Trending December 2023 # Review: Master & Dynamic Mw07 Offer Great Sound In True Wireless In # Suggested January 2024 # Top 19 Popular

You are reading the article Review: Master & Dynamic Mw07 Offer Great Sound In True Wireless In updated in December 2023 on the website We hope that the information we have shared is helpful to you. If you find the content interesting and meaningful, please share it with your friends and continue to follow and support us for the latest updates. Suggested January 2024 Review: Master & Dynamic Mw07 Offer Great Sound In True Wireless In

As you’d expect of an Apple-focused tech writer, I do have one or two pieces of Apple kit. The current tally comprises my 15-inch MacBook Pro, Apple Thunderbolt Display, 11-inch MacBook Air (only really kept as an emergency backup these days), 10.5-inch iPad Pro, iPhone X, Apple Watch Series 4 (as of later today) and three HomePods.

But there’s one Apple product I dismissed as soon as I tried it: AirPods. That’s not to say I hate them. I love the idea of something that portable. I adore the pairing system. The battery-life is great. There’s a lot to love about them. But not, to my ears, the audio quality …

Don’t get me wrong: AirPods sound quality isn’t terrible; it’s just very, very average. I found the same thing with the first truly wireless in-ear headphones I tried: the Bragi Dash. I loved the portability, but the sound was again only adequate.

I’m no audiophile, but I do love music, and want to listen to it through headphones which really do justice to the sound. So much as I love the idea of just tucking a charging case in my pocket and having headphones available on demand, I’d resigned myself to the idea that good quality audio on the move means carrying on-ear headphones, with my Bowers & Wilkins P5 Wireless my usual choice.

I’d been equally impressed by the Master & Dynamic MH40 headphones I tried back in 2014. I was somewhat surprised that they equalled B&W headphones because the company didn’t have anything like the same heritage. It was founded just a year earlier, and yet the audio quality was spot-on.

So when I heard the company was launching its first wireless in-ear headphones, it caught my attention straight away. I was able to test them on the day they launched, and first impressions were good. I’ve been using them as my primary headphones since then, so can now give a more considered verdict.


The first thing that struck me about the MW07 design was not the earphones themselves, but the charging case.

It’s polished stainless steel, with three tiny LEDS to indicate the battery status of the case itself and each earphone: green for fully-charged, orange for medium charge and red for low.

The case is, quite simply, beautiful. It looks like a cufflinks case for a designer brand. It feels equally good in the hand, and magnets mean it closes with the same satisfying snap as the AirPods case. It looks and feels absolutely as you’d expect from a premium audio brand.

The earphone aesthetics are a more subjective taste. A stainless steel sound enclosure is surrounded in handcrafted acetate, designed to give the best balance of audio quality and weight. The design is a marbled one, available in a choice of four colors. Personally, I have to say that the aesthetics don’t do much for me, so I’d opt for the anonymity of the black ones, but if marbling is your thing, the blue, grey and tortoise-shell versions are more distinctive.


The MW07 come with four different tips and two sizes of ‘wing’ to help achieve the perfect fit. I seem to be blessed with standard-sized ear canals, as the default tips always seem to work for me no matter the brand.

The ‘wings’ are designed to hold the headphones securely in place. You twist the earphones in from about a 45-degree angle, and the wing slips behind the part of the ear which Google suggests is called the antitragus.

Once inserted, they proved 100% secure. They never fell out or even felt like there was the slightest chance they would.

They are perfectly comfortable for me. My longest continuous usage was a two-hour train journey, and I scarcely knew they were there.


There are two volume buttons on the left earphone, and the usual multi-function button on the right. Press once to play/pause, twice for next track, three times for previous track and hold down to activate Siri.

Their small size makes the controls slightly awkward to use – even a week in, I was still having to feel for them – but I think that’s unavoidable at this size, and it’s not like AirPods controls are great either. With anything this small, you’re going to be better off using your phone as the controller.

There’s also the same auto-pause/play feature as AirPods, pausing when you remove one from your ear and resuming when you reinsert it. However, while AirPods will happily resume if you reinsert one a minute or two later, the MW07 times out after eight seconds – which is too short when you remove one to answer a question, listen to a PA announcement or similar. It’s not like it’s a big deal to hit the play button to resume, but auto-pause and -resume is a really nice feature, so I’d prefer a significantly longer timeout.

Sound quality

I don’t think it’s physically possible for in-ear headphones to create the same level of immersive sound you get from on-ear or over-ear ones. Driver size is far from the only thing that matters with headphones, but it is one factor. Even when you place a tiny driver right inside your ear, there’s only so much air it can push.

AirPods, for example, have 9mm drivers, compared to the typical 40mm drivers of decent on-ear headphones. The MW07 have 10mm drivers, which are the largest you’ll find in-ear, and they are made of beryllium, a material used only in high-end Hi-Fi equipment. The stiffness and lightness help compensate for the size, and the cost of the material likely goes a long way toward explaining the $299 price tag.

