Deep Learning with TensorFlow and Intel – a hardware and software guide for beginners

Deep learning is the next big thing in tech, with applications in image processing, speech processing and natural language processing taking off, all based on the simple idea of an artificial neural network. There’s no dearth of pre-existing frameworks with which to implement a range of pre-designed models, many of which even have pre-trained weights – so you can jump right into the inference stage. But in order to add even a little customizability to your deep learning application, it will probably be necessary to have the right training and inference hardware.

Deep learning is a computationally intensive process, particularly during training, but also for inference. It involves a huge number of linear algebra operations, and this means that a well-planned hardware setup can go a long way in making your application perform well at the scale you need it to. But this is easier said than done, especially when on a budget. A faster computer should have a CPU with more cores and higher clock speed, right? But wait, what about GPUs or Graphical Processing Units, aren’t they supposed to be better at linear algebra operations? And what are these new kids on the block – FPGAs and ASICs?

Here is a step-by-step guide to setting up a workstation to get started with Deep Learning without breaking the bank. While there are many libraries for implementing deep learning topologies, this article is geared towards one of the most popular deep learning frameworks in use today – TensorFlow.

What you will need

Typically, a fully functional deep learning rig for non-commercial purposes – including research, competitions and even most initial-stage startups – will require the following:

A processing unit for desktop and development (a CPU with enough cores and frequency will suffice)

OR

a high-powered scalable CPU that can really crunch the numbers

OR

A processing unit for the deep learning computations (such as a CPU or GPU)
CPU or GPU cooling systems
Plenty of RAM (in which to load your potentially huge datasets)
A hard drive (preferably a Solid State Drive or SSD)
A power supply unit (PSU)
A motherboard and perhaps a case (if you want things to look pretty)
A monitor (or two, or even three!)

 

The choice of hardware

For deep learning at scale, the CPU is quickly becoming a more and more viable option. Intel’s latest range of CPUs is optimized for deep learning calculations using the Intel® Math Kernel Library for Deep Neural Networks (Intel® MKL-DNN), and it’s actually an excellent one-size-fits-all solution for a standard deep-learning-enabled processor.

However, it’s not so cut and dry when you mention CPUs because there are a vast number of choices when it comes to processors and it really comes down to what kind of application you wish to implement your deep learning project on. Building a deep learning model consists of two broad phases – training and inferencing. If you look at the entire lifecycle of a DL model then your hardware will spend most of its time on the inferencing part, so if you’re starting out and don’t want to invest in specialized hardware for every stage of the process, then the most prudent course of action is to invest in hardware that’s focused on inference but can also handle the training part of it. It might not appear the right choice but current gen CPUs such as the Intel® Xeon® Scalable Processor family are built specifically to run high-performance AI workloads. 

Field Programmable Gate Arrays are typically used in deep learning only for very specific optimizations for very high speed and aren’t likely to be necessary for your applications. And ASICs (Application-Specific Integrated Circuits) are a kind of accelerated chip that can be customized for AI operations. So you really need to figure out what kind of deep learning model you wish to deploy if you want the best performance. In fact, Intel has a wide portfolio of specialised hardware such as the Intel® FPGAs, Intel® Movidius™ Myriad™ VPUs and Intel® NCS 2, which are suited for different AI workloads. Since we’re starting out, we’ll stick with CPUs for now.

How to pick a processor

Your desktop Intel Core Processors can run inference workloads. In fact, the Intel Xeon Scalable Processors (previously known as Skylake-SP) are designed with AI and HPC workloads in mind. With improved memory and I/O bandwidth, your DL models face fewer bottlenecks. One of the key advantages of going with the new processor family is the introduction of the Intel® Advanced Vector Extensions 512 (Intel® AVX 512) instruction set that increases parallelism and vectorization which are important for AI workloads. 

