p>[1] <a href="https://web.physics.wustl.edu/cmb/">Carl Bender</a> and Steven Orszag, <a href="https://www.amazon.com/dp/0387989315">Advanced Mathematical Methods for Scientists and Engineers I: Asymptotic Methods and Perturbation Theory</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b12a0adde8ca" width="1" height="1" alt=""><hr><p><a href="https://blog.gwlab.page/series-solutions-to-an-ode-ordinary-point-and-regular-singular-point-the-frobeniuss-method-b12a0adde8ca">Series Solutions to an ODE, Ordinary Point, and Regular Singular Point — the Frobenius’s Method</a> was originally published in <a href="https://blog.gwlab.page/">GWLab</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[什麼是量子力學？ 為什麼叫作“量子”力學？]]></title>
            <link>https://blog.gwlab.page/%E4%BB%80%E9%BA%BC%E6%98%AF%E9%87%8F%E5%AD%90%E5%8A%9B%E5%AD%B8-%E7%82%BA%E4%BB%80%E9%BA%BC%E5%8F%AB%E5%81%9A-%E9%87%8F%E5%AD%90-%E5%8A%9B%E5%AD%B8-1f424b8cddf?source=rss-ac44afc6d355------2</link>
            <guid isPermaLink="false">https://medium.com/p/1f424b8cddf</guid>
            <category><![CDATA[量子力學]]></category>
            <category><![CDATA[introduction]]></category>
            <category><![CDATA[physics]]></category>
            <category><![CDATA[quantum-mechanics]]></category>
            <category><![CDATA[quantum]]></category>
            <dc:creator><![CDATA[Cory Chu]]></dc:creator>
            <pubDate>Wed, 15 Dec 2021 18:04:40 GMT</pubDate>
            <atom:updated>2026-09-19T11:17:40.975Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uoMBK0yJjTVPflHeew6W-A.jpeg" /><figcaption>圖為量子力學中，路徑積分（Path Integral）的推導</figcaption></figure><p>量子力學是一個能用來解釋微觀世界的理論，其中的薛丁格方程式（Schrödinger Equation）就像是牛頓定律的 <em>F=ma</em>，然而牛頓定律（古典力學）沒辦法完全適用於如原子、電子、中子、質子這般微小的東西，而量子力學提供了我們一個能夠預測這些微小粒子行為的理論。</p><h3>為什麼叫作 “量子” 力學?</h3><p>如果這是你第一次學習量子力學，或是有人問你什麼是量子力學，第一個你想問的問題大概就是為什麼他叫“量子”力學？ 一個簡單的答案是：<strong>量子力學是一個可以用來解釋各種量子化之實驗結果的理論。</strong>而這些量子化的現象是無法被古典力學所解釋的。“量子”力學因其可以“解釋量子化的現象”而得名，<strong>這個命名方式有點像“湯”鍋是可以用來“煮湯”的鍋子</strong>；反觀“牛頓”力學是因為由牛頓提出而得名，這比較像“大同”電鍋是指由大同發明的電鍋。</p><p>看到這裡，你可能會問，那什麼是量子化的現象呢？所謂的<strong>量子化</strong>就是指一個一個可以數的、離散的、非連續的概念，比如你可能學過在氫原子中，電子有許多的軌域（<em>1s</em>, <em>2s</em>, <em>2p</em>, 3s, 3p, 3d, ……），並對應到各個能階，這些能階是離散的、非連續的。並且，只有處於這些特定的能量（或稱作能階，因為像台階一樣一階一階的）時，電子才是穩定的環繞在原子核周圍。所以我們說，被氫原子核的庫倫吸引力束縛住的電子有著量子化的本徵能量（能階）。</p><p>不過話說回來，湯鍋雖然可以煮湯，但不見得要用來煮湯，同理，量子力學可以用來解釋量子化的現象，但不一定會給出量子化（離散化）的結果。事實上，離散的、非連續的概念並非被建構在量子力學的基本架構中，而是量子力學在討論特定系統時可能出現的一種結果。更具的講，<strong>量子化的現象是發生在當我們用量子力學來討論束縛系統時會出現的結果，或者你可以說量子化是來自於系統的邊界條件</strong>。比方說，氫原子的電子能階是量子化的，是由於電子被氫原子核的庫倫吸引力給束縛著。又或者你可能聽過電子繞行原子核的軌道角動量是量子化的，這是由於空間旋轉的週期性邊界條件所致（也就是如果我們把世界旋轉 360° 應該要長得差不多一樣，或用數學表示就是 exp{<em>i2π</em>}=exp{<em>i0</em>}=1）。事實上對於不受力的自由粒子來說，在沒被束縛住的情況下，也沒有特定位能造成的邊界條件，即便用上了量子力學來討論，他的能量與動量都不是量子化、不是離散化的，而可以是連續的數值！這就是一個量子力學不一定導致量子化結果的例子。</p><p>另外，由於邊界條件所造成的量子化或離散化並非那麼神秘，即便在古典力學中，也有這種因為邊界條件而導致離散化的例子。比方說兩端固定不動（固定不動是一種邊界條件）的小提琴弦，其容許震動的波長就是量子化（離散化）的，也就是你學過能夠形成駐波的那些波長。他發出的聲音，會是由頻率為基頻（最低的容許頻率，波長為兩倍弦長，零個節點）與其整數倍的頻率（泛音，在弦上有 1~N 個節點的駐波）之聲波的疊加。這完全是個量子化（離散化）發生在古典力學的例子，我們沒有用到量子力學！就好像大同電鍋也是能煮湯一樣，煮湯不是湯鍋的專利。不過在這裡要強調一下，湯鍋與大同電鍋只是是個比喻，並不代表古典力學也能解釋氫原子量子化的能階。</p><p>然而，在這裡我們要非常小心我們的用字遣詞，目前為止，我不斷地交替使用量子化的（Quantized）與離散化的（Discrete）這兩個字。而我們在此所討論的是 “量子化的（Quantized）” 這個字本來字面上的意思，當今，物理學家幾乎賦予了他一個新的意思，就是用 “量子化的（Quantized）”來代表 “量子力學化的（“Quantum-Mechanicalized”）”。因為後者聽起來太長了！所以在大多的情況下，如果你聽到什麼是量子化（Quantized）的，或把一個系統量子化（Quantization），有 99% 的機會他們指的是 “量子力學化的” 與 “量子力學化”。舉個例子，一個物理學的專有名詞 — — <strong>正則量子化（Canonical Quantization）</strong>指的是讓位置算符 <strong><em>x </em></strong>與動量算符 <strong><em>p </em></strong>不可對易的過程，也就是讓 <strong><em>xp</em></strong><em>-</em><strong><em>px </em></strong>= <em>i</em>ℏ 。這個 <em>xp-px</em><strong><em> </em></strong>在古典上應該要等於零，因為在古典力學中，<em>x </em>與 <em>p</em> 是數字，具有乘法交換律；透過使其乘法交換會差一個 <em>i</em>ℏ ，也就是讓 <em>x </em>與 <em>p </em>的乘法變得不具交換律，可以迫使他們變成算符，或是矩陣，進而 “量子力學化” 我們討論的系統。</p><p>順帶一提，你可能會想，那鼎鼎大名的光子（photon），也就是量子化的電磁波，是不是也是量子力學給出的一個結果呢？如果你帶著想搞懂光子的心，去上量子力學的課，可能會有點小失望，因為在大多物理系量子力學正課的討論範疇中，我們並沒有處理光子的問題。<strong>然而光子，或稱光量子，其實在歷史上正是整個量子力學或說量子論的起點！</strong>量子力學的提出，是啟蒙於普朗克（Max Planck）對黑體輻射的討論，再加上愛因斯坦對光電效應的思索，這些都是嘗試推翻當年認為光是一種古典波動的認知，並試圖用光是量子化的“光子”來解釋實驗現象的努力，進而發展成今天的量子力學。不過時至今日，我們發現要深入了解什麼是光子，其實需要仰賴一個比量子力學更進階的理論，稱作量子場論（Quantum Field Theory，QFT）。量子場論是量子化的古典場論。其中，針對帶電粒子的電磁交互作用，我們有一個量子場論的分支叫做量子電動力學（ Quantum Electrodynamics，QED），也就是量子化的電動力學（電磁學）。在 QED 中，光子是量子化的電磁場，而電子是量子化的「電子的狄拉克場（Dirac Field）」， <strong>這也解釋了為什麼全世界的電子都長得一樣（即電子們是全同粒子 Identical Particles ），因為所有的電子都是同一個量子化的狄拉克場的激發態！</strong></p><h3>量子力學是機率性的</h3><p>另一方面，你可能會聽過量子力學是機率性的（Probabilistic），在<strong>量子的世界，我們無法預測單一觀測的結果，我們只能預測單一觀測出現某個結果的機率</strong>。而這個機率性不像離散化只是個可能的結果，<strong>機率性是深深植入量子力學的基本架構當中的</strong>。讓我們稍微深入討論一下這是怎麼一回事。</p><p>在古典力學中，以一維運動為例，我們可以把一個正在運動的粒子狀態用 (<em>x</em>, <em>v</em>) 表示，其中 <em>x </em>是該粒子的位置，而 <em>v</em> 則是他的速度。對一個不受外力質量為 <em>m </em>的粒子而言，他會根據慣性定律等速運動向前行，若其起始位置是 <em>x₀</em>，隨著時間 <em>t</em> 增加，他的狀態可以表示成 (<em>x₀ + vt</em>, <em>v</em>)。