- #74 [長長影子], 18-07-16 12:07
- #73 [icefox_2001], 18-07-16 11:23
- #72 [r33_kevin], 18-07-12 18:51只講聽感,冇咩數據支持.同一隻sacd碟rip wav同dsd落cas聽,好多時wav係好聽過dsd.聽dsd個低音好似焗住左咁,wav反而冇野.唔知係唱片公司懶定慳錢冇將唔同format再做較聲定係其他原因.
最後修改時間: 2018-07-12 18:52:24 -
- #71 [Dynavector], 18-07-12 18:34Lossless, first and ever.
最後修改時間: 2018-07-12 18:50:01 - #70 [長長影子], 18-07-12 17:42咁有咩客觀證據證明 dsd 會好d呢?
Tidal MQA 係用邊個grade 呢?
MQA 都已經發表咗唔少量測數據作ref。
最後修改時間: 2018-07-12 17:57:17 - #69 [MLfans], 18-07-12 17:40Response to #66.
I have a Lumin A1 and listen to MQA files through Tidal. My experience so far is that MQA is not a sure guarantee of better sound quality.
If the sound quality of the original source is not very good then MQA cannot make it sound better. If the original source quality is very good then just standard CD quality streaming from Tidal sounds good. I have heard better sounding standard files than some MQA files from Tidal.
What I am saying is that it is the quality of the source (files or analogue tapes) that is the determining factor. There is no proof yet MQA sounds better than DSD or PCM files from the same source. It is just one of the many encoding algorithms, nothing more. - #68 [長長影子], 18-07-12 16:55傳統 PCM, DSD 格式被認為係 lossless,但其實只局限於保留整個重放系統中 A/D, D/A 之間嘅數據是否一致。 但實制應用上不同嘅錄/放器材本身已存在差異,當中 A/D、D/A 處理數/模訊號轉換部份由其關鍵,A/D、D/A 本身時/頻域表現差異亦足以影響音質嘅因素,但往往被用家所忽略。 唔同廠商生產嘅器材除咗可能使用唔規格嘅 A/D、D/A 外當中亦有使用唔同嘅 DSP 處理去達到自家牌子嘅音效或突顯某些技術規格,因此傳統 PCM, DSD 所謂嘅 lossless 並不足以保留理想中嘅原音重現。
有時候為了有好嘅錄放播放原音,用家只能花好多錢買昂貴嘅 A/D、D/A 器材,或花費買錄放監聽一樣嘅器材去追求原音,不過世上咁多唔同錄音室所用嘅器材又唔一樣。 不過其實即使係普通專業器材只要透過低成本但正確有效嘅數碼修正技術就可以有效降低器材之間嘅差異。 情況就好似用數碼相機影相,然後喺電腦 screen 睇返張相之後曬返出嚟。 如果相機、電腦 screen、曬相機本身冇獨立做校正。 相機 sensor、相機鏡頭、曬相機本身質素固然重要,但有做校正色準嘅器材先至可以有效降低器材顏色之間嘅偏差。
既然同一首音樂,即使用 PCM、DSD 格式去記錄,PCM、DSD file 數據量亦唔相同,重播出嚟都係唔同聲,咁兩種所謂 lossless 嘅 format 嘅 lossless 意義又何在呢? 當兩個鐘顯示嘅時間唔同,即係起碼其中一個鐘係唔準? 唔同格式固然有自身嘅優/缺點,但理應要殊途同歸。 雖然當年有份研發 DSD 嘅 James Angus 喺我以前讀電聲間 U 嘅老師,我有特別去留意佢發表嘅 paper,但以我個人所理解,DSD 嘅優點固然係能夠保留音樂嘅時域訊號特徵,但最大缺點係檔案大,又缺乏後期 DSP 處理彈性,宜家越來越多播放系統都不乏針對喇叭嘅 DSP 處理,轉換 DSD 過程中亦會產生不必要嘅干擾,而且 DSD out-band noise 亦會影響其他器材,又係影響音質嘅原兇。 PCM 要極高嘅取樣頻率先足以保留音樂嘅時域訊號特徵,錄放過程中唔同器材之間嘅 anti-aliasing filter 疊加後會劣化音樂嘅時域訊號特徵。 因為 MQA 就係針對以上嘅問題表現喺 PCM/DSD 中間嘅折衷方案,既能保留音樂嘅時域訊號特徵又降低所需嘅儲存數據量,music file 儲喺 NAS 裡面要一個大容量嘅 HDD/SSD 都冇 free lunch。 而且 MQA 可以縮窄錄/放器材之間嘅偏差,喱D係傳統 PCM、DSD 格式中所欠缺。 一開始都聲明 MQA 唔係 perfect 嘅 solution,每種格式亦各有優/缺點,綜合唔同 H/W、S/W、DSP 等等因素,MQA 有傳統 PCM、DSD 目前音質及應用上所缺乏嘅優點。
[圖] Envelope of spectrum for a closely recorded high piano note (G-sharp 72) and recording background noise are compared to hearing threshold and noise floors of four different channels: 1-bit 2.4224 MHz, 8-bit 192 kHz, 24-bit 96 kHz, and 24-bit 192 kHz. Acoustic gain is set to 120 dB SPL and retains level of the recording.