But while no in-ear headphone can match the performance from their larger brethren, the MW07 come closer than any others I’ve yet tried. The sound is simply stunning for the size.

The audio is beautifully neutral, especially noticeable in live performances where you can really feel you are there in the room.

Both treble and mid-tones are clear and detailed, and the bass is really impressive. But while the bass has real presence, it doesn’t feel in any way artificial – the balance is spot-on.

There’s also plenty of volume. Even on public transit, I never use them at maximum volume, which I think is another impressive achievement from in-ear units. But if you do crank them up to the max, there’s no distortion.

For pure audio quality, these are hands down the best in-ear headphones I’ve ever used – and that includes B&O H5s.

For the first time with in-ear headphones, they’ve hit the sweet-spot. Previously I’ve had to significantly compromise on audio quality if I want the freedom achievable with in-ear headphones – there when you want them, unnoticeable in a pocket when you don’t. But with the MW07, I feel like the compromise in audio quality is so small that the trade-off works.

For plane travel, where I want longer battery-life and the option of wired connection to the in-flight entertainment, I’ll take my P5s. But out and about in London, I’ve now switched to these – and that really says a huge amount about their quality.

Battery life

There is one area where they fall short compared to AirPods, and that’s battery-life. While AirPods claim up to 5 hours per charge, and 24 hours total with in-case recharging, MW07 claim 3.5 hours per charge and 14 hours total. That’s a significant difference.

For some, that may be important. For me, it’s not. On flights, I’m going to take my P5s. Any other time, I’m never going to be listening for more than a couple of hours at a time, and I’m never going to get anywhere close to 14 hours in a day.

I haven’t been able to test the 3.5 hours claim simply because I’m never going to listen that long in one go, but I did hit around 2.5 hours on one journey and they were still going strong.

Price and conclusions

Having previously been impressed by Master & Dynamics headphones, I was expecting these to be good. But not as good as they turned out to be. I really didn’t expect to be switching to these as my daily drivers when mobile in London.

We all have different ears. I have true audiophile friends who spend four-figure sums on their headphones (and would probably spend five if they could). At the other end of the scale, you may not hear any difference between AirPods and higher-quality headphones.

$299 is quite a lot of money at almost twice the price of AirPods, but if – like me – your ears fall into Bang & Olufsen or Bowers & Wilkins territory, the MW07 justify every penny of the difference.

Master & Dynamic MW07 cost $299.99, and are available in a choice of two colors.

Check out 9to5Mac on YouTube for more Apple news:

FTC: We use income earning auto affiliate links. More.

You're reading Review: Master & Dynamic Mw07 Offer Great Sound In True Wireless In

Teclast Master T10 Review – A Good Tablet!

Teclast Master T10


The Teclast Master T10 is a 10.1″ tablet with midrange specifications, an MTK8176 hexacore processor, 4GB of RAM, and a high resolution 2560×1600 IPS LCD. There aren’t many tablets made anymore, let alone large 10.1″ tablets, and this is one of the few options released recently.

Is it Good?

Teclast Master T10


Processor MediaTek MTK8176 Hexacore

Display 10.1″ Sharp 2560×1600 IPS LCD


Storage 64 ROM

Operating System Android 7.0 Nougat

Cameras 8MP, 13MP Camera

Battery 8100mAh

Physical Dimensions 553g, 23.90 x 16.70 x 0.80 cm

Big thanks to Gearbest for providing this review unit.

Teclast Master T10


The tablet doesn’t stand out with a rather generic silver metal body, it looks like every other tablet out there. However, the metal body does feel fairly good in the hand but unfortunately fingerprints are somewhat visible on the tablet. Flip the tablet around and we see fairly small bezels on this tablet, not bezelless no, as that would render the tablet unholdable. We have a fair amount of ports here, a microUSB port, a microSD card slot, and a micro HDMI port as well.

Fairly generic look

Teclast Master T10


Gizchina News of the week

Join GizChina on Telegram

Teclast Master T10


The tablet has stereo speakers, one on each side of the tablet and it does provide stereo sound to a certain extent, and audio volume is quite loud, loud enough for most situations. Audio quality is just average, there is a bit of bass but there is distortion at max volume that disappears if you lower it a notch.

Teclast Master T10


I also performed a video playback test and was able to playback 720p video at 200nits of brightness for 9 hours and 22 minutes before the tablet died, definitely a good media consumption device.

Good battery life

Teclast Master T10


The tablet runs stock Android 7.0 Nougat but unfortunately performance is not silky smooth. Swiping between home screens is quite smooth, definitely more than smooth enough for normal use, but its not as smooth as a high end Android phone or even the Mi Pad 3. There is a bit of stutter opening apps but multi-tasking is great with 4GB of RAM.