如果我們在時間 <em>t₁</em> 時同時測量他的位置、動量、與動能，毫無疑問我們會得到 <em>x₀ + v t₁</em>、<em>mv</em> 、 與 <em>mv²/2</em>。另外值得一提的是，經過測量，該粒子依舊依循慣性不斷向前行，他的狀態 (<em>x₀ + vt</em>,<em> v</em>) 始終沒有因為測量而改變。不過，有趣的是，這在量子力學中就不是這麼一回事了。</p><p>在量子力學中，以氫原子中的電子為例，我們假設該電子處於能量最低的軌域，也就是你學過的 <em>1s</em> 軌域，我們可以用 | <em>1s</em> 〉來表示該電子的狀態。這個狀態向量 | <em>1s</em> 〉全權描述了這個粒子在空間中移動的狀態，就好像上面古典例子中的 (<em>x₀ + vt</em>, <em>v</em>) 完全描述了該古典粒子的移動狀態。（在這裡我們暫時不處理粒子自轉的部分）。如果你試圖測量這個電子的總力學能，也就是動能加位能，沒意外的，你會得到 E₁s = −13.6eV 這個你在課本上學過的氫原子基態能量值。而且在測量之後，這個電子依舊處於 | <em>1s</em> 〉狀態。看到這裡，你可能會想，這不是跟古典力學的例子一樣嗎？沒什麼新奇的東西。沒錯，到目前為止，的確看不是出量子力學特別的地方，但這是因為| <em>1s</em> 〉 狀態是測量力學能的<strong>本徵狀態</strong>（eigenstate） 或者說| <em>1s</em> 〉 是哈密頓算符（能量算符）的本徵態，即 | <em>1s</em> 〉 is an <strong>eigenstate </strong>of<strong> </strong>the Hamiltonian operator.</p><p>現在，我們準備要來量這個電子的位置了，這也就是有趣的開始。</p><p><strong>第一</strong>，如果你對這顆處於| <em>1s</em> 〉狀態的電子做一個測量位置的動作，你確實會得到一個位置值，就假設你量到的是 <em>x₁</em> 好了。然而在量子力學中，你沒辦法在實際測量前預測你會得到哪個位置值，你只能預測你在某個位置找到電子的機率。在測量前你可以預測你量到該 | <em>1s</em> 〉電子位置處於 <em>x₁</em> 的機率為 P₁s(<em>x₁</em>)=|〈 <em>x₁</em>|<em>1s</em> 〉|² ，其中〈 <em>x₁</em>|<em>1s</em> 〉代表了兩個狀態向量 |<em>1s</em> 〉與 | <em>x₁ </em>〉的內積，稱作的 |<em>1s</em> 〉到 | <em>x₁ </em>〉的<strong>轉移振幅</strong>（Transition Amplitude），或者說是 |<em>1s</em> 〉在 | <em>x₁ </em>〉方向上的分量。需注意，這裡的向量 |OOXX 〉是狀態向量，不是我們熟悉的三維向量，而是在一個無限維度的抽象空間中的向量。</p><p>再強調一次，你只能預測 |<em>1s</em> 〉電子出現在某個位置 <em>x₁ </em>出現的機率，而無法在真正進行某次位置測量前就先知道你鐵定會在某時刻量到某個位置值。這個量子力學中的機率性，正是說明你沒辦法畫出電子的“路徑”或“軌跡”，這也就是為什麼你在課本上總是看到 <em>1s</em>, <em>2s</em>, <em>2p</em>, … 這些電子軌域被畫成一坨電子雲包圍著原子核的樣子。而沒有人把他畫成像地球繞太陽那樣的橢圓軌道。可以參考下面的圖片：</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/527/0*PnczEo_dPikM4bQN.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*INiwHnq3PMvcpgbm.png" /><figcaption>s與p軌域的圖示 Source: <a href="https://chemistrygod.com/atomic-orbital">https://chemistrygod.com/atomic-orbital</a></figcaption></figure><p>於此容我岔題一下，你可能會想這些軌域 <em>1s</em>, <em>2s</em>, <em>2p</em>, …… 好像都是呈現不會動的一種機率分佈，難道在量子力學中，電子就只會這樣根據一個不隨時間而變得機率分佈隨機出現嗎？事實上，這些你學過的軌域是所謂的能量與軌道角動量的本徵態，但你應該也聽過有個東西叫做量子疊加，最著名的例子大概就是<a href="https://zh.wikipedia.org/wiki/%E8%96%9B%E5%AE%9A%E8%B0%94%E7%8C%AB">薛丁格的貓</a>。如果你考慮一個電子狀態是 1<em>s</em> 與 <em>2p_z</em>（三個 <em>2p</em> 軌域中的一個） 的疊加，或寫作 |<em>ψ</em> 〉=|1<em>s</em> 〉+|2p_z 〉，這個 |<em>ψ</em> 〉所對應的電子位置機率分佈其實是會隨時間改變的，可以參考下列動畫：</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/512/0*0ZrYETHe49O-EMvw.gif" /><figcaption>1s 與 2p_z 的疊加之電子位置機率分佈隨時間演化 source:<a href="http://web.uvic.ca/~dharr/courses/chem347/hsuperposition/hsuperposition.htm">http://web.uvic.ca/~dharr/courses/chem347/hsuperposition/hsuperposition.htm</a></figcaption></figure><p>至於如何計算他該怎麼動，可能就要真的讀一下量子力學了 🥺 😢。好，拉回我們故事的主軸。</p><p><strong>第二</strong>，剛剛我們對顆處於| <em>1s</em> 〉狀態的電子做一個測量位置的動作而得到了 <em>x₁ </em>這個位置值，就在我們測量的過程，與古典力學不同的是，我們無可避免的改變了這顆電子的狀態，在我們得到 <em>x₁ </em>這個位置值後，該電子的狀態即從| <em>1s</em> 〉<strong>塌縮</strong>（collapses）或稱跳到 | <em>x₁ </em>〉的狀態。也就是說，我們的測量改變了電子的狀態，該電子不再是以 <em>1s </em>軌域 “繞行” 原子核，而變成待在 <em>x₁</em> 位置的電子並記做 | <em>x₁ </em>〉。值得注意的是，這個 | <em>x₁ </em>〉狀態是位置測量的一個<strong>本徵態</strong>，而 <em>x₁</em> 是這個 | <em>x₁ </em>〉狀態對應位置測量的<strong>本徵值</strong>（eigenvalue）。</p><p>在這裡我們學到量子力學中關於測量的一個重要的論述：</p><p>當我們對一個狀態|<em>ψ</em> 〉進行 <strong>A</strong> 的測量，會迫使系統從 | <em>ψ</em> 〉塌縮成 <strong>A</strong> 的其中一個本徵態 | <em>a1</em> 〉，至於會塌縮哪個本徵態是機率性的，出現 | <em>a1</em> 〉的機率為 |〈 <em>a1</em>|<em>ψ</em> 〉|²，出現另一個 |<em>a2</em> 〉本徵態的機率是|〈 <em>a2</em>|<em>ψ</em> 〉|²，出現第 <em>n </em>個本徵態 | <em>an </em>〉的機率是 |〈 <em>an</em>|<em>ψ</em> 〉|² 。但是當我們真的進行 <strong>A</strong> 的測量，這些所有可能的結果中只會有一個出現，就像樂透頭獎人人有機會，但是最後中獎的就是那麼一位！假如出現的本徵態是 |<em>a1</em> 〉，同時就代表我們量到的測量結果是本徵值是 <em>a1</em>。</p><p>套用回剛剛的例子，|<em>ψ</em> 〉就是剛剛的|<em>1s</em> 〉，<strong>A</strong> 的測量就是位置 <em>x</em> 的測量，如果我們量到位置本徵值 <em>x₁</em>，就代表在這測量的當下，系統由|<em>1s</em> 〉塌縮成 | <em>x₁ </em>〉。</p><p><strong>第三</strong>，如果我們在剛剛量完位置得到 <em>x₁</em> 後立馬再量一次位置，我們還是會得到 <em>x₁</em>，或說有 100% 的機率會得到 <em>x₁</em> ，這是因為剛剛量完位置後，電子的狀態變成了 | <em>x₁ </em>〉，而 | <em>x₁ </em>〉是位置測量的本徵態。這就好像我們一開始說如果你對| <em>1s</em> 〉狀態的電子做力學總能測量，你有 100% 的機會會得到 E₁s = −13.6eV。另外一個觀點是，出現非 <em>x₁ </em>的位置 <em>x₂</em> 的機率是 |〈 <em>x₂</em>|<em>x₁</em> 〉|²=0。</p><p><strong>第四</strong>，因為電子現在處於 | <em>x₁ </em>〉，是位置測量的本徵態，而非力學能的本徵態。如果你在此時進行力學能測量，不見得會得到 −13.6 eV。相信讀到這裡，你已經能猜到，測量到能量為 E₁s = −13.6eV 的機率是|〈 <em>1s</em>|<em>x₁</em> 〉|²。</p><p>到這裡讓我做個小結。我們已經學到在量子力學中，測量結果呈現機率性，而且在你量之前你沒辦法預測哪個結果會出現，只能預測各個可能的結果出現的機率。這個荒謬的特性，可說是量子力學中最讓人難以接受的。在量子力學出現之前，我們總是相信我們想要測量的物理量，是一種系統的本質或稱物理真實（Physical Reality），不論我們去測量他與否，都應該要在那裡，而不是像在量子力學裡似乎測量的結果是因為測量的動作所致。這個荒謬之處，可以用一句名言來凸顯，就是 “<strong><em>如果你不看月亮，月亮還在天上嗎？</em></strong>”。事實上，連愛因斯坦都為這種機率性的量子力學感到不滿，並與 Podolsky 和 Rosen 共同發表了一篇俗稱 EPR paper 的經典論文：<a href="https://journals.aps.org/pr/abstract/10.1103/PhysRev.47.777"><em>Can Quantum-Mechanical Description of Physical Reality Be Considered Complete?</em></a><em> </em>，標題翻作中文就是 — 量子力學對“物理真實”的描述是完備的嗎？強烈建議有興趣的人讀一下這篇 paper，看一下大師如愛因斯坦如何吵架（X）提出質疑（O），最有趣的是，波耳（Niels Bohr）在五個月後用<a href="https://journals.aps.org/pr/abstract/10.1103/PhysRev.48.696"><strong>一篇同樣標題 Can Quantum-Mechanical …… </strong>的論文</a>來回覆愛因斯坦等人！