Forget about discrete audio cards. Integrated is enough for everyone. Sound card diagnostics Why the built-in sound is not very good
There was a time when a question of necessity sound card didn't stand at all. If you need a sound in your computer that is a little better than the grunting of the speaker in the case, buy a sound card. If you don't need it, don't buy it. However, the cards were quite expensive, especially while they were being made for the prehistoric ISA port.
With the transition to PCI, it became possible to shift part of the calculations to CPU and also use RAM for storing music samples (in ancient times, this need was not only for professional musicians, but also for normal people, because the most popular music format on computers 20 years ago was MIDI). So soon sound cards entry level dropped a lot in price, and then built-in sound appeared in the top motherboards Oh. It's bad, of course, but it's free. And this dealt a severe blow to sound card manufacturers.
Today, absolutely all motherboards have built-in sound. And in expensive ones it is even positioned as high quality. That's straight up Hi-Fi. But in reality, unfortunately, this is far from the case. Last year I collected new computer, where I installed one of the most expensive and objectively best motherboards. And, of course, they promised high-quality sound on discrete chips, and even with gold-plated connectors. They wrote it so well that I decided not to install a sound card and make do with the built-in one. And he got by. About a week. Then I disassembled the case, installed the card and didn’t bother with any more nonsense.
Why is the built-in sound not very good?
Firstly, the issue of price. A decent sound card costs 5-6 thousand rubles. And it’s not a matter of manufacturers’ greed, it’s just that the components are not cheap, and the requirements for build quality are high. A serious motherboard costs 15-20 thousand rubles. Is the manufacturer ready to add at least three thousand more? Will the user get scared without having time to evaluate the sound quality? It's better not to take risks. And they don't take risks.
Secondly, for truly high-quality sound, without extraneous noise, interference and distortion, the components must be located at a certain distance from each other. If you look at the sound card, you will see how unusually much free space there is on it. But on the motherboard there is only enough space for it, everything has to be placed very tightly. And, alas, there is simply nowhere to do it really well.
Twenty years ago, consumer sound cards cost more than a computer, and they had memory slots (!) for storing music samples. In the photo, the dream of all computer geeks in the mid-nineties is Sound Blaster AWE 32. 32 is not a bit depth, but maximum amount simultaneously playing streams in MIDI
Therefore, integrated sound is always a compromise. I have seen boards with seemingly built-in sound, which, in fact, hovered from above in the form of a separate platform connected to the “mother” only by a connector. And yes, it sounded good. But can such sound be called integrated? Not sure.
A reader who has not tried discrete sound solutions may have a question: what exactly does “good sound in a computer” mean?
1) He's simply louder. Even a budget-level sound card has a built-in amplifier that can “pump” even large speakers or high-impedance headphones. Many people are surprised that the speakers stop wheezing and choking at maximum. This is also a side effect of a normal amplifier.
2) The frequencies complement each other, and do not mix, turning into mush. A normal digital-to-analog converter (DAC) well “draws” the bass, mids and highs, allowing you to very accurately customize them using software to suit your own taste. When listening to music, you will suddenly hear each instrument separately. And the films will delight you with the effect of presence. In general, the impression is as if the speakers were previously covered with a thick blanket, and then it was removed.
3) The difference is especially noticeable in games.. You'll be surprised that the sound of the wind and dripping water doesn't drown out the quiet footsteps of your opponents around the corner. That in headphones, not necessarily expensive ones, there is an understanding of who is moving, where from and at what distance. This directly affects performance. It simply won’t be possible to sneak up/drive up to you on the sly.
What kind of sound cards are there?
When did this type of component become of interest only to connoisseurs? good sound, of which, unfortunately, there are very few, there are very few manufacturers left. There are only two – Asus and Creative. The latter is generally a mastodon of the market, having created it and set all the standards. Asus entered it relatively late, but it still hasn’t left.