I was able to play many games on the tablet, however Need For Speed: No Limits did not run properly because the entire screen was black except for the road signs. Many other games ran well also such as Hearthstone and Clash of Clans chúng tôi fingerprint sensor is not that great, its fairly accurate but its not very fast.

Teclast Master T10


WiFi performance is blazing fast, hitting incredible speeds however WiFi range is poor, cutting it in and out even a mere 15 feet away.

Bluetooth works fine, and GPS performance is fairly slow.

Teclast Master T10


The camera is actually fairly decent, especially for a tablet. The 8MP rear camera takes some fairly OK photos with some fair detail and colour saturation. The rear 13MP camera also takes some good photos as well, definitely more than good enough for video calls and selfies (but selfies with a tablet, seriously?).

Teclast Master T10


Good tablet

Teclast Master T10


Cooler Master Mm731 Mouse Review – Ultra Lightweight But Solid Performance

Cooler Master MM731 Gaming Mouse


Pixart PAW3370





Size (H x W x D)

122.3 x 69.0 x 39.1MM



How We Review Hands-on Review

Shop on Amazon

Check Price


Impressively lightweight

Attractive minimalist design

Reliable wireless performance


Less-than-premium materials


Tech Specs


Pixart PAW3370





Size (H x W x D)

122.3 x 69.0 x 39.1MM














2.4GHz, Bluetooth 5.2, USB type-C

Unboxing & setup

The mouse comes packaged in a utilitarian box with no bells or whistles, just enough packing materials to keep the mouse safe during its travels. The setup is a simple and uncomplicated affair with it lighting up as soon as you plug it in. The wireless functionalities are enabled via the switch on the underside of the mouse and are equally easy to operate. It’s worth noting that the pairing button must be held down for the mouse to become discoverable on your device of choice. It’s not necessary to download the Cooler Master MasterPlus+ software suite to use the MM731, but it’s recommended in order to get the most out of it.


The word ‘innocuous’ comes to mind. The MM731 is devoid of sharp angles and aggressive lines and instead favors smooth ergonomics and an understated color scheme. It strongly resembles the venerable G-Pro from Logitech. We were provided with the matte black version, but a minimalist white version is available too. There is a refreshingly restrained implementation of RGB lighting present here, taking the form of the soft hexagon of Cooler Master’s logo, without the name of the company within. The hexagon entirely disappears when the lighting isn’t active. 

Taking a look at the underside you’ll find a small switch to flip between wired, 2.4GHz, and Bluetooth connectivity. Alongside this, there is a pairing button and a button to cycle through DPI presets. The decision to locate a button to improve gameplay performance on the underside of the mouse is unintuitive and inconvenient. There is no obvious reason why Cooler Master couldn’t have positioned it between the mouse buttons as other manufactures like Corsair and Razer usually do.

Overall, though, the design is smart and professional. It will blend into any setup with ease and will sit on your desk incognito until the lighting activates. Cooler Master has taken no chances with the design here and there’s nothing to criticize about the aesthetic.

Build quality

The mouse is impressively sturdy given its weight or lack thereof. The materials feel decent but more is expected given the price, and while they exhibit very little flex, the sides do bend inwards under unrealistically extreme pressure. Under normal gaming conditions, you can be confident that the materials will hold up. The cable has very soft, robust-feeling fabric shrouding that creates so little friction that it’s easy to forget that this mouse isn’t wireless.


This mouse is designed for both claw-users and palm users, though it favors a palm grip as the lower right-hand side of the mouse extends slightly too far out for ideal pinky finger placement. It’s not uncomfortable by any means, but if you’re a picky claw user, this mouse isn’t for you. 


The mouse was tested in 2.4GHz, Bluetooth, and wired modes. No perceptible lagging or loss of signal occurred in any of them, although top-tier pro gamers may still prefer wired mode. The mouse wheel smoothly slots back and forth which allows for precise usage in gaming – you’ll be hard-pressed to switch past the intended weapon selection with this wheel. The rubber-textured cover over the mouse wheel could benefit from being more robustly attached to the wheel itself as we found that it was easy for your finger to roll the rubber cover around the wheel, instead of actuating it.


Final verdict

The MM731 is a strong contender for the hybrid wired/wireless gaming mouse market. It boasts a minimal, professional design that belies a killer sensor and impressive wireless capabilities. The mouse is designed to accommodate users of all grip types and given the minute weight of just 59g, packing all these features in is an impressive feat of design and engineering.

Let’s Master Pdf Optimization & Compression

When it comes to optimizing content on your site for search engine rankings, nothing beats standard web pages. While PDFs are never a replacement for a web page, there are some definite benefits to using PDFs in the right situations.

Why and When Should You Use PDFs on the Web?