</p><h3>波動力學 vs. 矩陣力學</h3><p>你可能在高中物理課本上依稀讀過量子力學有兩種不同的表示方式（ Formulations），一個是 1926 年薛丁格（Erwin Schrödinger）提出的<strong>波動力學</strong>（Wave Mechanics），另一個則是 1925 年海森堡（Werner Heisenberg）提出的<strong>矩陣力學</strong>（Matrix Mechanics），第一次看到的你可能會想，波動力學大概就是有個波動方程式用來描述物質波怎麼隨時間演化，但是拿矩陣來解力學是什麼東西？！</p><p>事實上，這兩種表示方式是等價的，你學過的物質波，或稱波函數 <em>ψ</em>(<em>x</em>) 可以被想成是一種抽象空間中的向量 <em>ψ</em>，其中 <em>ψ</em>(<em>x</em>) 的值可以被視為向量 <em>ψ </em>的 “第 <em>x </em>個” 分量。覺得太抽象嗎？讓我們慢一點，首先，三維的向量 <strong><em>v</em></strong> 可以寫成 <strong><em>v</em></strong> = (v_x, v_y, v_z)，更進一步，我們可以把函數 <em>f</em>(<em>x</em>) 想成是向量 <em>f</em> = (<em>f</em>(0), <em>f</em>(1), <em>f</em>(2)…. )，如果再切細一點就變成 <em>f</em> = (<em>f</em>(0.0), <em>f</em>(0.1), <em>f</em>(0.2)…. )，切成無限細，我們就說 <em>f</em>(<em>x</em>) 是 <em>f</em> 向量的 “第 <em>x </em>個” 分量，當然 <em>x</em> 不一定是整數，所以說第 <em>x </em>個有點奇怪，但這就是物理學家或數學家怎麼把函數想成是向量的方式。 而且你可能已經注意到，在函數 <em>f</em>(<em>x</em>) 中的變數 <em>x</em> 可以是從 -∞ 到+∞ 的任意實數，因此，這個對應的 <em>f </em>向量有無限個分量，是一個無限維度的向量。更進一步，我們可以把作用在函數上的微分寫成矩陣（也是無限乘無限維的矩陣），從此，微分就變成了矩陣乘法；解微分方程式就變成了解線性代數（矩陣）方程式。透過這樣，我們就能把薛丁格用來描述波函數的波動方程式改寫成對應的無限維度矩陣方程式，後者就是在海森堡的矩陣力學中會出現的東西！而波函數 <em>ψ</em>(<em>x</em>) 成了狀態向量| <em>ψ</em>(<em>x</em>)<em> </em>〉，其實就是本文從一開始就一直在用的符號，只是一直沒有解釋而已。</p><h3>後記</h3><p>本文主為中文版的 <a href="https://blog.gwlab.page/what-is-quantum-mechanics-3811309f3ee7">What is Quantum Mechanics?</a> ，並補上更多的描述，希望能讓高中生也能閱讀。</p><p>另外，如果你讀完後好像解決了一些疑惑，但是又出現了更多的疑惑，那我的目的大概就成功了 😂 🥳 😅</p><p>最後，如果想更認識量子力學，很推薦觀看由臺大科學教育發展中心CASE所錄製的：<a href="https://www.youtube.com/watch?v=eHC_6WcgagU&amp;list=PLil-R4o6jmGjuQAX05pw-CI1rgI0kBPno">探索講座第9期 — 沒人懂的量子力學</a>，一系列演講。特別是當中第一集，高涌泉老師精湛的解說了歷史上科學家如何從光到底是粒子還是波的辯論，到最後光量子的出現，成為量子力學的開端。這都是本文所沒有提及但重要的思辨過程！</p><h3>延伸閱讀</h3><ul><li><a href="https://blog.gwlab.page/%E4%BB%80%E9%BA%BC%E6%98%AF%E6%BA%AB%E5%BA%A6-460f69e36fc1">什麼是溫度？</a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1f424b8cddf" width="1" height="1" alt=""><hr><p><a href="https://blog.gwlab.page/%E4%BB%80%E9%BA%BC%E6%98%AF%E9%87%8F%E5%AD%90%E5%8A%9B%E5%AD%B8-%E7%82%BA%E4%BB%80%E9%BA%BC%E5%8F%AB%E5%81%9A-%E9%87%8F%E5%AD%90-%E5%8A%9B%E5%AD%B8-1f424b8cddf">什麼是量子力學？ 為什麼叫作“量子”力學？</a> was originally published in <a href="https://blog.gwlab.page/">GWLab</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What is Quantum Mechanics? Why it is called Quantum? — An Introduction to Quantum Mechanics]]></title>
            <link>https://blog.gwlab.page/what-is-quantum-mechanics-3811309f3ee7?source=rss-ac44afc6d355------2</link>
            <guid isPermaLink="false">https://medium.com/p/3811309f3ee7</guid>
            <category><![CDATA[quantum]]></category>
            <category><![CDATA[mechanics]]></category>
            <category><![CDATA[physics]]></category>
            <category><![CDATA[schrodinger]]></category>
            <category><![CDATA[quantum-mechanics]]></category>
            <dc:creator><![CDATA[Cory Chu]]></dc:creator>
            <pubDate>Thu, 09 Dec 2021 16:29:42 GMT</pubDate>
            <atom:updated>2022-01-01T21:33:48.540Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*pVIvUDrN3iHmBk_XgHmHKw.jpeg" /></figure><p>Note: a <a href="https://corychu.medium.com/%E4%BB%80%E9%BA%BC%E6%98%AF%E9%87%8F%E5%AD%90%E5%8A%9B%E5%AD%B8-%E7%82%BA%E4%BB%80%E9%BA%BC%E5%8F%AB%E5%81%9A-%E9%87%8F%E5%AD%90-%E5%8A%9B%E5%AD%B8-1f424b8cddf">Chinese version of this article is available here</a>.</p><p>Quantum Mechanics is a theory that describes how tiny stuff like electrons move around. It’s like Newton’s Law: <em>F=ma</em>, which only works for the macroscopic objects, but is still valid in the microscopic world.</p><h3>Why is it called “Quantum” Mechanics?</h3><p>If you are learning Quantum Mechanics for the ﬁrst time, or you are asked by a friend who wants to know about Quantum Mechanics, the ﬁrst question you might have probably is why it is called “Quantum” Mechanics? A quick answer to this question is that<strong> Quantum Mechanics is a theory that can explain various quantized phenomena</strong> we encountered in experiments, which cannot be explained by Classical Mechanics. For example, the allowed energy level of the electron in the Hydrogen atom is quantized, i.e., we have <em>1s</em>, <em>2s</em>, <em>2p</em>, …… all those discrete allowed orbitals with quantized or discrete allowed energies.</p><p>Although “Quantum” is one of the possible outcomes of Quantum Mechanics, quantization is not built inside the Quantum Mechanics. <strong>The “quantized” property comes from the fact that you are asking the eigenvalue problem of a bounded system, or you can say it comes from the boundary condition.