My wife and daughters love to sing. A gift in the form of a karaoke system was simply asking for itself. But I wanted to not just “poorat”, but also to teach children something else. Therefore, immediately considered the option of karaoke with the assessment.
I did not find any available karaoke systems with an adequate score by grades, therefore I was born with such a homemade item.
Under the cat a lot of text and photos, videos, wiring diagrams, a little DIY, and a small guide to creating songs.
Spoiler - you can sing karaoke with points without shopping and handicraft, having only a computer with a microphone.
IdeaKaraoke classical systems in essenceThey are a combination of a video player and a microphone mixer. The video recorded illuminated letters with seascapes in the background, and the sound track to the video is the "minus" of the song. The signal from the microphone is mixed into the track being played and output to the speaker system.
Any computer can handle video playback; Moreover, for karaoke there is a special format karwhich takes up less space than video.
The basis of the karaoke system, therefore, canis a computer. In my case, an old but quite working Asus EEE 900 netbook was chosen. Of course, it also needs a speaker system and is highly desirable - a TV or a large monitor, or even better - a projector.
Mixing the signal from the microphone to the track turnedmore difficult. The simplest option seemed to use the microphone input of the computer, and the output connected to the amplifier of the speaker system. The system mixer can be configured so that the microphone signal is mixed into the music being played.
However, practice has shown that in this casethe microphone signal is noticeably late, which makes the idea unusable. Who ever heard an echo in telephone conversations, understands what it is. To speak, much less sing, with the echo is almost impossible.
As I understand it, the delay is due toimplementation of the sound paths of most "ordinary" PC - they are completely digital. That is, the signal from the input is digitized, then mixed in digital form with the reproduced sound, and only then is it converted to analog form and fed to the output. Since buffers are used during digitization, a signal delay occurs.
Without a hardware "supplement" in the karaoke system, it did not work out.
Hardware partA help came our Eastern brothers, who found such microphone amplifier board - mixer. Two microphones and a music source can be connected to this board, and a mixed signal is taken from its output, and without any delay due to the purely analog nature of the circuit. The fee was immediately ordered. Separately, I will say that now the seller is more expensive (I took for $ 12.87) and differs in photo from the one I ordered.
In appearance, the board was disappointing. A crooked soldering, different types of parts, and an abundance of Chinese fingerprints on the side of the tracks - this could be expected for $ 1-2-5, but this price is more expensive. However, she earned from the first inclusion.
Scheme and reflections on it
This is how the mixer circuit looks like:
The scheme is simplified - only one of the two channels of the microphone amplifier is shown, since it is identical to the second. Also only the left channel of the output stage is shown, and the power supply circuits are not shown.
Microphone preamps, timbre and final amplifier, assembled on dual operational amplifiers NE5532P. The board also contains an echo effect unit on the chip. M65831AP and analog timbrelock with separate adjustment of the bass and treble timbres.
The board is powered from a bipolar 12V source. To form the required voltages, stabilizers are installed. 7812 and 7912 for the organization of a bipolar power of 12 volts, as well as a stabilizer and 7805 5 volts to power the echo effect chip. Food patterns are typical and of no particular interest. It's funny that all stabilizers are labeled by different manufacturers - Motorola, ON Semiconductor and STMicroelectronics, although I suspect that it is still a fake.
To power the mixer, I used the purchased tricky transformer with double winding on 12V with a midpoint. I installed the transformer in the case from the power supply unit, combined with the mains plug. This made it possible to remove the transformer from the circuit and thereby reduce interference.
Improvements of the boardFor evaluation of points when singing onthe computer should be output a signal with the voice of the singer, without the "minus". The signal was fed to the microphone input of the netbook - from one microphone to the left channel, from the other to the right one.
It should be noted here that in modernNotebooks with a combined audio jack microphone input is mono (single-channel), and using some game modes that require separate signals from each of the microphones (competition, duet) will not work.
Board in the original version of the signal outputthe microphone does not have, therefore, required refinement. In fact, the refinement has been reduced to finding the point from which the signal was taken, and to soldering to this point. The point found by the method of scientific reasoning and testing is shown in the diagram. The dot is shown for one microphone - the input amplifiers of both of them are absolutely identical.