New models are released extremely rarely, and old ones are sold for a long time, 5-6 years. The fact is that in terms of sound you can’t improve anything there without a radical increase in price. And few people are willing to pay for audiophile perversions in a computer. I would say no one is ready. The quality bar is already set too high.
The first difference is the interface. There are cards that are intended only for desktop computers, and they are installed into the motherboard via the PCI-Express interface. Others connect via USB and can be used with both large computers and laptops. The latter, by the way, have disgusting sound in 90% of cases, and an upgrade certainly wouldn’t hurt it.
The second difference is the price. If we're talking about internal maps, then for 2-2.5 thousand Models are sold that are almost similar to built-in sound. They are usually purchased in cases where the connector on the motherboard has died (alas, a common phenomenon). An unpleasant feature of cheap cards is their low resistance to interference. If you place them close to the video card, background sounds will be very annoying.
The golden mean for built-in maps is 5-6 thousand rubles. It already has everything to please a normal person: interference protection, high-quality components and flexible software.
Behind 8-10 thousand The latest models are sold that can reproduce 32-bit sound in the 384 kHz range. This is right here top top. If you know where to get files and games in this quality, be sure to buy them :)
Even more expensive sound cards differ little in hardware from the already mentioned options, but they acquire additional equipment - external modules for connecting devices, companion boards with outputs for professional sound recording, etc. It depends on the actual needs of the user. Personally, I have never needed the body kit, although in the store it seemed like it was needed.
For USB cards, the price range is approximately the same: from 2 thousand alternative to built-in sound, 5-7 thousand strong middle peasants, 8-10 high end and beyond that everything is the same, but with a rich body kit.
Personally, I stop hearing the difference at the golden mean. Simply because cooler solutions also require hi-fi speakers and headphones, and to be honest, I don’t see much point in playing World of Tanks with thousand-dollar headphones. Probably, every problem has its own solutions.
Several good options
Several sound cards and adapters that I tried and liked.
PCI-Express interface
Creative Sound Blaster Z. It's been on sale for 6 years now, it costs about the same on different computers, and I'm still very happy with it. The CS4398 DAC used in this product is old, but audiophiles compare its sound to CD players in the $500 range. The average price is 5500 rubles.
Asus Strix Soar. If everything in the Creative product is shamelessly geared towards games, then Asus has also taken care of music lovers. The ESS SABRE9006A DAC is comparable in sound to the CS4398, but Asus offers more fine tuning parameters for those who like to listen to Pink Floyd on their computer in HD quality. The price is comparable, about 5500 rubles.
USB interface
Asus Xonar U3– a small box, when inserted into a laptop port, takes the sound quality in it to a new level. Despite the compact dimensions, there was even room for a digital output. And the software is simply surprisingly flexible. An interesting option to try is why you need a sound card at all. Price 2000 rubles.
Creative Sound BlasterX G5. The device is the size of a pack of cigarettes (smoking is evil) and its characteristics are almost indistinguishable from the internal Sound Blaster Z, but there is no need to climb anywhere, just plug the plug into the USB port. And immediately you have seven-channel sound of impeccable quality, all sorts of gadgets for music and games, as well as a built-in USB port just in case you don't have enough of them. Having space made it possible to add an additional headphone amplifier, and once you hear it in action, it’s hard to get out of the habit. The main functions of the software are duplicated by hardware buttons. The issue price is 10 thousand rubles.
Play and listen to music with pleasure! There are not so many of them, these pleasures.
A sound wave represents areas of high and low pressure that are perceived by our hearing organs. These waves can travel through solid, liquid and gaseous media. This means that they easily pass through the human body. Theoretically, if the pressure of the sound wave is too high, it could kill a person.
Any sound wave has its own specific frequency. The human ear is capable of hearing sound waves with frequencies ranging from 20 to 20,000 Hz. The level of sound intensity can be expressed in dB (decibels). For example, the intensity level of the sound of a jackhammer is 120 dB - a person standing next to you will not receive the most pleasant sensation from a terrible roar in the ears. But if we sit in front of a speaker playing at a frequency of 19 Hz and set the sound intensity to 120 dB, we will not hear anything. But sound waves and vibrations will all affect us. And after a while you will begin to experience various visions and see phantoms. The thing is that 19 Hz is the resonant frequency for our eyeball.