</strong> For example, the allowed total energy of the electron in a Hydrogen atom is quantized is due to the fact that the electron is bounded by the Coulomb potential from the proton. Also, the quantization of a component of Spin (e.g., the eigenvalue of <strong><em>S_z</em></strong>) is due to the periodic boundary condition of spacial rotation (i.e., exp{<em>i2π</em>}=exp{<em>i0</em>} or the world should look similar after a 360° rotation). For free particles in Quantum Mechanics, there is nothing really quantized. The momentum and the energy of a free particle are continuous.</p><p>Being quantized as a result of the boundary condition is not that mysterious and is not exclusive to Quantum Mechanics. In classical physics, we also have a very similar situation. For example, the vibration modes of a classical string with two fixed ends are quantized (e.g., the string in a violin). It makes the sound waves which are the superposition of its base frequency and all harmonics that are frequencies equal to the integer multiples of the base frequency.</p><p>Of course, here I’m talking about the original meaning of the word “Quantum.” Nowadays, people just use the word “quantization” as a technical term referring to “Quantum-Mechanicalization.” For instance, the jargon <strong>Canonical Quantization</strong> in Quantum Mechanics means that we take the commutator [<strong><em>x</em></strong>, <strong><em>p</em></strong>] ≡ <strong><em>xp</em></strong><em>-</em><strong><em>px </em></strong>= <em>i</em>ℏ, which was equal to zero in the classical case. In other words, it’s a process that makes the order of performing position and momentum operators matter, thereby making the system “Quantum-Mechanicalized.”</p><p>By the way, you might guess that the existence of photon, which is the quantized electromagnetic waves, is also a result of Quantum Mechanics. However, problems involving photons are not usually addressed in the standard Quantum Mechanics. In fact, to fully understand what photon is, we need a theory called Quantum Field Theory (QFT), which is the quantum version of classical field theory. (Especially, the Quantum Electrodynamics, the QED, which is the quantum version of Electrodynamics.) In QED, photons are treated as the quantized electromagnetic waves, while electrons are treated as the quantized Dirac field of electron. <strong>This also explains why electrons are identical particles because all electrons are the excitation of a single quantized Dirac field.</strong></p><h3>Probabilistic</h3><p>On the other hand, there is another well-known feature about things that happen in the quantum world. That is, everything is probabilistic.</p><p>In Classical Mechanics, considering a 1-dimensional case, we can describe the state of a particle by its position <em>x</em> and its velocity<em> v</em> as (<em>x</em>, <em>v</em>). For a free particle with mass <em>m</em> moving with a constant velocity <em>v</em>, its initial state is (<em>x₀</em>, <em>v</em>), and it is evolved with time <em>t</em> as (<em>x₀ + vt</em>, <em>v</em>). If you try to measure the position, momentum, and kinetic energy of such particle simultaneously at time <em>t₁</em>, you will deﬁnitely get <em>x₀ + v t₁</em>, <em>mv</em>, and <em>mv²/2</em> with a hundred percent conﬁdence with no doubt. Also, the state of the particle remains (<em>x₀ + vt</em>,<em> v</em>) after the measurement. The particle keeps moving along its original path. The measurement does not affect the state of the particle. However, this is not the case in Quantum Mechanics.</p><p>For the electron in a Hydrogen atom that is orbiting around the proton in the orbital <em>1s</em>, we can say the state of the electron is | <em>1s</em> 〉. The state | <em>1s </em>〉 completely describes how the electron is moving around the space just like the state (<em>x₀ + vt</em>,<em> v</em>) we encountered for the classical particle (Although neither of them described the rotational property of the electron or the classical particle we mentioned). If you try to measure the total energy of the electron, that is ﬁne. You will always get E₁s = −13.6eV. Also, after the measurement, the electron stays in the state | <em>1s</em> 〉. It seems nothing looks surprising, but that is because the state | <em>1s</em> 〉 is an <strong>eigenstate</strong> of the total energy measurement or the Hamiltonian operator.</p><p>Nevertheless, if you try to measure the position of this electron, something mysterious happens. First, although you will get a position of the electron, say <em>x₁</em>, you can’t predict which position you will get before you actually measure it. The only thing you are able to predict is the probability of ﬁnding this electron at position <em>x₁</em>, which can be written as P₁s(<em>x₁</em>)=|〈 <em>x₁</em>|<em>1s</em> 〉|². The notation 〈 <em>x₁</em>|<em>1s</em> 〉 means the inner product of two state vectors | <em>x₁ </em>〉and| <em>1s</em> 〉and if you are interested about what is that mean exactly, you’ll need to read a Quantum Mechanics textbook. The fact that you can only predict the probability of ﬁnding the <em>1s</em> electron in a certain position <em>x₁</em> is equivalent to that there