The place of the signal from the microphone
The signal was shot before all treatments (adjustmentvolume, timbre, echo effect). In an amicable way, it was necessary to use a buffer amplifier to eliminate the influence of the microphone input on the signal, and also to use an attenuator, since the microphone input level at this point is too high and the input is clearly overloaded: something like that is observed clipping. However, experience has shown that a direct connection is well suited for the required application, and overload has been very useful for singing with points.
After the first enthusiasm for the inclusion of the board dissipated, I noticed the echo effect. First, the echo was with pronounced repetitions, while I wanted something closer to reverbwhen you can not hear the individual "reflections". Also, when the echo level regulator was unscrewed to “zero”, the “ghosting” of the voice from the microphone was clearly traced.
The first problem was solved quite simply -The echo effect chip provides for the installation of delay by echo levels at four terminals. There are even places for jumpers on the board, but the required levels were still set by the divorced tracks. According to datasheet Initially, the delay was set to 160 ms, after the adjustment it became 61 ms.
To solve the second problem, I had to take a closer look at the strapping of the echo effect chip. If you simplify the scheme, then it can be represented as follows:
The diagram shows that the input signal forkstwo ways - the main one (shown in red) and the echo signal (blue). The echo signal is a chip-delayed input signal. After a delay, the signals are mixed.
Part of the mixed signal is fed back to the input of the delay chip, thus forming the “repetitions” characteristic of the echo.
But according to this scheme, if the echo level control is set to “zero”, then a delayed signal will still pass to the output! This was the reason for the unpleasant “doubling” of voices.
It was decided to alter the scheme as follows:
In this embodiment, the operation of the circuit remains the same, but with the echo off "repeats" in the output signal will not be.
The schematic diagram has changed as follows (changes are shown in red):
An additional 22 kΩ resistor is needed in orderto even out the adjustment of the depth of the echo effect. As a result, it still turned out to be imperfect - for the most part of the regulator stroke, the echo is not audible, it appears only at the end of the turn. Moreover, when twisting to the maximum, the attenuation becomes too small and the echo does not fade out, but on the contrary, increases. All this can also be solved by installing and selecting resistors in the feedback circuit, but I left it as it is. In most cases, the echo is not needed, and the effect with the amplification of children even like it, indulge in a "demonic sound".
One more revision was needed when it turned out that one of the minus channels played much quieter than the second. In this case, the signal from the microphones went to both channels with the same volume.
The problem was in the condenser 10 microfarad immediately after the volume control. "People" transistortester showed a capacity of 1 microfarad. Here it is, the hero:
Replacing a capacitor with a similar one decidedthe problem. The internal perfectionist also demanded to replace the analogous capacitor from the second, “normal” channel, and also to measure the remaining capacitors on the board. Yielding to persuasion, he did so, receiving a feeling of deep moral satisfaction (which, however, had no effect on the quality of the sound).
The board was placed in a housing made ofold cd drive. The giblets are pulled out of the drive, the body is shortened, the front and back panels are made of PCB. The label on the front panel is paper glued onto the textolite with a printout over which the film is glued. It looks collective, please do not judge strictly. But the goal is satisfied.
Photos of the assembled device and power supply
And still life:
Photo housing inside
Now I would have done a different way. In particular, the output of the microphone jacks above the volume control was a bad idea, as the plug inserted interferes with the adjustment. It is better to display under the elements or to the side of them. Display connectors on the rear panel is tempting, but dangerous: if you touch the cord, then all the good will fly after him. If the connectors are on the front panel, there is a chance that the cord will simply come off. For me, such a danger is relevant - a two-year-old son likes to make extra money with a dancer when the sisters sing.
The labels of the adjustment elements are also better placed on top so that the handles do not close them.
MicrophonePay provides connectiondynamic microphones. This is important because the inputs of PCs, smartphones and tablets are mostly designed for connecting condenser microphones. If you connect a microphone of the wrong type, for which the input is not designed, then the sound, if it does, will be very quiet or distorted. I did not have a dynamic microphone, so I was ordered such. Good or bad, I can not say, there is nothing to compare. In fact it suits me, the sound according to my feelings is quite clear.