In most instances, a standard web page will give you the flexibility that you need without disrupting the visitor’s on-site experience. But a web page won’t work for every situation, especially if you need or want your visitors to be able to download/print precisely formatted content. With modern technology, most web pages can be set up to print in an appealing format, but you still won’t have the control that a PDF document gives you.

The following types of documents are typical–and usually beneficial–to be published as a PDF:

Ebooks: Typically much longer than a standard web page, often with illustrations and guides. PDFs allow you to ensure the accompanying images are placed and sized correctly. They also allow you to add cover art, along with headers and footers, just like a normal book, which you can’t do from a web page.

Specification Documents: Specification docs are often very format heavy, relying on images and illustrations, which makes the PDF format valuable. However, unless you know your customers frequently print this material, a web page will often suffice, making the PDF unnecessary.

White Papers: Similar to ebooks, white papers are often long-form content that is better provided as a download for printing rather than reading on a screen.

Any Offline Content: Any content that is better read offline rather than online can be made into a PDF. In fact, any info-heavy blog post can benefit by being turned into a downloadable PDF in addition to being available online for those who want to read it away from their screens. However, if taking this route, be aware of the potential for duplicate content. You’ll want to make sure these PDFS are not indexable by the search engines.

Part 1: How to Compress PDFs to Reduce File Size

First and foremost, any optimization you should be doing is for your visitors. With PDFs, optimizing for people largely comes down to reducing the file size of your PDF documents to increase the speed at which it is downloaded. Larger documents will always take a longer time to download, but you can do your audience a huge favor by reducing that time as much as possible.

But first things first. Before we begin optimizing for speed, let’s make sure it’s not already.

Check if your PDF is optimized:

Open your PDF in Adobe Acrobat

Make sure you’re on the Description tab

At the bottom, you’ll see “Fast Web View.” If it says “No,” your PDF needs to be optimized!

So, let’s get to work optimizing!

How to Optimize Fonts

Use Standard Fonts: All PDF readers support a set of standard fonts: Times, Helvetica, Courier, Symbol, and Zapf Dingbats. Any font you use outside of these will likely have to be embedded, which increases the file size of your PDF. If you don’t embed the fonts, Adobe will automatically substitute one of the standard fonts. This, of course, causes you lose one of the main control benefits of using PDFs: Visual layout control.

Use Fewer Fonts: Every font used in your document adds kilobytes to your PDF file size. Do what you can to keep your font variables down to as few as possible.

Limit Font Styles: Not only does using more fonts increase the size but every font style is treated as a different font. So if you have even just one usage of italics, bold or bold italics, you’re adding kilobytes. This becomes especially important when you’re using non-standard fonts that have to be embedded. Each style variation is a new font embed.

Want to know what fonts you’re using and whether or not they are embedded? Follow this procedure:

Open your PDF in Adobe Acrobat

This shows you each font used and if it is embedded.

How to Optimize Images

Use Vector-Based Images: When available, use vector-based image files. These images are smaller and of higher quality than bitmap images.

Use Monochrome Bitmap Images: If you can’t use vector-based images and are using bitmaps instead, make them monochrome rather than color. This will keep bitmap images as small as possible.

How to Reduce PDF Size When Saving From Microsoft Word

Save as Minimum Size: When saving a word document to PDF format, follow this procedure to ensure the PDF is created in the smallest size possible:

Under Save as Type, select PDF

Select Minimum Size

Make sure “ISO 19005-1 compliant (PDF/A)” and “Bitmap text when fonts may not be embedded” are not checked.

How to Optimize PDFs in Adobe Acrobat

You can further optimize your already created PDFs in Adobe Acrobat.

Open PDF in Acrobat

From here you have several options for reducing the file size of your PDF document. The PDF will likely already be set to Standard mode. You can use that to select this mode, Mobile mode, or create your own custom presets. If you change any of the preset standard options, the Settings automatically change to Custom.

Adobe provides a very helpful tutorial on how to utilize the different panels, so no sense in recreating the wheel here. Adjust the custom settings to your liking for maximum speed optimization without sacrificing quality, and you’re all set. Then it’s just a matter of saving and naming your custom settings if you want to retain those for future use.

It’s probably a good idea to have different custom settings based on the content of the PDF. Image-rich PDFs will probably need different optimization settings than text-heavy PDFs.

Fast Web View: There is one last step before you save your newly optimized PDF, and that is to make sure Fast Web View is enabled. This restructures the PDF so pages can be downloaded one at a time, rather than forcing the entire document to be downloaded at once. This is especially important for large PDF documents that are likely to be viewed in the browser. Here is how:

Under Save Settings, make sure Save As optimizes for Fast Web View is checked

Now when you save, you can go back to view the properties, and you’ll see a “Yes” next to Fast Web View.

Still not convinced you have the smallest PDF file possible? There are a number of tools available such as Smallpdf and PDFCompressor that you can use to try to compress your PDF even further, though your results may vary.

