Albert给出的解决方案，只要不经常调用所提供的函数，并且只用于大小适中的数组，就能很好地工作，否则效率非常低。下面是一个增强的普通JavaScript解决方案，它适用于Node和浏览器，具有以下优点:

•有效工作于所有八字节数组大小

•不生成中间丢弃字符串

•在现代JS引擎上支持4字节字符(否则“?”将被替换)

var utf8ArrayToStr = (function () {
    var charCache = new Array(128);  // Preallocate the cache for the common single byte chars
    var charFromCodePt = String.fromCodePoint || String.fromCharCode;
    var result = [];

    return function (array) {
        var codePt, byte1;
        var buffLen = array.length;

        result.length = 0;

        for (var i = 0; i < buffLen;) {
            byte1 = array[i++];

            if (byte1 <= 0x7F) {
                codePt = byte1;
            } else if (byte1 <= 0xDF) {
                codePt = ((byte1 & 0x1F) << 6) | (array[i++] & 0x3F);
            } else if (byte1 <= 0xEF) {
                codePt = ((byte1 & 0x0F) << 12) | ((array[i++] & 0x3F) << 6) | (array[i++] & 0x3F);
            } else if (String.fromCodePoint) {
                codePt = ((byte1 & 0x07) << 18) | ((array[i++] & 0x3F) << 12) | ((array[i++] & 0x3F) << 6) | (array[i++] & 0x3F);
            } else {
                codePt = 63;    // Cannot convert four byte code points, so use "?" instead
                i += 3;
            }

            result.push(charCache[codePt] || (charCache[codePt] = charFromCodePt(codePt)));
        }

        return result.join('');
    };
})();

在NodeJS中，我们有缓冲区可用，使用它们进行字符串转换非常容易。更好的是，它很容易将Uint8Array转换为Buffer。试试这段代码，它为我在节点基本上任何转换涉及Uint8Arrays:

let str = Buffer.from(uint8arr.buffer).toString();

我们只是从Uint8Array中提取ArrayBuffer，然后将其转换为适当的NodeJS Buffer。然后我们将Buffer转换为一个字符串(如果你愿意，你可以使用十六进制或base64编码)。

如果我们想从一个字符串转换回Uint8Array，那么我们会这样做:

let uint8arr = new Uint8Array(Buffer.from(str));

注意，如果你在转换为字符串时声明了一个像base64这样的编码，那么如果你使用base64或任何其他你使用的编码，你就必须使用Buffer.from(str， "base64")。

这将在没有模块的浏览器中不起作用!NodeJS的Buffer在浏览器中是不存在的，所以除非你在浏览器中添加Buffer功能，否则这个方法是行不通的。这其实很容易做到，只要使用像这样的模块，它既小又快!

来自Encoding标准的TextEncoder和TextDecoder，由stringencoding库填充，在字符串和ArrayBuffers之间转换:

var uint8array = new TextEncoder().encode("someString");
var string = new TextDecoder().decode(uint8array);

在节点“缓冲区实例也是Uint8Array实例”，所以buf.toString()在这种情况下工作。

我正在使用这个函数，它对我有用:

function uint8ArrayToBase64(data) {
    return btoa(Array.from(data).map((c) => String.fromCharCode(c)).join(''));
}

class UTF8{
static encode(str:string){return new UTF8().encode(str)}
static decode(data:Uint8Array){return new UTF8().decode(data)}

private EOF_byte:number = -1;
private EOF_code_point:number = -1;
private encoderError(code_point) {
    console.error("UTF8 encoderError",code_point)
}
private decoderError(fatal, opt_code_point?):number {
    if (fatal) console.error("UTF8 decoderError",opt_code_point)
    return opt_code_point || 0xFFFD;
}
private inRange(a:number, min:number, max:number) {
    return min <= a && a <= max;
}
private div(n:number, d:number) {
    return Math.floor(n / d);
}
private stringToCodePoints(string:string) {
    /** @type {Array.<number>} */
    let cps = [];
    // Based on http://www.w3.org/TR/WebIDL/#idl-DOMString
    let i = 0, n = string.length;
    while (i < string.length) {
        let c = string.charCodeAt(i);
        if (!this.inRange(c, 0xD800, 0xDFFF)) {
            cps.push(c);
        } else if (this.inRange(c, 0xDC00, 0xDFFF)) {
            cps.push(0xFFFD);
        } else { // (inRange(c, 0xD800, 0xDBFF))
            if (i == n - 1) {
                cps.push(0xFFFD);
            } else {
                let d = string.charCodeAt(i + 1);
                if (this.inRange(d, 0xDC00, 0xDFFF)) {
                    let a = c & 0x3FF;
                    let b = d & 0x3FF;
                    i += 1;
                    cps.push(0x10000 + (a << 10) + b);
                } else {
                    cps.push(0xFFFD);
                }
            }
        }
        i += 1;
    }
    return cps;
}