is no “path” or “trajectory” of a quantum particle. This is also why you always see the ﬁgure of <em>1s</em> orbital on the textbook as some “electron cloud” around the proton instead of a clear trajectory following an elliptic orbit like how the Earth is orbiting around the Sun. Second, after the measurement, the state of this electron <strong>collapses</strong> or jumps into the state | <em>x₁ </em>〉. So, it’s no longer orbiting around the proton! Third, if you try to measure the position of the electron again right after the previous measurement, you will still get <em>x₁</em> with a hundred percent probability. This is because the electron is now in the state | <em>x₁ </em>〉, which is an eigenstate of the position measurement. Fourth, since this electron is no longer in the state | 1s 〉 but the state | <em>x₁ </em>〉, if you try to measure its total energy again at this point, you won’t (always) get −13.6 eV. That is because the state | <em>x₁ </em>〉is not a total energy eigenstate. In fact, the probability you get the value −13.6eV is |〈 <em>1s</em>|<em>x₁</em> 〉|²</p><p>The fact that the results of measurements are probabilistic and the exact answer of a given measurement is only known after you really did the measurement is one of the biggest mysteries of Quantum Mechanics. Because prior to Quantum Mechanics, we always think the quantities we are going to measure are the physical reality of the system. They should be already there no matter you are going to measure it or not. This probabilistic nature of Quantum Mechanics also bothers Einstein a lot, such that he actually wrote the famous paper — <a href="https://journals.aps.org/pr/abstract/10.1103/PhysRev.47.777"><em>Can Quantum-Mechanical Description of Physical Reality Be Considered Complete?</em> — with Podolsky and Rosen (a.k.a. the EPR paper)</a>. This also leads to the famous quote, “Is the moon there when nobody looks?”</p><h3>Wave Mechanics vs. Matrix Mechanics</h3><p>You probably heard that there are two different formulations of Quantum Mechanics. One is the <strong>Wave Mechanics</strong>, which was mainly developed by Erwin Schrödinger in 1926. The other is the <strong>Matrix Mechanics</strong>, which was mainly developed by Werner Heisenberg in 1925. It’s hard to believe at the ﬁrst time that how the “matrix” can be used for doing Mechanics. However, they are actually equivalent. The wavefunction <em>ψ</em>(<em>x</em>), which is a function, can be represented by a vector. This can be done by identifying the value of <em>ψ</em>(<em>x</em>) as the “<em>x</em>-th” “component” of the corresponding vector. Since the variable <em>x</em> in <em>ψ</em>(<em>x</em>) is continuous and running from -∞ to +∞, there are infinite numbers of components in such vector. So, the vector we are dealing with here is essentially infinite-dimensional. Furthermore, the differential operators we encounter in Schrodinger’s wave equation can be represented by Matrices, which is also infinite-dimensional. So, differentiation becomes the matrix product, and solving linear differential equations in the Wave Mechanics becomes solving linear algebra problems in the Matrix Mechanics!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3811309f3ee7" width="1" height="1" alt=""><hr><p><a href="https://blog.gwlab.page/what-is-quantum-mechanics-3811309f3ee7">What is Quantum Mechanics? Why it is called Quantum? — An Introduction to Quantum Mechanics</a> was originally published in <a href="https://blog.gwlab.page/">GWLab</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Microcinema — Chicago]]></title>
            <link>https://corychu.medium.com/microcinema-chicago-5a1fc7b55312?source=rss-ac44afc6d355------2</link>
            <guid isPermaLink="false">https://medium.com/p/5a1fc7b55312</guid>
            <category><![CDATA[davinci-resolve]]></category>
            <category><![CDATA[film]]></category>
            <category><![CDATA[chicago]]></category>
            <category><![CDATA[microcinema]]></category>
            <category><![CDATA[filmmaking]]></category>
            <dc:creator><![CDATA[Cory Chu]]></dc:creator>
            <pubDate>Fri, 03 Dec 2021 08:49:14 GMT</pubDate>
            <atom:updated>2021-12-03T08:49:14.317Z</atom:updated>