With four lineups – bronze, silver, gold and platinum – the Intel® Xeon® Scalable Processor lineup gives you a wide range of processors to suit your investment. The bronze lineup is extremely affordable with the 6-core Intel® Xeon® Bronze 3104 Processor being one of the most popular SKUs and as your requirements scale, you can simply upgrade to a Silver or Gold CPU since the sockets are the same. 

Getting a GPU

In some cases, offloading compute workload to a GPU might execute a task quicker. There are a few simple points to keep in mind while getting a GPU, the first of which is compatible with your CPU and motherboard. However, before you do that, you should figure out what kind of workloads you will be running.
One thing to note is that most of the mainstream desktop CPUs in Intel’s portfolio come with Intel® HD Graphics and they’re quite useful if you wish to run inference workloads. You can make use of clDNN (Compute Library for Deep Neural Networks) which is a library of kernels to accelerate deep learning on Intel Processor Graphics. They’re based on OpenCL and are able to accelerate many of the common functions calls made to popular topologies such as AlexNet, VGG, GoogLeNet, ResNet, Faster-RCNN, SqueezeNet, and FCN. So based on the workload, might not even need a discrete accelerator.

RAM

The first important point to note is the RAM speed is one of the least useful things you can spend your money on. Speed is essential in moving data sets from the cold storage to your memory and capacity plays a role in how much of the dataset you can store closer to the CPU. Moving data frequently from the cold storage to the RAM isn’t ideal. It would be better if you got an SSD as a replacement for your hard drive, to begin with. Investing in SSDs such as the Intel® Optane™ SSD DC P4800X would help in this regard. Coming back to the RAM aspect, a good rule of thumb is to populate all memory channels first and have as much capacity as your dataset requires.  

CPU Cooling Systems

The importance of this component of your rig cannot be overstated. An overheating CPU is not only going to be throttled but will also have a shorter lifetime of usage – and worst of all, might cause a fire! 

We’ve mentioned how you can even run inference workloads on the integrated graphics present on CPUs. It’s natural to question whether this would require additional cooling. Since this inference workload is performed on the IGP (Integrated Graphics Processor) which is present on the CPU die, you don’t have to make separate arrangements for cooling. You can simply refer to the datasheet for the said processor and buy a cooler that can handle the TDP that the CPU is rated for. As long as you are not overclocking your CPU, you do not need a custom cooling solution.

Hard Drive (preferably a Solid State Drive)

Good data-feeding practices are key to making the most of a hard drive. Reading data from disk at run-time is a universally slow option and should be avoided at all costs. An asynchronous call to the same will be orders of magnitude faster. Your hard drive can really be as basic as it gets, and these simple principles will hold true.

A solid-state drive, however, is a nice high-performance addition that does give you a useful speed boost. Most systems combine an SSD with a traditional hard drive, using the larger and slower hard drive for storing data, and the SSD for high-productivity tasks and frequently accessed data that one would prefer not to always have in RAM.

The Intel® Optane™ family offers a nice set NVMe solid state drives. These make use of 3D XPoint™ memory and are great to replace your existing hard drives. Intel® Optane™ SSD 900P Series offers low latency coupled with high IOPS. 

Power Supply Unit

There’s not much to say here. Investing in a PSU with a high rating for efficiency is always a good idea. This will also prolong the life and efficiency of your hardware. Do make sure that your PSU has enough connectors for all the components you are using. Also, the PSU should have enough wattage on the 12V rails to handle the requirements of the CPU. There are processors with C-states that might be incompatible with older PSUs so you should always check your CPU datasheet to know its wattage requirements and also check the PSU datasheet for compatibility with the processor family.

Motherboard

There are two major factors to take care of while selecting a motherboard. Firstly, make sure there are enough PCIe slots to connect all your devices and that it can support the devices you intend to use. Whether you are using GPUs or FPGAs such as the Intel® Stratix® 10, you need to ensure that there are enough PCIe lanes to satisfy the requirements of each add-in card. 