The microphone is light, it sits in your hand well, comfortably. The cord is thick, but very flexible. The cord is heavy with unaccustomed, especially in comparison with the microphone itself, but it does not delay the hand. Nobody complained of severity, even 6-7 year old children.
Since I don't understand audio equipment fromthe words "absolutely", I will describe the microphone in simple words. The microphone is designed to “hear” sounds only near the microphone itself. At the same time, the directivity pattern is such that it is necessary to sing into the “top of the head” of the microphone, and not to the side. As I understand it, this is how all microphones of this type are arranged.
Despite the fact that the board has two inputs, I first decided to order only one microphone, since it was not clear whether a toy would “come in” or not.
Sound quality The system being developed is notMeans neither music lovers nor audiophiles. I doubt that it will give acceptable sound quality for these categories. From the narrow-minded point of view, the system does not introduce noticeable distortions into the sound. The only thing that was noticed was a slight background hiss that occurs when the supply voltage is applied to the board. But it is much weaker than music and voice, and is heard only in complete silence. If you do not pay special attention to it, then it is invisible.
This noise is introduced by the microphone amplifier,as it disappears when unscrewing the signal level from the microphones to zero. Unfortunately, I have no equipment for measuring the noise level, frequency response and the like, so I cannot give any quantitative values.
So just a few examples. Recorded on the phone through a simple adapter with a voltage divider.
“Silence” test - the first 5 seconds were recorded with the microphone turned off (the switch on the microphone itself was used), then the supply voltage was removed from the board: drive.google.com/open?id=16TSbTLNTQ17zeS-5TLgObyxZ1NqjBeMq
The program part of the highlight karaoke systemestimated score of singing in points. We already have a computer, a signal from a microphone is turned on to it, it remains to find software that can evaluate singing in points.
Short searches led to the game UltraStar Deluxe and her follower Vocaluxethat responded well to the goal. Games are made based on the Sony Playstation game and offers players to sing karaoke for points. A positive feature of them is that it displays a stylized “musical staff” and shows the key in which the player actually sings and whether he hits the notes or not right in the process of singing.
To analyze the player's singing, the signal from the microphone input is used - just the voice, without mixed music.
Choosing between the two listed, stoppedon UltraStar Deluxe, since Vocaluxe requires Windows 7, and on the used netbook, Windows XP was already installed and I did not want to install anything else. In addition, despite the lack of development, the latest versions of the game are quite stable and playable.
It turned out to be a big problem to find Russian songs for the game. A search on the Internet made it possible to download several dozens of options, but there were only a few of the normal ones to check.
It should be said here how these songs are organized. The basis is a text file, which contains the title, artist, genre, etc., links to the audio file, covers and background video, as well as information about words and notes. Details about the file format can be read, for example, here.
If the cover (dress up half-nakedperformers) and the lyrics are easy to correct, then the information about the words and notes is more difficult. In short, this information comes down to the fact that for a given position in the song, the next syllable and tone of the note corresponding to it are indicated. And if with the syllables in the found songs everything was more or less in order, then the tonalities often came across complete thrash. This was clearly seen (or rather, audibly) if you load the song into the editor and turn on the mode when the notes set in the file are played over the song. Even with my bearish ears, I obviously understood that the tone was not the same. I had to get into the fascinating world of creating and editing songs for these games.
Fascinating world with a small video tutorial
The game includes tools for creating songs (UltraStar Creator) and organizing them (UltraStar Manager); In addition, there are several third-party editors (SongEasy, Yass, UltraStarEditor). I liked Yass and UltraStarEditor the most. Though they are not trivial, they are quite comfortable and allow you to edit the song, unlike UltraStar Creator and SongEasy, which are aimed at quickly creating a song.
In these editors you can clearly place the syllablesin the right position in the song, set the pitch, listen to the result - in this mode notes sound over the track. It was in them that I was convinced that most of the Russian songs I found in the public domain have the wrong notes.