            <content:encoded><![CDATA[<h3>Microcinema — Chicago</h3><p>It’s my recent tiny filmmaking work:</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FXUue2APw-Bo%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DXUue2APw-Bo&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FXUue2APw-Bo%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/c125fab4b12e268d603da9fb2d6babf9/href">https://medium.com/media/c125fab4b12e268d603da9fb2d6babf9/href</a></iframe><h4>Equipment</h4><ul><li><strong>Camera</strong>: Nikon D7500</li><li><strong>Lens</strong>: Nikon AF-S DX NIKKOR 16–80mm f/2.8–4E ED VR</li><li><strong>Recorder</strong>: Blackmagic Design Video Assist 7&quot; 12G</li><li><strong>Microphone</strong>: Line Audio CM4</li></ul><h4>Post-production</h4><ul><li>Final Cut Pro</li></ul><h4>Future Task</h4><ul><li>Learn how to use Davinci Resolve for color grading by following the <a href="https://www.blackmagicdesign.com/products/davinciresolve/training">Official Training Materials</a>, especially <a href="https://documents.blackmagicdesign.com/UserManuals/DaVinci-Resolve-17-Colorist-Guide.pdf">The Colorist Guide to DaVinci Resolve 17</a></li><li>Add more sound-design</li><li>Establish the post-production workflow between Final Cut Pro, Davinci Resolve, and Logic Pro</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5a1fc7b55312" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[錄音筆記－遠端錄音]]></title>
            <link>https://corychu.medium.com/%E9%8C%84%E9%9F%B3%E7%AD%86%E8%A8%98-%E9%81%A0%E7%AB%AF%E9%8C%84%E9%9F%B3-c676fa23034a?source=rss-ac44afc6d355------2</link>
            <guid isPermaLink="false">https://medium.com/p/c676fa23034a</guid>
            <category><![CDATA[remote]]></category>
            <category><![CDATA[錄音]]></category>
            <category><![CDATA[recording]]></category>
            <category><![CDATA[logic-pro-x]]></category>
            <category><![CDATA[遠端工作]]></category>
            <dc:creator><![CDATA[Cory Chu]]></dc:creator>
            <pubDate>Mon, 22 Nov 2021 22:39:10 GMT</pubDate>
            <atom:updated>2021-11-29T05:03:45.443Z</atom:updated>
            <content:encoded><![CDATA[<p>今天要來分享如何透過網路遠端錄音的工具。在後疫情時代，遠端工作、合作已經成為大家生活無法分割的一部份，而遠端錄音除了解決因為疫情無法讓大家群聚在錄音室的問題，也開啟了更多跨地區甚至跨國合作的可能。不論是錄 Podcast 還是流行樂，都能或多或少受惠。</p><p>然而錄音要如何遠端進行呢？想必大家關在家裡開了那麼多線上會議，一定聽膩了 Zoom、Skype 那種可怕的音質。所以會想，透過網路傳輸高品質的聲音是不是天方夜譚？其實，這些通訊軟體的音質會那麼有限，是因為為了要有好的多方通話體驗，追求低 Latency 的結果。</p><p>我們不也是常聽線上串流的 Spotify、Apple Music 嗎，他們的聲音相較之下就好很多，所以其實只要容許多一點點的 Latency，透過網路傳輸高品質的聲音是絕對可行的！</p><p>應用到遠端錄音上，最好的狀態當然是雙方都在本機錄音再將檔案傳給對方，可是在合作上，有時候又不是那麼即時方便。另一方面，如果能將 DAW 中的分軌 Playback 給對方，同時遠端錄下對方的聲音，就算有個 1000ms 的 Latency 其實也無傷大雅，只要在 DAW 裡將錄音往回拉個一秒就行了。</p><p>就在這個背景下，有許多高品質、稍延遲的線上傳音軟體相繼誕生，並能實現遠端錄音的可能。在這列出一些我有聽過的軟體，並以其中一個免費的軟體配上 Logic Pro 來分享給大家，其他類似的軟體配在各種 DAW 裡面的方式也都大同小異。</p><ul><li><a href="https://www.source-elements.com/products/source-connect/">Source-Connect</a><br>據說最穩、但是最貴 QQ</li><li><a href="https://audiomovers.com/wp/">Audiomovers Listento</a><br>只傳聲音、沒有其他功能，但是很穩，介面非常簡單，音質與 Latency 也可以調整，另外除了可以用 App 接收外，也可以直接在網頁打開就能聽。訂閱方案中提供 3.99USD/week 的方案我覺得大推！需注意接收端不需付費，但傳送端需要訂閱。<br>更：漲價了ＱＱ，而且weekly方案消失了嗚嗚</li><li><a href="https://www.sessionwire.com/">Sessionwire</a><br>據說目前沒那麼穩，前陣子有大改版，持續進步中，同時提供視訊與Screen Sharing 等功能，算是方便的 All-in-one，價錢 15USD/mo 尚可接受。</li><li><a href="https://sonobus.net/">SonoBus</a><br><strong>免費、</strong>open-source<strong>！！！</strong>而且我覺得用起來蠻喜歡的，就是介面上不像 Listento 那麼直覺，偏複雜，音質與 Latency 都可以自訂，上手後我覺得很棒。唯二缺點大概就是傳輸沒有加密以及設定上麻煩點。</li></ul><h3>SonoBus + Logic Pro</h3><p>免費的 SonoBus 提供 Standalone app 與 plugin（AU、VST、AAX）版本，並支援 macOS、Windows、Linux，可以於官網下載：<a href="https://sonobus.net/">https://sonobus.net/</a></p><p>以下將以用 Standalone App 當作 Sender，並在網路的另一端使用 Logic Pro + AU plugin 作為 Recever 來說明。</p><h4>發送方 Sender（以 Standalone App 為例）</h4><p>在安裝完 SonoBus 後即可打開他的 Standalone App：</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*grmznuskD6bTzhwnO4bKTA@2x.png" /></figure><p>一開始可以點選上方的 Connect ，就會進到連線頁面：</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*qCWASFTz7FfiWaSN_r-7HQ@2x.png" /></figure><p>如果你是第一個人需要先決定 Group Name 與選擇性的 Password，建議選擇 Private Group ，也就是別人要知道你的 Group Name 才能連進你的 Group，另外底下的預設 Server 只是用來讓大家能透過 Group Name 找到對方，聲音的傳輸不經過該伺服器，而是點對點傳輸。</p><p>Connect 完成後就可以回到主畫面，在左上角的齒輪中可以選擇 input device，在這裡可以選擇你要傳送到遠端的錄音介面。</p><p>接著就能等待接收方登入你的 Group 了！</p><h4>接收方 Receiver（以 Logic Pro + SonoBus AU-Plugin 為例）</h4><p>要在 Logic Pro 裡面接收聲音比較複雜一點，你可能會想，只接在你想要錄進去的 Channel 將 Input 設成 No Input，並掛上 SonoBus Plugin 就能錄下遠端的聲音了吧！不不不，你會發現根本沒聲音，據說這是因為如果軌道沒有 Input ，Logic Pro 會自動忽略該軌道上的 Plugin。那該怎麼辦呢？基本上只要把該軌的 Input 掛著如麥克風之類的 Input，並在 SonoBus 的 Plugin 裡將 Input Mute 掉就可以了。</p><p>另外一個推薦的做法是開個 Aux Channel，在該 Aux 上只掛上 SonoBus 的 Plugin 用來接收遠端聲音，並透過 Bus 將該Aux 的 Input 送回給 Recording 用的 Channel，在這個方法中 Aux Channel 的 Input 一樣不能是 No Input，可以隨意掛個麥克風，再於 SonoBus Plugin 中 Mute 掉 Input 即可，可以參考 <a href="https://audiomovers.com/storage/pdfs/Listento_Quick_Guide_Dec_2020.pdf">Listento 的 Guide</a> 有詳細的說明（對 😂 是 Listento 的Guide 不是 SonoBus 的）</p><p>實際上於 Logic Pro 的操作如下：</p><ol><li>按 [X] 開啟 Mixer</li><li>按 [Control]+[N] 新增 Aux Channel</li><li>選擇有聲音的 Input</li><li>掛上 SonoBus AU Plugin</li><li>設定 Output 到某個 Bus</li><li>將該 Bus 的 Output 設成 No Output 或直接 Mute 掉</li><li>將要用來錄音的軌道之 Input 選為剛剛的 Bus</li><li>開啟 SonoBus Plugin 將 Input Mute 掉</li><li>在 SonoBus Plugin 中透過前面設定的 Group Name 與 Sender Connect 到同一個 Group 裡面</li></ol><p>這樣應該就能正常錄音了！也可以參考下面兩個螢幕截圖：</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/964/1*0rdtVBNsvGpEJXI2NJel0A@2x.gif" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/804/1*IuZF_JU1_vtVgvoH0_hLiw@2x.gif" /></figure><p>另外可以發現 Receiver 可以設定 Jitter Buffer，並且 Sender 與 Receiver 都可以控制 Sending Quality。詳細的各種設定就有待你去試試看拉～～</p><p>最後，因為小弟其實很菜，只是剛好 Google 到 SonoBus 這個好東西所以分享給大家，如果文章有什麼可以改進的，或有各式想法，歡迎留言讓我知道喔！！感謝各位大大不吝惜指教 😬</p><h3>延伸閱讀</h3><ul><li><a href="https://corychu.medium.com/%E9%8C%84%E9%9F%B3%E7%AD%86%E8%A8%98-%E6%95%B8%E4%BD%8D%E9%9F%B3%E9%87%8F%E6%A8%99%E6%BA%96-dbfs-dbtp-lufs-c47ca4646b7f">錄音筆記－數位音量標準、 dBFS、dBTP、LUFS</a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c676fa23034a" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[[分享]在日本重力波實驗KAGRA的生活經驗]]></title>