Computer case

Ensure that your case can contain the full-length add-in cards that have been inserted. It may be helpful to assemble the setup on trial, take some measurements and only then order a case. If you have a cooling rig, especially with water-cooling, you may need extra space for this as well. Nothing is worse than ordering the perfect case only to find that your parts simply won’t fit. If you’re getting a workstation motherboard then there are specialised cases which fit dual-socket motherboards and have plenty of mounting options to install water cooling components. 

Another note of caution – water-cooling rigs will not work well if any of the tubes are bent at awkward angles! This is extremely important and can be a major safety hazard. Ensure that your case is large enough to have clearance on these tubes. 

Getting started with Deep Learning

Now that your hands are literally dirty from assembling your rig, it’s time to get your hands metaphorically dirty with the stuff you came here for – deep learning.

We’ll be discussing TensorFlow, one of the most popular deep learning packages available, most commonly used in Python, though written in C++ and in possession of a growing C API. We will, in this tutorial, be working with Python 3.4.

With TensorFlow, you can work on pretty much any application in deep learning, from image recognition and processing to speech processing to natural language processing. There is also no dearth of ready-made models for you to play with and explore. TensorFlow has rapidly become the industry standard, and it’s a great place to get started.

One point to note is that TensorFlow has a slightly unusual computation scheme which might be particularly intimidating to novice programmers. The computations are built into a ‘computation graph’ which is then run all at once. So to add two variables, a and b, TensorFlow would first encode the ‘computation’ a+b into a computation graph. Before running this graph, trying to access this graph will not give a result – the result hasn’t been processed yet! Instead, it will give you the graph. Only after running the graph will you have access to the actual answer. Bear this in mind, as it will help clear confusions in your later explorations with TensorFlow.

One of the best things about using TensorFlow with Intel hardware is the excellent hardware optimization available. Intel has worked to make the TensorFlow framework optimized using the Intel® MKL-DNN implemented on Intel processors. The Intel® MKL-DNN primitives make use of efficient algorithms and data-structure in conjunction with Intel’s hardware-specific upgrades, for significantly faster computations in deep learning.

The best part is that there is nothing new that you have to do to get access to these optimizations. The standard installation of TensorFlow that we will detail below will get you set with everything you need.

Installing TensorFlow (and Python before it)

Let’s start off by installing all the pre-requisites for our Linux system. If you don’t yet have Python, the best course of action would be to install Anaconda’s version of it. You can do this here. We’ll be working with Python 3.4. If you only have Python 2, you may need to create a virtual environment and install Python 3. If you have Python but not Anaconda, that’s alright – we’ll use pip.

Here’s how you install the Intel-optimized build of TensorFlow.

Open the Terminal and type:

● For Anaconda users on Windows:
conda install tensorflow
● For Anaconda users on macOS or Linux:
conda install tensorflow-mkl
● For non-Anaconda users on any operating system:
pip install intel-tensorflow

You can also download the tarball and install from source, or simply use a Docker image. More details on these methods can be found here.

Building your first model

One of the most popular first examples with deep learning is the MNIST digit classification application. This is a dataset of hand-written digits, the numbers from 0 to 9. The benchmark model is a convolutional neural network, a well-crafted system that helps detect the most useful edges and combinations of edges – thus identifying the digit in the picture.

We get started by importing the TensorFlow library:

import tensorflow as tf

We then import our dataset, which is thankfully included under TensorFlow’s in-built datasets.

mnist = tf.keras.datasets.mnist

We split our data into test and train sets:

(x_train, y_train),(x_test, y_test) = mnist.load_data()
x_train, x_test = x_train / 255.0, x_test / 255.0

We build our neural network by stacking layers one after the other. First, a ‘Flatten’ layer reduces the matrix of pixel values to a flat array, then a Dense network connect fully to all nodes. A Dropout layer helps to reduce overfitting, and another Dense layer will give us our output probabilities.