This is interesting: Scientists learned that 19 Hz is the resonant frequency for our eyeball under rather interesting circumstances. American astronauts, when ascending into orbit, complained of periodic visions. Detailed studies of the phenomenon have shown that the frequency of operation of the engines of the first stage of the rocket coincides with the frequency of operation of the human eyeball. At the required intensity of sound, strange visions arise.
Sound with a frequency below 20 Hz is called infrasound. Infrasound can be extremely dangerous for living beings, since organs in the human and animal bodies operate at infrasound frequencies. The superposition of certain infrasound frequencies on top of each other with the required sound intensity will cause disruptions in the functioning of the heart, vision, nervous system or brain. For example, when rats are exposed to 8 Hz infrasound, 120 dB causes brain damage. [wiki]. When the intensity increases to 180 dB and the frequency remains at 8 Hz, the person will not feel the best - breathing will slow down and become intermittent. Prolonged exposure to such sound waves will cause death.
This is interesting: The record for the loudest car sound system belongs to two engineers from Brazil - Richard Clarke and David Navone, who managed to install a subwoofer in the car with a theoretical sound volume of 180 dB. Needless to say, this system should not be used to its full potential?
During testing, the subwoofer, driven by electric motors and a crankshaft, reached a sound intensity of 168 dB and broke down. After this incident, they decided not to repair the system.
Sounds belong to the section of phonetics. The study of sounds is included in any school curriculum in the Russian language. Familiarization with sounds and their basic characteristics occurs in the lower grades. A more detailed study of sounds with complex examples and nuances takes place in middle and high school. This page provides only basic knowledge according to the sounds of the Russian language in a compressed form. If you need to study the structure of the speech apparatus, the tonality of sounds, articulation, acoustic components and other aspects that go beyond the scope of the modern school curriculum, refer to specialized manuals and textbooks on phonetics.
What is sound?
Sound, like words and sentences, is the basic unit of language. However, the sound does not express any meaning, but reflects the sound of the word. Thanks to this, we distinguish words from each other. Words differ in the number of sounds (port - sport, crow - funnel), a set of sounds (lemon - estuary, cat - mouse), a sequence of sounds (nose - sleep, bush - knock) up to complete mismatch of sounds (boat - speedboat, forest - park).
What sounds are there?
In Russian, sounds are divided into vowels and consonants. The Russian language has 33 letters and 42 sounds: 6 vowels, 36 consonants, 2 letters (ь, ъ) do not indicate a sound. The discrepancy in the number of letters and sounds (not counting b and b) is caused by the fact that for 10 vowel letters there are 6 sounds, for 21 consonant letters there are 36 sounds (if we take into account all combinations of consonant sounds: deaf/voiced, soft/hard). On the letter, the sound is indicated in square brackets.
There are no sounds: [e], [e], [yu], [ya], [b], [b], [zh'], [sh'], [ts'], [th], [h] , [sch].
Scheme 1. Letters and sounds of the Russian language.
How are sounds pronounced?
We pronounce sounds when exhaling (only in the case of the interjection “a-a-a”, expressing fear, the sound is pronounced when inhaling.). The division of sounds into vowels and consonants is related to how a person pronounces them. Vowel sounds are pronounced by the voice due to exhaled air passing through tense vocal cords and freely exiting through the mouth. Consonant sounds consist of noise or a combination of voice and noise due to the fact that the exhaled air encounters an obstacle in its path in the form of a bow or teeth. Vowel sounds are pronounced loudly, consonant sounds are pronounced muffled. A person is able to sing vowel sounds with his voice (exhaled air), raising or lowering the timbre. Consonant sounds cannot be sung; they are pronounced equally muffled. Hard and soft signs do not represent sounds. They cannot be pronounced as an independent sound. When pronouncing a word, they influence the consonant in front of them, making it soft or hard.