最後修改時間: 2018-07-12 17:03:41 - #67 [sinner], 18-07-12 16:17Sony also provide free firmware update for adding MQA support to their DAP too.
- #66 [wklie], 18-07-12 15:28#61:
Although it's true there is no free lunch, some of the license fee is already absorbed by the manufacturer of your network player - Lumin, since it does not charge a MQA upgrade fee.
Since Tidal does not charge more for MQA access than its lossless tier subscription fee, the license fee is absorbed by Tidal as well.
I'll explain where the demand comes from, and why it is important for a network player to support it. For people who don't care about streaming, this is irrelevant.
From a Tidal user perspective, often users find that they get better SQ from Tidal Master streaming than the equivalent Tidal lossless CD streaming, regardless of whether it comes from a better master or the technology. (There are opposite reports as well, but far fewer.) The only other Hi-Res streaming alternative - Qobuz is available in far fewer countries and costs more. This is the main incentive for Tidal users to give MQA a try, and the reason why many Roon users asked Roon to provide MQA decoding for more than a year, since Tidal is the only service that is integrated into Roon.
Whether a new format is necessary or not, and whether it sounds good or not, I stay neutral in these and leave such questions for the users to decide. To enable users to make such a judgement, from my perspective it is important for the playback equipment to support these. I hope people will be able to hear it for themselves and say it's good or bad based on what they heard, it's far more convincing than saying they think it's good or bad because they merely read some claims on the internet and never heard it with proper equipment. - #65 [MLfans], 18-07-12 13:52I always have this question in mind:-
"Does the world need another proprietary (read - not open standard and need to pay license fee) lossy compression algorithm, while there are no shortage of open standard lossless compression algorithm in DSD and PCM formats?"
IMHO we need to compare MQA to existing DSD and PCM open standard formats using the SAME recorded source. If MQA does not sound better and have significant advantages then what is the point of paying license fee to get lower sound quality??? - #64 [長長影子], 18-07-12 11:48世上的確係冇 free lunch,dolby, aac, aptx...etc,就算IC上用嘅 I2C interface,全部都要畀 licence free。
雞尾包入面冇雞尾,波蘿包入面冇波蘿,全部都係呃人嘅。 :-D
最後修改時間: 2018-07-12 11:59:10 - #63 [長長影子], 18-07-12 11:36#50 MQA 之前有 reply,我手上有全文,但要搵下網上有冇已公開嘅資料。
最後修改時間: 2018-07-12 11:39:33 - #62 [長長影子], 18-07-12 11:27個人喜好就話係 preference,但 codec performance 係可以透過 objective 方法直接去量測驗證,因此 analog analog 之間世界上永冇絶對 lossless,MQA 喺 AES 上發表數據相關理論嘅出處都有 reference,真係唔明邊方面係做假,始終做假係一個好嚴重嘅指控。 加上我網上所見嘅 impulse response 錄音測試等結果同理論係符合預期,如果單純以 A/D D/A signal chain time correction solution 去睇 MQA,雖則 MQA performance 唔係最佳但已經係最簡單易用,好適合大多數冇 DSP 技術背境嘅使用者,下一步都係要等 MQA codec 到手先至可以自己去量度。
- #61 [MLfans], 18-07-12 11:26Agree with ching on #60.
Just to clarify my understanding is that MQA declared that it is a lossy compression algorithm. I don't think they have claimed to be lossless.
MQA operates is a "black box" and claim that what comes out of it is as good as the original (claiming it to be better is pure lie) with substantial reduction in file size.
Please note that MQA is a proprietary compression algorism and they charge license fees for using it. The license fee will ultimately be paid by us who subscribe to Tidal or buy MQA files. There is NO FREE LUNCH.
From the business point of view MQA is a lock-in format so that they can continue to receive license fees.
Just my 2 cents worth.