model = tf.keras.models.Sequential([
  tf.keras.layers.Flatten(),
  tf.keras.layers.Dense(512, activation=tf.nn.relu),
  tf.keras.layers.Dropout(0.2),
  tf.keras.layers.Dense(10, activation=tf.nn.softmax)
])

Compiling the model gets us ready to run the computation graph:

model.compile(optimizer=’adam’,
              loss=’sparse_categorical_crossentropy’,
              metrics=[‘accuracy’])

And finally, we actually train and then evaluate the model:

model.fit(x_train, y_train, epochs=5)
model.evaluate(x_test, y_test)

Of course, there are plenty of ways to play around with these models. You can play with the deep learning architecture, the optimization algorithm, the type of loss function used, and much more. In terms of evaluating performance, you can keep track of the loss per epoch, or select different metrics of accuracy.

Be sure to look over the tutorials on the TensorFlow website for much more. The learning curve is pretty easy to scale, and you can be working on custom models on your slick custom deep learning rig in very little time

In conclusion

Getting started with deep learning is really not so hard, even for cool new research or for working on a pet project or startup. If you wish to just test out a fancy new DL library then your existing desktop PC is more than enough but if you are seriously considering on implementing a Deep Learning model then a workstation offers you the flexibility of hardware while keeping it on premises. Additionally, there are options wherein you can rent cloud instances for DL workloads. So you should always perform a cost-benefit analysis before investing. The cloud option is easier because there’s no significant upfront investment but depending on how fast your internet connection is, you may end up spending a lot of time moving dataset from a local on-prem storage to the cloud storage before you can use it in your training model. 

There’s a whole world of interesting deep learning applications to explore. Cutting edge technology like recurrent neural networks (RNNs) and Long Short-Term Memory (LSTM) are potential game-changers in this space. Here’s to the next disruptive technology potentially coming from your improvised, self-built workstation.

Source links:

https://software.intel.com/en-us/articles/intel-optimization-for-tensorflow-installation-guide
https://www.intel.in/content/www/in/en/processors/xeon/scalable/xeon-scalable-platform.html
https://github.com/IntelAI/modelshttps://en.wikichip.org/wiki/intel/uhd_graphics/630
https://www.tensorflow.org/install
https://www.tensorflow.org/install/source_windows

A Full Hardware Guide to Deep Learning


https://www.tensorflow.org/tutorials/
https://software.intel.com/en-us/articles/accelerate-deep-learning-inference-with-integrated-intel-processor-graphics-rev-2-0

[Sponsored Post]

Viewsonic M1 Projector Review: A feature rich portable projector with good audio

There are quite a few portable projectors available in the market and if you have a budget of about 35,000 there are plenty to choose from. One such projector recently made it to our test labs. The Viewsonic M1 is a battery powered projector with Harman Kardon speakers and it has quite a rich feature-set. It has an MRP of Rs 49,500 but is available online for Rs 35800. Is it a worthy contender for your money in the portable projector category?

Key specifications at a glance

Weight: 689 grams
Display resolution: 854×480
Brightness: 250 Lumens
Screen Size: 100-inches from 9 feet
Input: HDMI, USB-C, USB Type A, MicroSD,
Audio Out: 3.5mm
Built-in speakers: 2 X 3W
Built-in battery life: up to 6 hours (company claims)
Wireless: No
Price: MRP: 49,500 . Available online for Rs. 35,800

Build and design

Kicking things off with the build and design of the projector, it isn’t the smallest or the lightest portable projector we have seen but it is compact and portable. It won’t fit in your pocket like the Rayo i5 mini projector we reviewed earlier (read our review here) and that’s ok as the bigger size brings with it a set of interesting features. The projector has a hard plastic shell with some metal thrown into the mix. It has a gunmetal finish giving it a premium look. It is quite well built considering its size and form factor. The plastic grill on the top of the projector houses the speakers. There is a flap on the side which houses the connectivity options. The bottom of the projector has a standard mounting port. 