Initial View Settings: Adobe also gives you options to set how you want the PDF to be seen upon being opened. Here you can set whether or not the navigation tab is open by default, the visible page layout, window size, and other options.

Play around with the settings to your liking

That’s it for optimizing for speed.

If you want your PDFs to be indexed by the search engines–and have any hope of ranking–you have to implement some additional optimization strategies. But since you can’t edit the code of a PDF like you can a web page, there are a few unique steps that you have to go through.

Use Text-Based Files: Just like a web page, if you want to get your PDFs indexed, you need to make sure they are text rather than imaged-based. Images are great, and can be used in your PDF documents for visual flair or illustrations, but the bulk of the document needs to be plain text.

Search-Friendly File Name: The PDF file name is akin to the URL of a web page. As with web page URLs, make sure the file name is keyword relevant and search friendly. In most cases, the file name will become part of the URL used to access the document. Where your original word document might be “Complete Guide to PDF Optimization.docx” your PDF file name should read, “complete-guide-pdf-optimization.pdf.”

Optimize Your Content: Again, just like any other optimized web page, you want to optimize your PDF content for the keyword topics you want to rank for. In fact, follow all the standard content optimization procedures regarding keyword usage, heading tags, etc.

Add alt text to images: You can add alt text to your PDF images just like a web page. Well, not just like it, because the process is different, but the value is the same. Here’s how:

Open the Tools menu (it’s usually one of a set of links on the far right of the header)

Select Action Wizard

Select Make Accessible action

This will find all the images in the document with alt text and allow you to add it. Be succinct, but descriptive, using keywords when warranted.

Set Reading Language: While you’re in there, go ahead and set the reading language to whatever suits you. I have no idea if search engines use this or not, but it can’t hurt.

Optimize PDF Properties: Where web pages have meta information, PDFs have properties that you can optimize:

Open document in Adobe Acrobat

Make sure you’re on Description tab

The remaining fields have little to no search impact and can be ignored. However, you may want to treat the subject similar to a meta description. You never know if the search engines will choose to use that to support your optimized title when displaying in search results.

One thing you may want to utilize here, however, is the copyright status. Change that to Copyrighted or Public Domain, based on what fits the document, and add a copyright notice in the box below or URL. None of these have any impact on search, but it can’t hurt to fill these in for your own publishing protection.

Link Out of Your PDF: Since you’re optimizing the PDF for web viewing (as well as download), you might as well add some strategic links in the PDF to your website. While a PDF doesn’t have a navigation like your site does, that doesn’t mean you can’t add relevant links in the content. Be sure to use relevant anchor text so, once your PDF is indexed, the links will count like standard web page links.

One note to this: If your PDF is primarily for off-line viewing, you will want to use the link URL rather than (or in addition to) the linked keyword text. This will at least give printed-page readers a way to visit the reference if they are inclined to type the full URL into their browser later.

Check Compatibility: It’s a good idea to avoid saving your PDF compatibility with the most recent version of Acrobat. Go a couple of versions lower than the latest to ensure most readers or search engines won’t have issues viewing the document.

Link to Your PDF: Search engines find documents and pages via links in other web pages. Be sure to strategically link to your PDF just as you would any other page to ensure it gets found by the search engines.

Any good content strategy will undoubtedly include content that is best turned into a PDF. I mentioned above several reasons why PDFs are better than HTML pages, and there are many more.

While PDFs are often created for various web marketing campaigns, optimization of those PDFs is often overlooked. Unfortunately, this is to the overall detriment of the optimization campaign’s success.

Any PDF that isn’t behind a subscription wall should be optimized for both search engines and visitors alike. Optimizing your PDFs creates a nearly seamless web experience for any visitor who comes across these valuable documents.

Image Credits

Master The ‘Houseparty Lingo’ With These Tips

As the world strives to find a sense of normalcy during the ongoing COVID-19 pandemic, video calling applications have emerged as our sole, irrefutable lifelines. From Google Duo to Skype — the selection of traditional video calling applications is quite extensive. Sadly, none of these do enough to take your mind off mundane conversations and offer something a bit more engaging.

So, if you’re looking for a bit more fun and a little less worry, you’d be wise to take the road less taken and explore Houseparty — an app that has successfully fused video calling with interactive games.

Sure, there have been a few reports of compromised security, but none of them were conclusively proven. The company even held a press conference, clearing the doubt and assuring no data breach had occurred.

Now with that out of the way, let’s focus back on today’s discussion: getting you accustomed to the Houseparty language. We’ll share pretty much everything you need to know to get around the app and eventually make the most out of it.

Related: How to Use Houseparty Privately

What is different about Houseparty?

In an attempt to stand out from the crowd of conventional video calling applications, Houseparty uses innovative terms and phrases to denote even the simplest of tasks.