private encode(str:string):Uint8Array {
    let pos:number = 0;
    let codePoints = this.stringToCodePoints(str);
    let outputBytes = [];

    while (codePoints.length > pos) {
        let code_point:number = codePoints[pos++];

        if (this.inRange(code_point, 0xD800, 0xDFFF)) {
            this.encoderError(code_point);
        }
        else if (this.inRange(code_point, 0x0000, 0x007f)) {
            outputBytes.push(code_point);
        } else {
            let count = 0, offset = 0;
            if (this.inRange(code_point, 0x0080, 0x07FF)) {
                count = 1;
                offset = 0xC0;
            } else if (this.inRange(code_point, 0x0800, 0xFFFF)) {
                count = 2;
                offset = 0xE0;
            } else if (this.inRange(code_point, 0x10000, 0x10FFFF)) {
                count = 3;
                offset = 0xF0;
            }

            outputBytes.push(this.div(code_point, Math.pow(64, count)) + offset);

            while (count > 0) {
                let temp = this.div(code_point, Math.pow(64, count - 1));
                outputBytes.push(0x80 + (temp % 64));
                count -= 1;
            }
        }
    }
    return new Uint8Array(outputBytes);
}

private decode(data:Uint8Array):string {
    let fatal:boolean = false;
    let pos:number = 0;
    let result:string = "";
    let code_point:number;
    let utf8_code_point = 0;
    let utf8_bytes_needed = 0;
    let utf8_bytes_seen = 0;
    let utf8_lower_boundary = 0;

    while (data.length > pos) {
        let _byte = data[pos++];

        if (_byte == this.EOF_byte) {
            if (utf8_bytes_needed != 0) {
                code_point = this.decoderError(fatal);
            } else {
                code_point = this.EOF_code_point;
            }
        } else {
            if (utf8_bytes_needed == 0) {
                if (this.inRange(_byte, 0x00, 0x7F)) {
                    code_point = _byte;
                } else {
                    if (this.inRange(_byte, 0xC2, 0xDF)) {
                        utf8_bytes_needed = 1;
                        utf8_lower_boundary = 0x80;
                        utf8_code_point = _byte - 0xC0;
                    } else if (this.inRange(_byte, 0xE0, 0xEF)) {
                        utf8_bytes_needed = 2;
                        utf8_lower_boundary = 0x800;
                        utf8_code_point = _byte - 0xE0;
                    } else if (this.inRange(_byte, 0xF0, 0xF4)) {
                        utf8_bytes_needed = 3;
                        utf8_lower_boundary = 0x10000;
                        utf8_code_point = _byte - 0xF0;
                    } else {
                        this.decoderError(fatal);
                    }
                    utf8_code_point = utf8_code_point * Math.pow(64, utf8_bytes_needed);
                    code_point = null;
                }
            } else if (!this.inRange(_byte, 0x80, 0xBF)) {
                utf8_code_point = 0;
                utf8_bytes_needed = 0;
                utf8_bytes_seen = 0;
                utf8_lower_boundary = 0;
                pos--;
                code_point = this.decoderError(fatal, _byte);
            } else {
                utf8_bytes_seen += 1;
                utf8_code_point = utf8_code_point + (_byte - 0x80) * Math.pow(64, utf8_bytes_needed - utf8_bytes_seen);

                if (utf8_bytes_seen !== utf8_bytes_needed) {
                    code_point = null;
                } else {
                    let cp = utf8_code_point;
                    let lower_boundary = utf8_lower_boundary;
                    utf8_code_point = 0;
                    utf8_bytes_needed = 0;
                    utf8_bytes_seen = 0;
                    utf8_lower_boundary = 0;
                    if (this.inRange(cp, lower_boundary, 0x10FFFF) && !this.inRange(cp, 0xD800, 0xDFFF)) {
                        code_point = cp;
                    } else {
                        code_point = this.decoderError(fatal, _byte);
                    }
                }

            }
        }
        //Decode string
        if (code_point !== null && code_point !== this.EOF_code_point) {
            if (code_point <= 0xFFFF) {
                if (code_point > 0)result += String.fromCharCode(code_point);
            } else {
                code_point -= 0x10000;
                result += String.fromCharCode(0xD800 + ((code_point >> 10) & 0x3ff));
                result += String.fromCharCode(0xDC00 + (code_point & 0x3ff));
            }
        }
    }
    return result;
}

`