Transcription of the word
Transcription of a word is a recording of the sounds in a word, that is, actually a recording of how the word is correctly pronounced. Sounds are enclosed in square brackets. Compare: a - letter, [a] - sound. The softness of consonants is indicated by an apostrophe: p - letter, [p] - hard sound, [p’] - soft sound. Voiced and voiceless consonants are not indicated in writing in any way. The transcription of the word is written in square brackets. Examples: door → [dv’er’], thorn → [kal’uch’ka]. Sometimes the transcription indicates stress - an apostrophe before the stressed vowel.
There is no clear comparison of letters and sounds. In the Russian language there are many cases of substitution of vowel sounds depending on the place of stress of the word, substitution of consonants or loss of consonant sounds in certain combinations. When compiling a transcription of a word, the rules of phonetics are taken into account.
Color scheme
In phonetic analysis, words are sometimes drawn with color schemes: letters are painted in different colors depending on what sound they represent. The colors reflect the phonetic characteristics of sounds and help you visualize how a word is pronounced and what sounds it consists of.
All vowels (stressed and unstressed) are marked with a red background. Iotated vowels are marked green-red: green means the soft consonant sound [й‘], red means the vowel that follows it. Consonants with hard sounds are colored blue. Consonants with soft sounds are colored green. Soft and hard signs are painted gray or not painted at all.
Designations:
- vowel, - iotated, - hard consonant, - soft consonant, - soft or hard consonant.
Note. The blue-green color is not used in phonetic analysis diagrams, since a consonant sound cannot be soft and hard at the same time. The blue-green color in the table above is only used to demonstrate that the sound can be either soft or hard.
Question: is it worth buying a sound card if it has a built-in sound card?Your question needs to be divided into two categories: hardware and software and the actual sound quality.
There is optical drive. If the transmission is through optics, there is a difference with
built-in sound card, or from a separate, cool sound card?
1. Hardware and software:
If we are not talking about built-in software codecs of the AC97 and HDaudio standard, then a sound card in a PC is necessary mainly for the implementation of numerous sound algorithms like EAX (from Creative, for example), which add realism, volume, take into account the real-time characteristics of the visual environment and correct the corresponding them sound parameters. For example, you are walking along a corridor in some horror story and the sound corresponds to the characteristics of reflection from concrete walls, it literally walks and is tangible. Then go out into a large hall and immediately the reverberation changes, the equalization characteristics shift, etc. and so on. This is not as noticeable as the visual effects, but in games with high quality soundtrack adds a significant amount of drama. Specialized gaming audio cards process all these effects at the hardware level using chips like EMU10K, EMU20K, etc., freeing the CPU from additional calculations of effects. If the game engine does not detect such a device on your PC, then it sets up a simplified sound effects scheme, which may not differ in actual parameters from EAX, or may be much inferior to it. It’s up to you to decide whether it’s necessary, although you can output sound in games via the audio device, and music via an external USB DAC, by switching in the audio device manager or directly in the software player (some have this option);
2. Sound quality. Modern top-end (and expensive) GAMING sound cards (there is also a category of professional sound cards such as those produced by LYNX, M-AUDIO, etc.) in principle, sound on musical material at the level of cheap ones external USB DACs. To some extent, they are saved by ASIO drivers, if there are any for your sound card model, which allow the audio stream to bypass the Windows software meat grinder (Asio4all is a software crutch that does not solve this problem). As for sound output via outdated optical interfaces SPDIF (Sonny Philips interface), TOSLINK (Toshiba Link), etc., their only advantage is the limited and completeness of any options. How would it be more correct to describe this: “You can buy an advanced food processor with a bunch of gadgets and adjustments for using which you need at least an understanding of the process, or you can load everything into one cup and press one button, where the knives will chop your vegetables into a certain guaranteed mass, but you can immediately forget about all sorts of neat “cubes” and “straws.” In fact, these interfaces are a standard connection option that guarantees that the digital stream will