研究左一下點玩好 Room EQ Correction
先由結果講起, 做 Room EQ Correction 個分別係好大
可以好大幅度修正喇叭特性, 彌補喇叭不足既地方, 較正房間/擺位問題

正題:
基於唔接受 Signal Path 有任何 AD 轉換, 但做 EQ Correction 一定要用到 DSP 既 Parametric EQ, 所以得出兩個方向

1. Hardware DSP (with AES or Coaxial Input/Output)
2. Software DSP

1 - Hardware DSP 通常一個 Bundle 搞店晒, 但價格貴

Trinnov:
https://www.trinnov.com/en/products/product-applications/pro-audio/

平 D 既有 MiniDSP
但入面 DSP sample rate 得 96kHz, 唔係好符合現代 Hi-Res 192kHz 標準趨勢

MiniDSP:
https://www.minidsp.com/

2 - Software DSP:
發現係性價比最高既方法, 雖然做法較為複雜
但勝在更強大既 DSP 功能, 調較精細度更高, 升級空間也更大
硬件需求門檻也更低 (只需測量咪, Audio Interface 就搞店)

原理:
Digital 訊號進入 Audio Interface
Audio Interface 經 USB 連接電腦, Windows/MAC OS 也可以
將訊號接入 VST Host/DAW軟件, 掛載 Parametric EQ plugins
人手輸入經測量的 Room EQ Correction 的較正數據
最後, 再經由 Audio Interface 將 Digital 訊號輸出

每個 Band 既 Frequency, Gain, Q, 什至乎 Slope 都可以詳細輸入

https://upload.review33.com/avforum/202407/202407100656249340.jpg

Setup Demo, 片中示範操作系統任何應用程式都可以用到 Room EQ Correction 功能:

可以玩盡對喇叭既表現
Room EQ Correction 呢方面, 大家又會點做 ?



最後修改時間: 2024-07-10 07:04:03

Great thanks for the insight
I do know room EQ is a very important (perhaps the most?) aspect to good sound.
Like you I do not want too much unnecessary processing D＞A by the DAC then A＞D by the DSP then D＞A again.
It always is something that I am looking out for but never had the 'mind openness' to try it.

Would love to hear your thoughts how much improve EQ has brought to your listening experience.

I think so
Room EQ not only corrects room acoustic problem, but also compensate the inaccuracy of the speakers
As a result, an accuracy and corrected sound wave is perceived at the listening position.

The improvement is much larger for implementing Room EQ Correction than swapping different brands of players, DACs, audio cables, etc (Some of them are really IQ taxes)

最後修改時間: 2024-07-10 09:03:52

Actually, for mid to hi-end speakers, I don't recommend using EQ to adjust the entire frequency range. Personally, I can only accept using EQ for frequencies below 500 Hz. The adjustments should be limited to identifying peaks at the listening position using REW and flattening them with EQ. It's not advisable to forcefully boost the dips.

I personally use Roon or Foobar/MathAudio Room EQ, especially MathAudio Room EQ, as I find its room correction method to be the most sensible and entirely digital. Of course, this approach is limited to streaming or file playback.

最後修改時間: 2024-07-10 09:43:44

You can use MathAudio Room EQ for your all software applications.
Just like me, I can use the room eq correction for listening music, watching movies or gamings, even using microphone for recording and monitoring, just everything.
All you need is a VST3 Host software or any DAW (e.g. Pro Tools, Cubase, Ableton Live, etc), but I think DAW is too heavy weight.
As MathAudio Room EQ has VST3 version plugins, you can use it for system audio output (ALL applications).

The way to do this is you might need an audio interface.
In my case, my system audio output goes to the audio interface (digital) virtual output.
Then, route the signal to the interface's loopback input directly.
using ASIO driver, the signal enters the VST3 host software, processed by the EQ plugins, is sent out to the interface's physical digital coaxial output for my DAC for decoding.

我今晚試左 MathAudio Room EQ 了...
效果沒有 REW + Fabfilter Pro-Q 3 理想

MathAudio Room EQ 做 Measurment 其實幾廢下....
精度也不夠...

兩個軟件 Gen 既 Curve
https://upload.review33.com/avforum/202407/202407110304298700.jpg

然後講 Correction, REW 得出既 Correction Data 好清楚, 要修正既 Frequency 都好清晰
再到 Correction, MathAudio 果條 Vertical Slide 叫 user 自己較, 自己聽咩係最自然...
我 diu 佢啦 ! 我依家要做較正, 你叫我自己聽.... on99 的 :o)

Parametric EQ 最緊要係較 frequecy, gain, bandwidth Q
MathAudio 佢就乜都無...
人地 Fabfilter 要較既都有得較
真係行內標準, 就係行內標準, 無得講

最後, 我要問 CHATGPT 4o 分析我 MATHAUDIO 幅圖, 點 Set 好....?!
人地 REW 成個 Predicted Target 圖整左出黎, 跟住 SET

最後修改時間: 2024-07-11 03:43:38

無得比

MathAudio 仲要賣100 鎂
智商稅, 留番 D 錢買 Fabfilter 好過佢

Wow manyung1987
You are really good at it.
I am afraid it is too complicated for me.