Happy listening :)
最後修改時間: 2018-07-12 11:28:39 - #60 [MetalNuts], 18-07-12 10:10Whether MQA sounds better is a subjective opinion and depends on one's preference.
However, MQA has not given any proof of his technical claim so far and to the worse, it has been proved that many of its technical claims are in fact lies, such as lossless format. - #59 [長長影子], 18-07-12 01:24Listening environment, ~1.8m away from LSPKs, freq response 10Hz ~ 50kHz +/-6dB (1/3rd oct.) without equalisation, with time response & freq response + preference response equalized ~+/-3dB
最後修改時間: 2018-07-12 01:28:32 - #58 [長長影子], 18-07-12 01:18Listening environment reverberation time ~0.2s
- #57 [長長影子], 18-07-12 01:09Test bench
[49M] 44.1kHz 16-bit
最後修改時間: 2018-07-12 01:11:48 - #56 [長長影子], 18-07-12 01:00Test bench...
[742MB] 352kHz 24-bit
最後修改時間: 2018-07-12 01:11:04 - #55 [長長影子], 18-07-12 00:52Test bench...
[96MB]
最後修改時間: 2018-07-12 01:03:13 - 下一頁 (2 of 5)
- 返回 ...
https://www.audionet.com.tw/thread-10819-1-1.html
部份節錄如下:
"MQA Core: ...簡單說,當採用能對應MQA的「軟體」播放時(軟體解碼),能讓MQA音訊展開(unfold)第一層折疊,例如在電腦端採用Tidal HiFi、Audirvana或Roon這些支援MQA的軟體,就可以展開MQA的第一層折疊,其解析度最高不超過96kHz/24bit或88.2kHz/24bit。例如,你用Tidal HiFi播放一個原始編碼為192kHz/24bit的MQA檔案,最終播放出來的效果會是96kHz/24bit,而這種展開MQA第一層折疊之後輸出的音訊,也被稱為MQA Core。..."
"Full Decoder: ...如果是採用能對應MQA的「硬體」裝置播放(硬體解碼),除了能將音訊做第一層展開(Origami unfold to Core),接著還進行更多層的展開,和各種相關處理,最終展現出最完整的MQA效果,展現音訊的最佳解析度。例如,你採用具備高級MQA解碼能力的硬體裝置,播放一個原始編碼為352.8kHz/24bit的MQA檔案,最終播放出來的效果就會是352.8kHz/24bit,不打折,而此時就是MQA最佳的播放狀態,聲音質感也是最佳的。..."
"...MQA在解碼過程中還有一個特性值得一提,就是其解碼過程是一層層展開的,前一層展開後,會挾帶讓後一層展開的必要資訊,例如MQA Core展開之後就挾帶了必要資訊,能夠進行下一階段Full decoder或是MQA Renderer的完整解碼。
那是不是沒有達到Full decoder等級,就等於解碼不完全、玩半套,效果會不好呢?MQA的官方文件中有提到,不管是No decode、MQA Core、Full decoder所產生出的聲音效果,錄音師都可以在製作端進行個別檢視,以確保三種產出結果都是處於最佳狀態,也就是說,只是受到播放條件限制而有不同的聲音解析度,但效果都是能做到最佳化的,這是MQA官方的說法。..."
"...當以軟體來解碼時,只能達成一次性的MQA Core解碼,一定會有人想問:有沒有辦法透過軟體解碼一次就達成Full decoder狀態呢?Bob Stuart曾在受訪時提到:這是不可能的,因為MQA是一個完整的音樂訊號編碼和解碼的過程,它的開始和結束都必須是類比訊號(MQA is an analog to analog process.),要做到Full decoder必須知道最末端的DAC為何種型態,是否支援MQA解碼,因此,PC上的軟體解碼只能做到MQA Core階段的第一層展開,無法做到Full decoder的完全展開。..."
"...我們先放下音響迷的執著,不要計較「是否達成Full decoder的狀態」,先用一般的使用情境來看。試想一下,未來隨著網路頻寬增加、數位檔案價格下滑,Hi-Res檔案勢必越來越多,如果你取得了一個352.8kHz/24bit FLAC檔案(我們以2L的MAGNIFICAT測試檔案為例),它的大小為410MB,不僅佔空間,而且只要手邊的軟體或硬體不支援,就是不能播,所以取得之後,只能放在某些有空間、而且支援352.8kHz取樣率的設備上使用。現代人能播放音樂的管道很多,能取得的檔案解析度各式各樣,有的能播有的不能播,其實是很麻煩的。..."