The physical controls along with an LED battery indicator rest at the back of the Viewsonic M1. The right of the device has vents to let out heat from the projector. On the left, we have the focus wheel as well.

The front of the projector has a lens cover that also doubles up as a stand. The stand feels sturdy, well built and helps angle the projector however you like. You don’t need to place it facing the wall directly. You can incline or decline the projector based on your table set up. This is nice if the wall is slightly higher or lower than the table where you place the projector giving it flexibility in terms of placement. The lens cover cum stand feels sturdy and it never felt like it would lose its grip on the projector after a lot of use. 

The projector also comes with a remote control in the box which is almost the same dimensions as the Amazon Fire TV remote control. It has all the essential functions like settings, volume, playback, vertical keystoning, and more. It runs on 2 AAA batteries and has a tapering design which adds to the ergonomics. I really like the build and design of the remote control except for one feature. The centre dial houses the rewind and fast forward function along with the vertical keystoning. There are quite a few times I used the keystoning buttons to try and navigate the projector and this takes some getting used to. A small con in an otherwise functional remote control.

Overall, the build of the projector is very good, better than other budget projectors that we’ve seen in the recent past. The remote control is well built, functional, and you don’t need to point it at the projector directly for it to function which is nice. 

Connectivity options and setup

When it comes to connectivity options, this projector has them all. On the left, you have the focus wheel. Below the focus wheel, you have a flap that hides a microSD slot, the DC in, a USB-C port, HDMI port, headphone out and a USB-A port. You can use the headphones out to connect the device to a pair of speakers but that’s something you may not need to do as the projector boasts of speakers by Harman Kardon. More on that in the performance section. 

At the back, you have the power cum enter button, volume controls along with navigating the projector and a back key. 

Setting up the projector is a breeze. You power it on and you are in the UI. there are no smart capabilities like Wi-Fi connectivity or apps which is okay. This is a projector meant to be used on the go with a host of connectivity options. When you fire up the projector, at the top you have the source selection options and at the bottom right you have things like settings, wallpaper and information. You also have a battery indicator in the top right corner. 

Setting up the projector is as easy as plugging in your desired source and hitting play. It also has 16GB of built-in storage so you can copy your favourite movies or TV shows or even music to it from your USB or SD card eliminating the need for an input source.  

Performance

When it comes to performance, we tested the projector using the HDMI and USB source. For the HDMI, we connected a Fire TV stick to access our streaming services and used the USB port to playback video and music files.

When it comes to the visuals, the projector can give you a hundred-inch screen from a distance of 9 feet, but we say, don’t go that far. The projector has a WVGA resolution of 854×480 pixels which isn’t even 720p. Sticking to around 50-inches works but we wish the projector were 720p at least as the trained eye can see the low resolution. 

While playing back videos from the Fire TV Stick, we saw movies like Mission Impossible: Ghost Protocol, John Wick and some Young Sheldon. The Viewsonic M1 projector does an excellent job of immersing you especially if you dim the lights. It will work well with a few lights in the room but you will need to bring up the brightness of the projector to full. Visually, the projector works quite well despite its low resolution. The red circle fight sequence in John Wick in a nightclub filled with blacks, blues and reds, mixed with a lot of gunfire was exhilarating to watch. Even the audio from the projector was engaging. It isn’t the clearest we have heard when thinking of the audio quality alone but consider the fact that it is packed into this tiny projector and you will be impressed. At a little less than 50 per cent volume, the speaker from this projector delivered quite a bang in the action sequences while maintaining clarity in dialogue. 

Moving over to shows like Young Sheldon, where the vocals are more important, you are met with clear dialogues which are loud, and very well audible. 

Moving over to music, with a little tap of the Harman Kardon button on the remote control you can change the audio preset of the projector or if you want, you can do so by delving into the settings of the projector. Once again, we have heard dedicated Bluetooth speakers that sound better than the speakers in this projector, but considering the package on offer, they sound pretty good. Even when you hear a song like Starboy by Daft Punk, the bass is punchy mixed with electronic music. It made me bob my head.