            <link>https://blog.gwlab.page/%E5%9C%A8%E6%97%A5%E6%9C%AC%E9%87%8D%E5%8A%9B%E6%B3%A2%E5%AF%A6%E9%A9%97kagra%E7%9A%84%E7%94%9F%E6%B4%BB-6e732c339037?source=rss-ac44afc6d355------2</link>
            <guid isPermaLink="false">https://medium.com/p/6e732c339037</guid>
            <category><![CDATA[physics]]></category>
            <category><![CDATA[gravitational-waves]]></category>
            <category><![CDATA[experiment]]></category>
            <category><![CDATA[kagra]]></category>
            <dc:creator><![CDATA[Cory Chu]]></dc:creator>
            <pubDate>Sun, 24 Oct 2021 01:42:03 GMT</pubDate>
            <atom:updated>2021-10-25T13:01:16.013Z</atom:updated>
            <content:encoded><![CDATA[<p>今天要來和大家分享在 KAGRA Site 工作的經驗，雖然聽起來不像在南極的實驗那樣酷炫，但還是蠻新奇的體驗。</p><p>KAGRA 的全名是神岡重力波探測器（KAmioka GRAvitational wave detector），他位在日本岐阜縣飛驒市神岡町的深山裡（Kamiokacho, Hida, Gifu, Japan），是一座邊長 3km 的 L 型大型雷射干涉儀。整個 KAGRA 的主體都在山洞裡面，L 的轉角處是所謂的 Center Area 或稱 Corner Station，有著當中最大型的洞穴，也是干涉儀分光鏡（Beamsplitter）與雷射源頭之所在。兩臂延伸出去被稱作 X-Arm 與 Y-Arm ，並連接到兩端的 End-Station，即 X-End 與 Y-End，可以參考下圖的電腦模型：</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*cp52AOCkUf3LaQ6hcSY95Q.gif" /><figcaption>Credit: ©KAGRA Collaboration / Rey.Hori (請勿轉載，如有使用需求請以英文連絡：<a href="/service/mailto:kagra-pub@icrr.u-tokyo.ac.jp">kagra-pub@icrr.u-tokyo.ac.jp</a>)</figcaption></figure><p>大家也許對這個地名不熟悉，但你可能聽過<a href="https://corychu.medium.com/%E7%99%BD%E5%B7%9D%E9%84%89%E5%90%88%E6%8E%8C%E6%9D%91-8f5bf8afe177">白川鄉（Shirakawa-go）的合掌造聚落</a>與<a href="https://www.snowmonkeyresorts.com/zh-hant/activities/tateyama-kurobe-alpine-route/the-snow-walls-of-tateyama-kurobe/">立山黑部（Tateyama Kurobe）的雪牆</a>，而 KAGRA 就位於這兩地的中間，到兩地分別約莫 60km 或 1.5 小時的車程。從那高聳的雪牆就不難想見地理上位於同樣區域的富山市到 KAGRA 當地，一路上的冬天降雪量可說是相當驚人。這是我當時下飛機後在路上的崩潰照，行李箱根本拉不動嗚嗚，最後只能放平像雪橇那樣拖：</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*7K0KeNTwintHZwK48HPHUg.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*htKzSCYy7hvnYxSwN-uTgg.jpeg" /><figcaption>Toyama 路上，攝於 2018 Feb. 12</figcaption></figure><p>當然也有 KAGRA 控制室外與隧道口美美的雪景拉：</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Pqxd0Pr1P56mbuEeBow_Eg.jpeg" /><figcaption>KAGRA 控制室外的雪景，控制室是一棟獨立的建築，距離 KAGRA 主要隧道口（Atotsu Entrance 跡津坑口）約 5km 車程，攝於 2017 Dec. 8</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uR2dVKRh-EGFemL19NGg8g.jpeg" /><figcaption>KAGRA 的主要隧道口（Atotsu Entrance 跡津坑口），攝於 2017 Dec. 12</figcaption></figure><p>不過如果你車子停在外面，到了晚上上面就是厚厚的積雪，雖然很美，但是也是蠻崩潰的：</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*KukDM7rL0O4H5h2-HaFzzw.jpeg" /><figcaption>車上滿滿的積雪，開車前鏟到懷疑人生，攝於 2017 Dec. 13</figcaption></figure><p>說個題外話，也許你以前在課本上學過日本四大公害病之一的痛痛病，也就是 1950 年發生在日本富山縣那世界最早的鎘（Cadmium）中毒事件。這事件發生的富山就是離 KAGRA 最近的城市，也是坐飛機或新幹線前往 KAGRA 必經之地，而當時被污然的河川就是 KAGRA 外面那條神通川（Jinzugawa），當年的污染源則是在當地採礦的神岡鉱業（好像以前叫三井金属，不太確定）。而神岡鉱業也就是協助東京大學在該地開鑿 KAGRA 與 <a href="http://www-sk.icrr.u-tokyo.ac.jp/sk/index-e.html">Super-Kamiokande</a> 所需之隧道的公司，畢竟挖洞是採礦公司的專長。另外如果你是做高能物理的人，應該對 Super-Kamiokande 這個微中子探測器而熟能詳吧，沒錯，他就跟 KAGRA 位在同一座山內，如果 KAGRA 是連接本壘到一三壘的 L 型，Super-Kamiokande 就大概在投手丘的位置。</p><p>扯太遠了哈哈哈，回歸主題</p><p>上面有提到 KAGRA 有兩個 3km 長的 Arm ，本部落格的封面照就是在 Arm 的隧道內拍攝的，就是這張：</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*In6W8FEbk_Y9jDuKhdl3WQ.jpeg" /></figure><p>像我那時參與的實驗部分有很多工作必須要在 End-station 完成，如果算上早上去一次，下午去一次，這來回兩趟共 12km 的路程就成了每天必經的道路。你可能會好奇，到底在這裡面穿梭是什麼感覺呢？可以看看這個坐在小電動卡車 <a href="https://www.mitsuoka-motor.com/global/lineup/special/like-t3/">Mitsuoka, Like-T3</a> 上拍攝的影片，就知道這是段多麽無聊的路程（中間都差不多，片尾有回到 Center Area 的樣子）：</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FZwGgvje9R64%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DZwGgvje9R64&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FZwGgvje9R64%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/0829b874c17756f06dd41c61de974331/href">https://medium.com/media/0829b874c17756f06dd41c61de974331/href</a></iframe><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*5jhxhVjQAhoxDlNYCVoYdw.jpeg" /><figcaption>小電動卡車 Mitsuoka, Like-T3 的樣子</figcaption></figure><p>不過通常如果不是要載器材的話，大家會騎電動輔助腳踏車，沒錯，電動只是輔助，還是要踩 😂😭，甚至有時候東西很輕的話也是會腳踏車載。讓各位看一下腳踏車的樣子（我知道你在想什麼，這部比較短了，只有錄一分鐘🤣）：</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2F2E5q8CaCsAI%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3D2E5q8CaCsAI&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2F2E5q8CaCsAI%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/54d5145009af6ac797b491a95a819767/href">https://medium.com/media/54d5145009af6ac797b491a95a819767/href</a></iframe><p>看到這，也許你會想，這隧道也太陰暗狹小了吧，又濕又冷又暗真的可以嗎？