However, it is still quite laborious to bring the “curve” song to the norm or create a new one. It takes me 2-3 hours per song.
Creating a song is to arrange syllables innecessary positions and setting the key for syllables. Placing syllables is simple - where it is sung, there it is “drawn”, but with tonalities it is more difficult. Musicians will probably be able to easily pick up a melody by ear, but this is not my case. I acted in different ways.
At first, I found the midi-file of the desired song, and opened it with the midi-editor. I used Aria maestrosa and MidiEditor. Then I found a channel with the main melody andwatched it in Piano Roll mode. This is actually the same display that is used in the game itself and in the editors. Then it remained just (but not quickly) to arrange the necessary tonalities to the syllables.
It is important that the midi tonality coincides withthe tone of the melody used. You can use that midi converted to mp3 as a melody. I did just that for one of the songs, for conversion I used this service. However, even if the tonality does not match, it is sufficient to correctly determine the initial note, and the rest is simply proportional to shift.
Last songs I did in the editorUltraStarEditor, I use it now. In it, the spectrogram of the track is displayed on top of the “music staff”, and if you use a full-fledged song with a voice as a track, then with a certain dexterity you can determine the desired note height. But this does not always work - often on the spectrogram the voice merges with the instruments, and the performers are not robots and not musical instruments, and their voices do not always sound exactly on the notes.
Later I searched the Internet for notes of songs and relied on them. It is important that the notes correspond exactly to the leading voice, i.e. Guitar chords are not suitable. Such notes are best suited:
In the "translation" notes in the editor helped a lot This one picture.
Another option is to find a video tutorial on the piano of the desired song. It immediately shows which notes to use.
I repeat, those who are familiar with musical diplomaat the level of more than "heard", most likely it will be much easier. But even if you, like me, have avoided musical education, there are still many ways to create songs.
Although I stopped at the end of the UltraStarEditor,I can not say that the editor is perfect. There are glitches, which at one time pretty spoiled my life. I will list some of them; Perhaps someone will save hours spent on nerves:
- does not like the paths and file names in Russian
- after creating the song it is necessary to register the name and author, then it is better to save the song and restart the editor
- sometimes stops copying, pasting, reboot helps
- when copying blocks sometimes notes fall out, it is necessary to recheck.
I try to make songs in two versions -"Full" and minus one. As practice has shown, people are often embarrassed to sing under the backing track, and sing along to the performer completely. The game does not get worse.
To make my life easier, I try to findthe backing track is exactly the same tonality and tempo as the original song. Understand whether the backing track fits the easiest way to run them in parallel. I used the editor for this. Audacity.
The easiest way to make a song is with a voice - and the words are audible, and tonality. Then the alteration under the backing track becomes trivial - it’s enough to replace the track and maybe move all the syllables a little.
Finally, another important point. Before you make a song, it is highly recommended to set the desired tempo (BPM). You can learn the tempo with MixMeister BPM analyzer or another similar.
There are a lot of nuances in editing, but it is worth starting, and all this is not so scary already. I hope it will be a little clearer after watching this video.
The songs I created are in the public domain. here.
Bonus - collection of kar fileswhich, probably, is already 15 years old. I don’t remember where I got it, it was back in the days before the Internet. There are a LOT of songs, Russian and English.
It looks like this process of the game. Sings a child, not a professional.
The quality of voice recording in this video is unimportant - notI managed to adjust the levels and launched the first voice recorder, which, as it turned out, was recorded at a sampling frequency of 16 kHz. So what about the quality of the microphone and the board for this video, please do not judge. Above are examples of normal recording.
How to sing without "iron" Intrigue of this whole ideathat for singing with points it is possible without any purchases and improvements. You need to use a microphone that connects directly to the computer. If you are lucky, and the sound path of your computer transmits the signal from the microphone without delay, you will get exactly the same effect for free and without SMS needlework! Even if there is a sound delay, you can sing without transmitting the sound from the microphone to the acoustics, the game from this one never ceases to be interesting.
That is, it is enough to install the game, connect a microphone, download songs and have fun!
Thanks for attention! Fascinating singing to you!
I will be glad to questions, comments and constructive criticism.