You also have a bunch of display modes to choose from such as Brightest, Standard, Theatre and Music. All of them are self-explanatory with the music mode disabling the display to give you better battery life. Speaking of battery life, Viewsonic claims 6 hours of playback and if you use the device conservatively, you can get there. However, bump up the brightness, crank up the volume and you have a portable experience that will last shy of 4 hours. That’s 2 Marvel movies I got through on the Viewsonic M1. 

Moving away from the audio-video performance, the projector has a few handy features worth nothing. Most important of them is auto keystoning. When you angle the projector, the image appears like a quadrilateral, or a trapezium or even a rhombus at times. You can manually adjust the corners (using the keystone alignment) to get it in the right rectangular shape, or you could let the projector do it for you by simply switching the “auto keystone” feature on. This is a really handy feature as it ensures you get the picture just right. 

By now you think we absolutely love this projector and you’d be right in assuming so. Considering its size, price and performance, it is impressive. However, there are a few niggles we have with the device. The first one is the placement of the keystoning button on the remote control. Pressing this by mistake automatically disables the auto keystone feature which has to be enabled manually by going into the settings. The other is the resolution, which we wish was 720p. The last is the fact that despite the vast array of formats the projector recognises, it didn’t read some of the audio from our movie rips. We played Spider-Man: Into the Spiderverse and we have a 4GB rip with Dolby 5.1 audio. The file is an h.264 mkv file and even though the video ran butter smooth, it failed to give us a sound output. Other files like MP4, AVI, etc. ran without a hitch. Since the UI is Android-based, we wish the device came with a version of VLC or MX player to decode all file formats. 

Bottom Line

For a street price of about Rs 35,000 online (MRP Rs 49,500), the Viewsonic M1 projector powered by Harman Kardon speakers is a treat for those looking to enjoy content on the go. It has good sound output, a host of connectivity options, and battery life that can get you through 2 movies with ease. It is well built too. On the downside, we wish the resolution were 720p at least and it recognised the audio format from some of our video files. 

Pros

Well built
Good audio output
Plenty of connectivity options

Cons

Not even 720p video output
Didn’t play certain MKV file audio formats

Best soundbars under Rs 5000 for enjoying the IPL on a budget

TVs today are getting slimmer and slimmer and when you are looking to pick up a TV to enjoy the IPL, especially on a budget, one of the biggest compromises is with the sound output of the TV. Most budget TVs today have two 10W speakers, which just about manage to get the job done. But if you want an immersive movie or sports watching experience on the TV, then you will have to invest in a soundbar. If you are on a tight budget, then worry not. We list the best soundbar that you can get for a budget of Rs 5000. Just so you know, we haven’t reviewed most of these soundbars. So we are listing them based on their features and specifications. 

Xiaomi Mi Soundbar – Buy Now

Xiaomi as a company has a reputation of disrupting the segment in which it launches a product. We have seen the company do this with smartphones and TVs. Xiaomi did the same with the launch of the Mi Soundbar. Put simply, the soundbar has very good sound reproduction for the price and has elegant looks as well. It has a good array of audio input options. The only downside is the absence of a remote control.

Portronics Sound Slick II POR-936 Wireless Bluetooth 40W Sound Bar – Buy Now 

The Portronics Sound Slick II boasts of 40W of power and the soundbar also comes with remote control. It has Bluetooth connectivity along with AUX-IN and USB. The soundbar weighs a mere 1.8 kgs and has a metal + plastic construction. It can be wall mounted or placed on a tabletop. The soundbar does not come with a subwoofer.

Philips DSP-475 U Soundbar Speaker – Buy Now 

This Philips soundbar comes with a subwoofer and remote control. It has a 5.2-inch subwoofer and 3-inch drivers in the soundbar. It has FM radio connectivity along with USB and 3.5mm connectivity. The product weighs 9kgs overall.