reach the DAC, and the amount of losses “along the way” will be minimized. This type of connection has been used for decades, all possible problems have been solved a long time ago and in general it is simpler and cheaper to implement. With a DAC of an outdated design or in a DAC where the manufacturer saved money on a high-quality USB receiver, this type of connection sometimes shows best result. But there is a very big BUT: the speed of these optical interfaces is very limited and we can’t even talk about any DSD or serious high-resolution (usually the speed is limited to 24 bit 48 kHz). USB connection has many implementation possibilities, this is a topic for a large separate article; on a PC with Windows OS it requires at least an understanding of the process and some user actions to programmatically configure the PC-USB DAC interface to provide the so-called. bit-to-bit transmission quality (some DACs even have a special indication of confirmation that this transmission mode has been achieved). It is important which USB receiver is installed in the DAC and the number of “dropouts” of digital fragments along the way depends on it. The trick is that it is the audio stream via USB that is transmitted in the outdated PCM format, which completely lacks such advanced features as data transfer by transaction, transfer of checksums of data packets, etc., and therefore in this case it makes sense as in high-quality USB receivers, as well as high-quality cables, methods for implementing data transfer (for example, top-end motherboards have specialized USB outputs for connecting to external DACs, in which the +5 Volt power supply line is DISCONNECTED, and the range of the logical zero and one signal is increased (in fact, zero and one in USB differ only in voltage)). As for DAC chips specifically, they should be the last thing you pay attention to! It doesn’t matter whether your device has a cheap Wolfson WM8741 or a top-end chip from Asahi Kasei, what is important first of all is the implementation and environment, which characterize 90% of the final sound. When they write about cool DACs and that “cheap” A produces a pathetic signal-to-noise ratio of 107 dB, and advanced DAC B produces as much as 120 dB, it becomes funny, since in most digital masters everything that lies below the 40 dB level is simply castrated ! Those. there is no musical information at all in this area. Of course, this does not apply to high-quality high-records made from analog media on high-quality hardware with direct hands, but you still need to look for such ones. Specifically, the Cambridge CXA80 is a worthy device, sounding in the usual intelligent “British manner” (although this is a misconception and the so-called “British sound” is also a lot and very different), implying in the general understanding timbral accuracy, as close as possible to the sound of the original, good spatial characteristics , provided by high-quality circuitry, acceptable dynamic and rhythmic indicators. Cambridge and Arcam are such “all-time” all-rounders, which may not evoke a storm of emotions with each soundtrack, but will provide listening pleasure. The USB DAC in this amp is built on the WM8740 chip, which 10-15 years ago was one of the most popular and received a lot of good reviews(IMHO well-deserved) due to neutrality, lack of digital sharpness, and besides, in this amplifier it is implemented at least like a human being, and not like a poor relative who is only invited to a funeral. Those. in a setup based on this amplifier, it is quite suitable for connection and is adequate to the level of the equipment. If you want more emotions and drive, less versatility, look to the Atoll 100SE. It doesn't have a DAC, phono preamp, or tone controls, but for the price it's one of the best sounding amps on the market. You can look for YBA - also excellent devices. Again, there are worthy competitors in the form of Rega Elex, Naim 5si (I would recommend Micromega, but the price for them now is just kind of crazy). In short, the choice is quite extensive. From the "Japs" you can pay attention to the good Denon 1520.
Let's figure out whether it's worth buying discrete or external sound cards. For Mac and Win platforms.
We often write about quality sound. In a portable wrapper, but we avoid desktop interfaces. Why?
Stationary home acoustics - subject creepy holivars. Especially when using computers as a sound source.
Most users of any PC consider a discrete or external audio card the key to high-quality sound. It’s all the fault of “conscientious” marketing, persistently convincing us of the need to purchase an additional device.
What is used in a PC to output an audio stream?
The built-in sound of modern motherboards and laptops significantly exceeds the auditory analysis capabilities of the average mentally healthy, technically literate listener. The platform doesn't matter.