Good to know what software for me to explore further.

我嘅做法係咁；

1。數字EQ好過Analog（analog要好貴先得），無咁偏向一邊，我自己用Roon，已經覺得好好
2。先搞好器材，線材，擺位，防震… 到最後先用EQ做少少調整
3。用同幾首歌，先在一套參考音響和環境聽，嗰個感覺做標桿，然後係自己地方盡量還原

呢位樓主師兄
你一開口就屌
咁就唔使討論落去
呢個topic你玩埋佢啦

首先睇你條curve
你玩嘅嘢都唔方高級得去邊

第二你連條curve係點樣都未搞得清楚
從來條room curve都唔會係條間尺咁直嘅直線

等三foobar跟mathaudio係唔使畀錢
連點玩你都未搞清楚

仲有真係識玩hi-fi唔會信支咪
咪就得一個收音頭
向天或者向前
人嘅耳仔打橫向左向右仲有兩個耳殼

最後我話你知chatgpt問資料有幾可靠你自己睇啦




最後修改時間: 2024-07-11 08:15:04

#11

師兄你所講嘅三點
樓主唔會有一點聽得明
同圖表派講聲音係冇得講
加上佢所謂圖表派連半桶水都冇

最後修改時間: 2024-07-11 08:16:41

1. 首先, 標題已經注明係 "分享試玩", 唔係技術教學

2. 我 diu 既係 MATH AUDIO, 對事, 唔係對人

3. 玩音響從來係平有平玩, 貴有貴玩
如果套音響系統係完美, 就不用多花心機時間去做測量較正

EQ Correction 目的就係彌補器材, 擺位, 環境既不足
蛋糕唔食, 食面包 ?
搞清楚 EQ Correction 既意義先


4. 條 sine wave curve 現實當然做唔到簡尺咁直
直線係理論值, 參考目標


5. Mic 擺位做較正, 係盡量接近 Listening Position
好多鑑聽系統什至用 Binaural Mic

只係精度既問題
唔會影響做測量較正既原意, 作用


6. MathAuodio 是收費軟件, 100 美金

我花了不少時間去研究, 如何盡量做好 Room EQ Correct
比教用唔同軟件做 Room Correct 既分別, 知道兩種做法的不同
目的係同大家分享, 討論邊樣係值得玩
又可以慳時間金錢, 效果又唔錯
如果師兄有更好, 更抵玩, 或者更有效既方法
歡迎分享

好似你咁劈到一條直線咁
一個字講哂.....錯

#11

如果玩黑膠, 我會用 Analog EQ

但現在所有音訊媒體都是數碼格式, 所以最好是在未經 DA 前就做 EQ
否則, DA完再做 AD, 再做 DA,
好比 Video, 用 VGA 轉做 DVI, 再轉番做 VGA
中間仲要做 processing (EQ)
搞唔店, 我真係接受唔到

另外 Digital 有 Parametric EQ,
可以提供好多 band, 仔細咁調教每個參數



最後修改時間: 2024-07-11 22:42:31

另外介紹一套軟件都幾正

Sonarworks:

https://www.sonarworks.com/soundid-reference

我又唔覺得黑膠唔可以加 corr.，有新嘅 phono 設計係訊號經 LPF 直入 ADC 配合 DSP 可以選擇唔同年代或 label 嘅 RIAA 曲線而等效噪音及失真仲低啲。 我自己有時會都用 ADC 轉黑膠訊號入電腦做 room/phase corr. 嘅 real-time playback，但大多數都會選擇先將啲舊黑膠錄成 file 去做 corr。 咁樣可以減少唱針、黑膠嘅損耗、減少洗碟及其他保養嘅成本有利保存舊版碟、白版碟、派台碟。 錄之前先將隻碟洗得乾乾淨淨除埋靜電，可以針對唔同碟嘅厚度去校準 VTA，錄嗰時又可以避免喇叭發聲等震動影響音質，錄完仲可以慢慢執走嚴重嘅炒豆聲，用 digital post-processing 嘅修正又可以避免 analog EQ 嘅 phase shift，又可以修正埋唱針嘅曲線...等等。 仲可以用 binaural 錄低佢或將錄音 file 加上自己專屬嘅 room effect，喺外邊用耳機聽都仲可以有屋企聽歌嘅感覺... 。

Long long hing

咁多年 都有幻想如果一個唱盤 冇咗喇叭音壓嘅影響嘅聲音表現會有啲乜嘢唔同有幾多唔同

有冇一些估算嘅數據 可以分享一下呢?

理論上可以睇到兩個檔案嘅分別, 唔知有冇人做過你這個簡單嘅實驗有一些有趣嘅數據 1)table direct to DAC without speaker
2) playing the record, listening to it, at the se time feed to DAC from a tape loop etc..


有一些人可能有一些資源可以做一個 小小的獨立唱盤房間, 如果是值得的, ,


Thx Thx.

我估可以做個簡單實驗比較吓，唱針放嗰支碟個盤唔需要轉動，喇叭播放喱一個 source on/off 比較錄出嚟嘅 files 甚至播放錄咗嘅 files 聽吓 pick 咗幾多訊號就應該有個大概比較。