Creative Stage Air – Buy Now

The Creative Stage Air does not come with a subwoofer and has a 3.5mm along with USB input and Bluetooth connectivity. It can work as your first soundbar for your PC or a small TV. It also has physical controls on the side. It also works wirelessly giving you up to 6 hours of playback. The soundbar weighs 912 grams. 

CELESTECH Boom bar – Buy Now 

The CELESTECH boom bar has 2 10w speakers and works with TF Card, Aux In and Bluetooth connectivity options. It also has physical controls on the bar. It also has a 2000mAh battery for cord-free playback.

Best soundbars under Rs 10,000 that will take your IPL experience a notch higher

Rs 10,000 is a sweet spot for getting your first soundbar on a budget. It will enhance your IPL audio experience. With a budget of 10k, you aren’t spending too much and you can get devices with decent connectivity options along with good sound output. There are still some that come without a subwoofer at this price point. We haven’t tested all the soundbars in this list and have made it based on the specifications and features each one brings to the table. This list is in no particular order. 

Samsung HW-K350 – Buy Now

This soundbar from Samsung comes with a passive subwoofer and can deliver 150W of output. It gives users a 2.1-channel setup and also boasts of 5 sound modes. The soundbar also features Bluetooth connectivity. The subwoofer isn’t wireless.  

Blaupunkt SBW-02 – Buy Now

The Blaupunkt SBW-02 comes with a wired subwoofer in the box and can deliver 100W sound output. It has a host of connectivity options such as HDMI ARC, Optical In, USB, Bluetooth & AUX-In. The soundbar has multiple sound modes including Music, Movie, News and 3D. It also comes with a remote control in the box and features physical controls on the side of the soundbar. 

Xiaomi Mi Soundbar – Buy Now

If you are on a tight budget and want to spend a lot less than 10k, then you can consider the Xiaomi Mi Soundbar. The soundbar has very good sound reproduction for the price and has elegant looks as well. It has a good array of audio input options. The only downside is the absence of a remote control.

JBL Bar 2.0 – Buy Now

Just like the Mi Soundbar, the JBL 2.0 is a soundbar that doesn’t come with a subwoofer. It does, however, feature built-in dual bass port to make up for the lack of a subwoofer. When it comes to connectivity options, the soundbar has HDMI output (with Audio Return Channel), AUX input, Optical, and Bluetooth (Version 4.2). It also comes with a remote control in the box. It also has physical controls on the top of the soundbar.

Philips HTL1193B/94 – Buy Now

This 2.1 channel system from Philips comes with remote control in the box and also has a display on the front to show you the source. The soundbar has 80W of output. In terms of connectivity, it features coaxial in, digital optical in, USB and AUX-In along with Bluetooth connectivity. The display on the soundbar has an auto-dimming feature so it doesn’t hamper your TV viewing experience. 

Best soundbars under Rs 20,000 to enjoy the IPL

If you have a budget of sub 20k for a new soundbar, you’ll get a good set of options when it comes to features and performance of a soundbar. You can expect a host of connectivity options, a wireless subwoofer and even classy designs. Some even offer HDMI pass through. If you are one looking to upgrade the sound experience from your TV for this IPL season and have a budget of Rs 20,000 then this list is for you. Do keep in mind that we haven’t reviewed all the soundbars in this list and have made it based on features and specifications offered. This list is in no particular order

JBL 2.1 Soundbar – Buy Now

The JBL 2.1 is on the higher side of the price range but is a worthy one for those looking for a good sound experience from their TV. The subwoofer connects to the soundbar wirelessly and the package offers 300W of sound output. For connectivity options, the soundbar has HDMI input as well as ARRC output, 1 Analog input, 1 Optical, and Bluetooth (Version 4.2). Thanks to ARC, you can control your TV and the soundbar with one remote control. The soundbar also comes with a remote control in the box. 