For newbies, mastering the language can be a little challenging, and that’s where we hope to come in. Check out the subsections below to learn about the many quirks of Houseparty.

Related: How to play Houseparty on PC

Houseparty Lingo: Tips to know

Passing a Note

In regular terms, it’s called sending a text message. Simply tap on one of your contacts and then hit ‘Pass a Note.’ After you’re satisfied with the message, hit send.


Houseparty, by default, notifies your friends when you log in to the app. And while that’s great for people you want to get in touch with, it can be a bit annoying if you’re looking to fly below the radar.

By Ghosting a specific contact, you could disable the notification we just mentioned for the individual, while still notifying your other important contacts. To do so, tap on the contact’s name, go to settings, and toggle on Ghosting.

House Party has changed ‘Ghosting’ in the latest update, making it more accessible for casual users. Instead of turning on ghosting, you’ll now need to toggle off ‘Send Notifications’ to ghost a contact. Similarly, you could toggle off ‘Get Notifications’ to stop getting alerts every time the user logs in to Houseparty.

Sneak into the house

This feature is quite similar to Ghosting but on a grander scale. While Ghosting allows you to evade a single individual, sneaking in lets you hide from all your friends. Do ‘Sneak into the house,’ press and hold the game icon, and tap on ‘Sneak into the house.’


This feature is identical to Facebook Messenger’s ‘Wave’ functionality. By waving, you notify a friend that you are available to talk. To wave, tap on a contact and hit ‘Say Hi.’

Send a Facemail

If you are familiar with Google Duo’s video messages, you’re going to feel right at home, here. Houseparty allows you to record short video messages — called a Facemail — and send it to as many contacts as you like.

Please note that the sender gets notified when the receiver watches their Facemail. So, it’s better to avoid reading Facemails when you’re sneaking in the house.

What is We Time?

Houseparty wishes to be the bridge between you and your contacts even you are thousands of miles apart. By offering this incentive called ‘We Time,’ the app encourages you to make more calls and stay within touching distance. This timer starts when you and your friend have been talking for 120 minutes.

Every minute you add beyond the 2-hour mark is added to your ‘We Time’ counter. Similar to the Snapchat streak, ‘We Time,’ too, needs regular maintenance. If you don’t talk with a contact for over 2 days, your We Time counter is reset and starts from 0.

Locking a Room

Houseparty allows your friends to see when you are online, who you are talking to, and whether you are available to chat. They can then simply join your room and start a video calling session. This feature makes Houseparty more accessible than other apps.

However, if you are looking for alone time with someone or don’t want to ruin your zen solo session, you could simply choose to lock your room. Friends would still be able to send a joining request, but you can easily ignore them. To lock your room, simply tap on the padlock button at the bottom of your screen.

If you want to learn deep dive into the world of Houseparty games, be sure to check out our dedicated article: “Cool Houseparty games to play right now.”

Master Dimensionality Reduction With These 5 Must


Singular Value Decomposition (SVD) is a common dimensionality reduction technique in data science

We will discuss 5 must-know applications of SVD here and understand their role in data science

We will also see three different ways of implementing SVD in Python


“Another day has passed, and I still haven’t used y = mx + b.“

Sounds familiar? I often hear my school and college acquaintances complain that the algebra equations they spent so much time on are essentially useless in the real world.

Well – I can assure you that’s simply not true. Especially if you want to carve out a career in data science.

Linear algebra bridges the gap between theory and practical implementation of concepts. A healthy understanding of linear algebra opens doors to machine learning algorithms we thought were impossible to understand. And one such use of linear algebra is in Singular Value Decomposition (SVD) for dimensionality reduction.

You must have come across SVD a lot in data science. It’s everywhere, especially when we’re dealing with dimensionality reduction. But what is it? How does it work? And what are SVD’s applications?

I briefly mentioned SVD and its applications in my article on the Applications of Linear Algebra in Data Science. In fact, SVD is the foundation of Recommendation Systems that are at the heart of huge companies like Google, YouTube, Amazon, Facebook and many more.

We will look at five super useful applications of SVD in this article. But we won’t stop there – we will explore how we can use SVD in Python in three different ways as well.

And if you’re looking for a one-stop-shop to learn all machine learning concepts, we have put together one of the most comprehensive courses available anywhere. Make sure you check it out (and yes, SVD is in there as part of the dimensionality reduction module).

Table of Contents

Applications of Singular Value Decomposition (SVD)

Image Compression

Image Recovery


Spectral Clustering

Background Removal from Videos

What is Singular Value Decomposition?

Rank of a Matrix

Singular Value Decomposition

Why is SVD used in Dimensionality Reduction?

3 Ways to Perform SVD in Python

Applications of Singular Value Decomposition (SVD)

We are going to follow a top-down approach here and discuss the applications first. I have explained the math behind SVD after the applications for those interested in how it works underneath.