事實上大部分的實驗工作並不會在 Arm 的中間，而是在 Beamsplitter 與 LASER 附近的 Center Area 或兩端的 End-Station 中。這些地方有著較寬闊的空間並有恆溫恆濕的控制，也有著不同等級的無塵空間，來確保光學元件的乾淨，相較之下長長的 Arm 中間只有如影片中左側那條真空的大管子，是雷射光往返兩個 End-Station 的路徑。</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*rHJpo2Ni45fM2OxKHhaA5g.jpeg" /><figcaption>Center Area 一隅，攝於 2019 Mar. 15</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*UCvg0H0h_mQJWvpZZCQTYw.jpeg" /><figcaption>整個 KAGRA 中大多數的實驗裝置可以透過電腦上叫做 <a href="https://epics.anl.gov/extensions/medm/index.php">MEDM</a> 的 GUI 經由網路來控制</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*E6I6jGbFCD0_YXjUVS2zYw.jpeg" /><figcaption>在無塵空間中的光學元件們</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*FlhEggu_bFRf5xlX9M-Aaw.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Z2yMTGmCNPjGddlT0eZxdg.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*aM2lAFGvXa60pu8ji8ipKA.jpeg" /><figcaption>在無塵區域工作必須穿著無塵衣、口罩、手套，並將攜入的器材用高壓空氣與酒精紙巾擦拭乾淨，若雷射在運作中，也必須戴上防雷射護目鏡以策安全</figcaption></figure><p>另外一日的行程大致就是七點左右起床，從住的地方開車前往 KAGRA Site，車子可能是自己租的，或是學校租的，我那時候沒有車，都是別人載，所以完全不能賴床😭。路上可能會停個全家或小七買杯咖啡、早餐、甚至午餐便當，因為 KAGRA Site 在的區域什麼店都沒有，吃的都要自己帶去。有些人會前一晚到超市買比較便宜的食物，像是冷凍的炒飯之類的，或自己準備沙拉。我則是很懶，通常就早上通勤時買便利商店的微波午餐，日本的微波食品算好吃，所以沒有想像中的糟，甚至非微波的熟食區還會有炸豬排之類的東西。唯一的遺憾大概就是沒有台灣的手搖飲，我就只能天天喝瓶裝飲料。</p><p>經過二三十分鐘的車程，就會抵達 Analysis Building，早上會在控制室先開一個簡單的行前會議，確認當天各個小組的工作內容，也確保大家的工作不會互相影響到，不然就要協調清楚，然後就從控制室開車到隧道口，把午餐先冰到隧道口旁的工作站內，就進入隧道工作，中午出來工作站吃午餐，再回去隧道裡工作到傍晚，最後開車回控制室做些不需要在隧道裡才能做的事。</p><p>而在安全上，在山洞裡面都要團體行動，至少兩人；在隧道內需配戴安全帽、頭燈、O2&amp;CO Meter、反光背心，並在進出個區域時在白板上留下名字與時間，以防意外發生且停電時，即便電腦都gg了，還是可以確定裡面有哪些人。</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*hahY_jhByxwdWyQ3SS2CZg.jpeg" /><figcaption>除了用門禁刷卡的紀錄外，也會在白板上寫下進出時間，以防警急狀況發生。</figcaption></figure><p>而在隧道裡面當然是不會有有手機信號，不過大家會帶上 PHS ，同時整個山洞裡面都有架設 PHS 的基地台，有問題就能互相聯絡。</p><p>講了這麼多隧道內的樣子，也帶大家來看一下控制室裡面的樣子：</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*KlA9SyQEJfn0hkdNc77c-g.jpeg" /><figcaption>週末沒什麼人的KAGRA 控制室，攝於 2019 Mar. 30</figcaption></figure><p>上面提過控制室其實位於隧道外約 5 公里遠的建築內，所以基本上控制室與被控制的東西都是透過網路來控制的，所以嚴格說起來，這就是比較多電腦跟螢幕的地方而已，功能上和隧道裡任一台在內網裡的電腦是一樣的。不過有的地方大家聚在一起有問題也比較容易互相討論。</p><p>而控制室所在的建築又稱 Analysis Building（解析棟），裡面也包含了辦公室、研究室、電路實驗室、機房等空間，在裡面大家都穿拖鞋，冬天地板還會發熱呢，是個非常日式的辦公空間。</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*qtNxRAVrt-q9yhEsFnWgZg.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*aRHpj09tXNgtjbzAd_DfFQ.jpeg" /><figcaption>Analysis Building 內的走廊，相當日式精緻的空間設計</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*pMpMxrso0UYT5ObEcNST-g.jpeg" /><figcaption>Analysis Building 內的 Circuit Room，可以將設備從山洞內帶來這裡做仔細的電路檢修</figcaption></figure><p>有時候也可以在 Analysis Building 內開個章魚燒 （Takoyaki）派對之類的，畢竟在山洞裡是不能吃東西的。</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*gwkngrczp_-ZiamBSkQHxg.jpeg" /><figcaption>Takoyaki Party</figcaption></figure><p>最後你一定會想，那在 KAGRA 這種連商店都沒有的山上工作，大家都住在哪裡呢？是的，有許多人會選擇住在離 KAGRA Site 約半小時車程的地方，像是我那時住的大沢野（Osawano）地區大概就位於富山市到 KAGRA Site 的中間，雖不像富山那樣是個城市，但至少有些商店、超市或餐廳，像是麥當勞或すき家，生活上比較方便。也讓大家看一下那時候住的房間，可說是相當標準的日式格局，最不習慣的大概就是沒有椅子，大家都是坐地上：</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-ACukz_H9VzB4nNSX1LzOA.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*XdAbWSLmxaJLhRcvl4nJDQ.jpeg" /><figcaption>當時東京大學租下來提供給短期訪問者的住宿，是三房一廳的格局，離 KAGRA Site 大約半小時車程</figcaption></figure><p>好了，這就是在KAGRA的生活簡記，希望接下來有機會寫一些關於 KAGRA 或是重力波研究的科普介紹，也許能包含更多喜歡物理的你可能更想了解的細節。如果有特別想知道什麼也可以留言告訴我，有時間的話也許能在未來的文章中回答你的問題～～謝謝大家😬</p><h3>其他網路上的資訊</h3><ul><li>KAGRA 內部的街景 VR（不確定是不是這樣叫XDDD）可以看看：<br><a href="http://www.icrr.u-tokyo.ac.jp/prwps/panorama/e/KAGRA/">http://www.icrr.u-tokyo.ac.jp/prwps/panorama/e/KAGRA/</a></li></ul><h4>一些KAGRA介紹的影片</h4><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FDUg0wqK9VoM%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DDUg0wqK9VoM&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FDUg0wqK9VoM%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/94b466ef8ee38769fad589cc63646699/href">https://medium.com/media/94b466ef8ee38769fad589cc63646699/href</a></iframe><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2Fx7LHXTtGei0%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3Dx7LHXTtGei0&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2Fx7LHXTtGei0%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/06b97cf7f4978c8c507f5024692707e3/href">https://medium.com/media/06b97cf7f4978c8c507f5024692707e3/href</a></iframe><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6e732c339037" width="1" height="1" alt=""><hr><p><a href="https://blog.gwlab.page/%E5%9C%A8%E6%97%A5%E6%9C%AC%E9%87%8D%E5%8A%9B%E6%B3%A2%E5%AF%A6%E9%A9%97kagra%E7%9A%84%E7%94%9F%E6%B4%BB-6e732c339037">[分享]在日本重力波實驗KAGRA的生活經驗</a> was originally published in <a href="https://blog.gwlab.page/">GWLab</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>