Sony HT-CT290 – Buy Now

The Sony HT-CT290 brings with it 300W of power along with a wireless subwoofer. The soundbar can be wall mounted or kept on the table and the subwoofer can be kept horizontally or vertically. For connectivity, the soundbar supports Bluetooth, HDMI ARC, Optical, USB and AUX. Thanks to ARC, you can control the soundbar and your TV with the TVs remote control. You also get a remote control in the box.

LG SJ3 – Buy Now

The LG SJ3 brings with it 300W of power along with a wireless subwoofer. The soundbar can be wall mounted or kept on the table. It also comes with a bunch of sounds presets which can be changed based on the type of content you are consuming. For connectivity, the soundbar supports Bluetooth, Optical, USB and AUX. You get a remote control in the box but if you have an LG TV, you can control the soundbar with the LG TV remote control. 

Cambridge Audio TVB2 – Buy Now

The Cambridge Audio TVB2 comes with a wireless subwoofer and boasts of 120W of sound output. For connectivity, the soundbar features a host of connectivity options including Bluetooth, NFC, 3 HDMI inputs, 1 HDMI ARC Out, Optical In and Aux-in. A great thing is that the soundbar supports 4K passthrough which is great if you have a 4K TV. The soundbar also comes with a remote control in the box. 

JBL Cinema SB350 – Buy Now

Another JBL soundbar in this list is the SB350. It features a wireless subwoofer and 320W of power. For connectivity, the soundbar has One HDMI output with Audio Return Channel (ARC), Bluetooth, analogue input, and optical digital input. The soundbar can be wall mounted or kept on a table top. 

Best soundbars under 30,000 to experience the IPL

For Rs 30,000, you’ll get yourself a pretty good soundbar experience and also a host of connectivity options and a wireless subwoofer. Some of these soundbars also offer 4K passthrough along with ARC support. While there are some soundbars that come with rear surround speakers in this price range, but we have spoken about those in a separate article, which you can check out here.. We haven’t reviewed all the soundbars on this list and there is in no specific order. 

Yamaha YAS-207BL – Buy Now

Yamaha claims that this is the first soundbar with DTS Virtual:X. The soundbar boasts of 100W of output and has HDMI, optical and analogue connectivity options. Sadly there is no HDMI passthrough. 

JBL 3.1 Sound Bar – Buy Now

Priced slightly above the Rs 30,000 price tag, the JBL 3.1 has a wireless soundbar and brings with it 450W of power. It has a 10-inch subwoofer for the bass. In terms of connectivity options, the soundbar features  3 HDMI inputs and one HDMI ARC port. It also has 1 Analog, 1 Optical, and Bluetooth 4.2 connectivity options. Thanks to HDMI ARC, you can control the soundbar and your TV with one remote control. The soundbar also comes with a remote control in the box. 

Bose Solo 5 Soundbar – Buy Now

If a simple soundbar without a subwoofer is what you are looking for, then you can consider the Bose Solo 5 Soundbar. It comes with a universal remote control in the box and for connectivity has AUX, optical and coaxial input in addition to Bluetooth. It comes with an optical cable in the box. The soundbar also has a dialogue mode. 

Polk Audio AM9644-A Command Soundbar – Buy Now

If you are looking for a smart soundbar, then you can take a look at the soundbar from Polk Audio. The soundbar comes with a wireless subwoofer and supports Alexa. You can not only control the soundbar with Alexa but also control your smartphone appliances. For connectivity, the soundbar boasts of 2 HDMI ports with 4K passthrough and one HDMI ARC port. It also has a USB port and optical in. You can also control your Fire TV stick with the soundbar. 

Philips Fidelio B1/94 Nano Cinema Speaker – Buy Now

The Philips Fidelio comes with a wireless subwoofer and supports Bluetooth aptX and AAC. It also supports 1 HDMI in and HDMI out via ARC. The soundbar has a compact form factor and a display up front to show you the source.