You just need to know four things to understand the applications:

SVD is the decomposition of a matrix A into 3 matrices – U, S, and V

S is the diagonal matrix of singular values. Think of singular values as the importance values of different features in the matrix

The rank of a matrix is a measure of the unique information stored in a matrix. Higher the rank, more the information

Eigenvectors of a matrix are directions of maximum spread or variance of data

In most of the applications, the basic principle of Dimensionality Reduction is used. You want to reduce a high-rank matrix to a low-rank matrix while preserving important information.

SVD for Image Compression

It minimizes the size of an image in bytes to an acceptable level of quality. This means that you are able to store more images in the same disk space as compared to before.

Here’s how you can code this in Python:

Python Code:


If you ask me, even the last image (with n_components = 100) is quite impressive. I would not have guessed that it was compressed if I did not have the other images for comparison.

SVD for Image Recovery

We’ll understand image recovery through the concept of matrix completion (and a cool Netflix example).

Matrix Completion is the process of filling in the missing entries in a partially observed matrix. The Netflix problem is a common example of this.

The basic fact that helps to solve this problem is that most users have a pattern in the movies they watch and in the ratings they give to these movies. So, the ratings-matrix has little unique information. This means that a low-rank matrix would be able to provide a good enough approximation for the matrix.

This is what we achieve with the help of SVD.

Where else do you see this property? Yes, in matrices of images! Since an image is contiguous, the values of most pixels depend on the pixels around them. So a low-rank matrix can be a good approximation of these images.

Here is a snapshot of the results:

Chen, Zihan. “Singular Value Decomposition and its Applications in Image Processing.”  ACM, 2023

SVD for Eigenfaces

The original paper Eigenfaces for Recognition came out in 1991. Before this, most of the approaches for facial recognition dealt with identifying individual features such as the eyes or the nose and developing a face model by the position, size, and relationships among these features.

The Eigenface approach sought to extract the relevant information in a face image, encode it as efficiently as possible, and compare one face encoding with a database of models encoded similarly.

The encoding is obtained by expressing each face as a linear combination of the selected eigenfaces in the new face space.

Let me break the approach down into five steps:

Collect a training set of faces as the training set

Find the most important features by finding the directions of maximum variance – the eigenvectors or the eigenfaces

Choose top M eigenfaces corresponding to the highest eigenvalues. These eigenfaces now define a new face space

Project all the data in this face space

For a new face, project it into the new face space, find the closest face(s) in the space, and classify the face as a known or an unknown face

You can find these eigenfaces using both PCA and SVD. Here is the first of several eigenfaces I obtained after performing SVD on the Labelled Faces in the Wild dataset:

As we can see, only the images in the first few rows look like actual faces. Others look noisy and hence I discarded them. I preserved a total of 120 eigenfaces and transformed the data into the new face space. Then I used the k-nearest neighbors classifier to predict the names based on the faces.

You can see the classification report below. Clearly, there is scope for improvement. You can try adjusting the number of eigenfaces to preserve and experiment with different classifiers:

Have a look at some of the predictions and their true labels:

You can find my attempt at Facial Recognition using Eigenfaces here.

SVD for Spectral Clustering

Clustering is the task of grouping similar objects together. It is an unsupervised machine learning technique. For most of us, clustering is synonymous with K-Means Clustering – a simple but powerful algorithm. However, it is not always the most accurate.

Consider the below case:

Clearly, there are 2 clusters in concentric circles. But KMeans with n_clusters = 2 gives the following clusters:

K-Means is definitely not the appropriate algorithm to use here. Spectral clustering is a technique that combats this. It has roots in Graph theory. These are the basic steps:

Start with the Affinity matrix (A) or the Adjacency matrix of the data. This represents how similar one object is to another. In a graph, this would represent if an edge existed between the points or not

Find the Laplacian (L) of the Affinity Matrix: L = A – D

Find the highest k eigenvectors of the Laplacian Matrix depending on their eigenvalues

Run k-means on these eigenvectors to cluster the objects into k classes

You can read about the complete algorithm and its math here. The implementation of Spectral Clustering in scikit-learn is similar to KMeans:

View the code on Gist.

You will obtain the below perfectly clustered data from the above code:

SVD for Removing Background from Videos

I have always been curious how all those TV commercials and programs manage to get a cool background behind the actors. While this can be done manually, why put in that much manual effort when you have machine learning?

Think of how you would distinguish the background of a video from its foreground. The background of a video is essentially static – it does not see a lot of movement. All the movement is seen in the foreground. This is the property that we exploit to separate the background from the foreground.

Here are the steps we can follow for implementing this approach:

Create matrix M from video – This is done by sampling image snapshots from the video at regular intervals, flattening these image matrices to arrays, and storing them as the columns of matrix M

We get the following plot for matrix M:

What do you think these horizontal and wavy lines represent? Take a moment to think about this.