Some motherboards have enough high-quality integrated sound. Moreover, they are based on the same means as in budget boards. Improvement is achieved by separating the sound part from other elements and using a higher quality element base. 
And yet, most boards use the same codec from Realtek. Desktop computers Apple is no exception. At least a decent portion of them are equipped Realtek A8xx.
This codec (a set of logic enclosed in a chip) and its modifications are typical for almost all motherboards designed for Intel processors. Marketers call it Intel HD Audio.
Realtek audio quality measurements
The implementation of audio interfaces largely depends on the motherboard manufacturer. High-quality specimens show very good figures. For example, RMAA test for audio path Gigabyte G33M-DS2R:
Frequency response unevenness (from 40 Hz to 15 kHz), dB: +0.01, -0.09
Noise level, dB (A): -92.5
Dynamic range, dB (A): 91.8
Harmonic distortion, %: 0.0022
Intermodulation distortion + noise, %: 0.012
Interpenetration of channels, dB: -91.9
Intermodulation at 10 kHz, %: 0.0075
All the figures obtained deserve the ratings “Very Good” and “Excellent”. Not every external card can show such results.
Comparison test results
Unfortunately, time and equipment do not allow us to conduct our own comparative testing of various built-in and external solutions.
Therefore, let's take what has already been done for us. On the Internet, for example, you can find data on double internal resampling of the most popular discrete cards in the series Creative X-Fi. Since they relate to circuitry, we’ll leave the check on your shoulders.
Here are the materials published one large hardware project allow us to understand many things. In testing of several systems from the built-in codec for 2 dollars before the audiophile decision for 2000, very interesting results were obtained.
It turned out that Realtek ALC889 does not show the smoothest frequency response, and gives a decent tone difference - 1.4 dB at 100 Hz. True, in reality this figure is not critical. 
And in some implementations (that is, motherboard models) it is completely absent - see the figure above. It can only be noticed when listening to one frequency. In a musical composition, after correct settings equalizer, even an avid audiophile will not be able to tell the difference between a discrete card and an integrated solution.
Expert opinion
In all of our blind tests, we were unable to detect any differences between 44.1 and 176.4 kHz or 16 and 24-bit recordings. Based on our experience, 16 bit/44.1 kHz ratio provides best quality sound that you can feel. The above formats simply waste space and money.
Downsampling a track from 176.4 kHz to 44.1 kHz using a high-quality resampler prevents loss of detail. If you get your hands on such a recording, change the frequency to 44.1 kHz and enjoy.
The main advantage of 24-bit over 16-bit is greater dynamic range (144 dB versus 98), but this is practically insignificant. Many modern tracks are in a battle for loudness, in which the dynamic range is artificially reduced at the production stage to 8-10 bits.
My card doesn't sound good. What to do?
All this is very convincing. During my time working with hardware, I managed to test a lot of devices - desktop and portable. Despite this, I use a computer with built-in chip Realtek.
What if the sound has artifacts and problems? Follow the instructions:
1) Disable all effects in the control panel, put on the green hole “ line output” in “2 channels (stereo)” mode.
2) In the OS mixer, turn off all unnecessary inputs, and set the volume sliders to maximum. Adjustments should only be made using the regulator on the speaker/amplifier.
3) Install the correct player. For Windows - foobar2000.
4) In it we set “Kernel Streaming Output” (you need to download an additional plugin), 24 bits, software resampling (via PPHS or SSRC) at 48 kHz. For output we use WASAPI Output. Turn off the volume control.
Everything else is the work of your audio system (speakers or headphones). After all, a sound card is, first of all, a DAC.
What's the result?
The reality is that, in general, a discrete card does not provide a significant gain in the quality of music playback (this is at a minimum). Its advantages lie only in convenience, functionality, and, perhaps, stability.
Why do all publications still recommend expensive solutions? Simple psychology - people believe that to change the quality of work computer system need to buy something advanced, expensive. In fact, you need to put your head to everything. And the result can be surprising.