The horizontal lines represent the pixel values that do not change throughout the video. So essentially, these represent the background in the video. The wavy lines show movement and represent the foreground.

We can, therefore, think of M as being the sum of two matrices – one representing the background and other the foreground

The background matrix does not see a variation in pixels and is thus redundant i.e. it does not have a lot of unique information. So, it is a low-rank matrix

So, a low-rank approximation of M is the background matrix. We use SVD in this step

We can obtain the foreground matrix by simply subtracting the background matrix from the matrix M

Here is a frame of the video after removing the background:

Pretty impressive, right?

We have discussed five very useful applications of SVD so far. But how does the math behind SVD actually work? And how useful is it for us as data scientists? Let’s understand these points in the next section.

What is Singular Value Decomposition (SVD)?

I have used the term rank a lot in this article. In fact, through all the literature on SVD and its applications, you will encounter the term “rank of a matrix” very frequently. So let us start by understanding what this is.

Rank of a Matrix

The rank of a matrix is the maximum number of linearly independent row (or column) vectors in the matrix. A vector r is said to be linearly independent of vectors r1 and r2 if it cannot be expressed as a linear combination of r1 and r2.

Consider the three matrices below:

In matrix A, row r2 is a multiple of r1, r2 = 2 r1, so it has only one independent row. Rank(A) = 1

In matrix B, row r3 is a sum of  r1 and r2, r3 = r1 + r2, but r1 and r2 are independent. Rank(B) = 2

In matrix C, all 3 rows are independent of each other. Rank(C) = 3

The rank of a matrix can be thought of as a representative of the amount of unique information represented by the matrix. Higher the rank, higher the information.

Singular Value Decomposition (SVD)

So where does SVD fit into the overall picture? SVD deals with decomposing a matrix into a product of 3 matrices as shown:

If the dimensions of A are m x n:

U is an m x m matrix of Left Singular Vectors

S is an m x n rectangular diagonal matrix of Singular Values arranged in decreasing order

V is an n x n matrix of Right Singular Vectors

Why is SVD used in Dimensionality Reduction?

You might be wondering why we should go through with this seemingly painstaking decomposition. The reason can be understood by an alternate representation of the decomposition. See the figure below:

The decomposition allows us to express our original matrix as a linear combination of low-rank matrices.

In a practical application, you will observe that only the first few, say k, singular values are large. The rest of the singular values approach zero. As a result, terms except the first few can be ignored without losing much of the information. See how the matrices are truncated in the figure below:

To summarize:

Using SVD, we are able to represent our large matrix A by 3 smaller matrices U, S and V

This is helpful in large computations

We can obtain a k-rank approximation of A. To do this, select the first k singular values and truncate the 3 matrices accordingly

3 Ways to Perform SVD in Python

We know what SVD is, how it works, and where it is used in the real world. But how can we implement SVD on our own?

The concept of SVD sounds complex enough. You might be wondering how to find the 3 matrices U, S, and V. It is a long process if we were to calculate these by hand.

Fortunately, we do not need to perform these calculations manually. We can implement SVD in Python in three simple ways.

SVD in NumPy

NumPy is the fundamental package for Scientific Computing in Python. It has useful Linear Algebra capabilities along with other applications.

You can obtain the complete matrices U, S, and V using SVD in numpy.linalg. Note that S is a diagonal matrix which means that most of its entries are zeros. This is called a sparse matrix. To save space, S is returned as a 1D array of singular values instead of the complete 2D matrix.

View the code on Gist.

Truncated SVD in scikit-learn

In most common applications, we do not want to find the complete matrices U, S and V. We saw this in dimensionality reduction and Latent Semantic Analysis, remember?

We are ultimately going to trim our matrices, so why find the complete matrices in the first place?

In such cases, it is better to use TruncatedSVD from sklearn.decomposition. You specify the number of features you want in the output as the n_components parameter. n_components should be strictly less than the number of features in the input matrix:

View the code on Gist.

Randomized SVD in scikit-learn

Randomized SVD gives the same results as Truncated SVD and has a faster computation time. While Truncated SVD uses an exact solver ARPACK, Randomized SVD uses approximation techniques.

View the code on Gist.

End Notes

I really feel Singular Value Decomposition is underrated. It is an important fundamental concept of Linear Algebra and its applications are so cool! Trust me, what we saw is just a fraction of SVD’s numerous uses.

I encourage you to check out this Comprehensive Guide to build Recommendation Engine from scratch to realize the power of SVD for yourself. Building this project will surely add value to your resume (and enhance your own skillset!).


Update the detailed information about Review: Master & Dynamic Mw07 Offer Great Sound In True Wireless In on the website. We hope the article's content will meet your needs, and we will regularly update the information to provide you with the fastest and most accurate information. Have a great day!