- 第 3 页 -
原帖 由 ddyykk_hello 于 09-11-26 23:09 发表
对于一个平均电平只有-20dB的混音作品来进行母带，提升很多的响度难道不会因为此而使音质下降呢
就好像原来的信号只有很小，我一下子gain了20db，不管底噪的问题，音质本身不会变化吗
找個10多軌工程看看, 如果軌道的平均電平大約在 -18/-20 dbfs , mastering buss 平均大概是 -10dbfs, peak 可以到 -6dbfs
, 這樣的聲音不應有問題..., 就算單軌 -20/-18dbfs 也不應有問題...要 安心一點的話, 錄音時可使用 24bit...
0vu / +4 / 1.23 的標準由來是低噪在 -75db 或以下時, 音量足以抵消人類對噪音相對的感覺....而硬件亦針對這個指標來設計, 所以在這音量工作最為理想, 。 Paul 在後面的貼建議用插件如壓縮, 音量可以在 -6 到-10 dbfs 之間...
如果因個別條件因素, 最便宜的方法當然是提高一點在錄音時的音量.....
要記著, 這是一個基本的指引而不是一個絕對的值, 只要在這概念下不要跑得太遠就已經是合理, 以及能達到一個較皆的效果...
正如第一貼所說, 如果 console 以 現在daw一般錯誤的概念( 以接近 0 dbfs) 訊號來混音, 聲音一樣糟糕。
...
[ 本帖最后由 himhui 于 12-2-18 14:01 编辑 ]
再來一段對話關高頻...........
Dean Roddey :
The high frequency thing seems to match my (admittedly limited but pretty heavily exploratory) experience. To get a sound that matches the kind of mixes I like, it seems like I'm cranking down the highs on lots of tracks. Not necessarily a low pass, sometimes it's just removing some of the presence frequencies, and leaving some of the air. Sometimes it's a low pass.
Particularly if you want to create any sort of sense of depth in the mix, you can't have things that are supposed to be behind the speakers all bright and sparkly. If all that high end is captured and it never gets lost, no many how many times you 'run the bits over the record head', they have to be removed manually sometimes.
Obviously compression does this job to some extent as well, since in some cases like drums the highest frequency content is often in that initial attack that will be reduced by any fast attack compression.
Paul Frindle :
The point you make here about freq content versus mix depth is exactly correct and very very important. :-)
In the natural world, stuff that is further away tends to have less HF content because of distance effects.
So getting depth into a production made from close mic'ed or direct sources does involve judicious EQ in the mixing process - remembering that the term 'equaliser' was actually invented by the early film industry for a device that equalised for distance effects. That was it's very purpose :-)
In the analogue world, losses here and there in the signal chain could often have the effect of providing a degree of this HF loss during production, multiple passes of tape, bouncing down etc etc. could sometimes effectively put the 'backing into the background'..
But in the lossless 'what you hear is what you get - always and forever' digital world no such effect happens by itself. So it's necessary for people to be fully aware of the art of mixing and apply it themselves :-)
One of the historical issues (now much less important as W/S get more powerful processing) is that to get this kind of stuff right you need to have EQ simultaneously running on many tracks while doing the mix as a whole - so you can tweak them against each other to get the correct depth impression.
Whilst simultaneous EQ on all channels was the norm in analogue console - in the early processing strapped digital world such a luxury was not available. So this fundamental art of mixing was either lost or never learnt
[ 本帖最后由 himhui 于 09-11-27 17:23 编辑 ]
Dean Roddey :
The high frequency thing seems to match my (admittedly limited but pretty heavily exploratory) experience. To get a sound that matches the kind of mixes I like, it seems like I'm cranking down the highs on lots of tracks. Not necessarily a low pass, sometimes it's just removing some of the presence frequencies, and leaving some of the air. Sometimes it's a low pass.
Particularly if you want to create any sort of sense of depth in the mix, you can't have things that are supposed to be behind the speakers all bright and sparkly. If all that high end is captured and it never gets lost, no many how many times you 'run the bits over the record head', they have to be removed manually sometimes.
Obviously compression does this job to some extent as well, since in some cases like drums the highest frequency content is often in that initial attack that will be reduced by any fast attack compression.
Paul Frindle :
The point you make here about freq content versus mix depth is exactly correct and very very important. :-)
In the natural world, stuff that is further away tends to have less HF content because of distance effects.
So getting depth into a production made from close mic'ed or direct sources does involve judicious EQ in the mixing process - remembering that the term 'equaliser' was actually invented by the early film industry for a device that equalised for distance effects. That was it's very purpose :-)
In the analogue world, losses here and there in the signal chain could often have the effect of providing a degree of this HF loss during production, multiple passes of tape, bouncing down etc etc. could sometimes effectively put the 'backing into the background'..
But in the lossless 'what you hear is what you get - always and forever' digital world no such effect happens by itself. So it's necessary for people to be fully aware of the art of mixing and apply it themselves :-)
One of the historical issues (now much less important as W/S get more powerful processing) is that to get this kind of stuff right you need to have EQ simultaneously running on many tracks while doing the mix as a whole - so you can tweak them against each other to get the correct depth impression.
Whilst simultaneous EQ on all channels was the norm in analogue console - in the early processing strapped digital world such a luxury was not available. So this fundamental art of mixing was either lost or never learnt
[ 本帖最后由 himhui 于 09-11-27 17:23 编辑 ]
我想瞭解的是，爲什麽要讓每個軌道的電平都在-10dBFS左右（DAW中）。
我們知道在模擬調音台上這樣做是爲了避免高電壓下的失真，這是模擬元器件特性所致。
那麼在DAW中，數字信號只有到0dBFS才會真正失真，而沒有像模擬設備一樣的失真過程。
且DAW main buss上的動態範圍一般來說足夠，在保證各軌及母線皆不失真的情況下，按理說就不應該出現信號劣化。
模擬summing是電流的summing，電壓越大，噪音越大，失真亦越大，所以控制各軌的電壓是非常必要的事。
所以我想知道，到底是什麽讓DAW中的聲音變”數字“的，因為如果所有的電平表都沒clip，按理說就沒有失真。
而且對於DAW的電壓summing，儘量保證大電平是降低噪音一個最主要的手段。
大家繼續討論。
我們知道在模擬調音台上這樣做是爲了避免高電壓下的失真，這是模擬元器件特性所致。
那麼在DAW中，數字信號只有到0dBFS才會真正失真，而沒有像模擬設備一樣的失真過程。
且DAW main buss上的動態範圍一般來說足夠，在保證各軌及母線皆不失真的情況下，按理說就不應該出現信號劣化。
模擬summing是電流的summing，電壓越大，噪音越大，失真亦越大，所以控制各軌的電壓是非常必要的事。
所以我想知道，到底是什麽讓DAW中的聲音變”數字“的，因為如果所有的電平表都沒clip，按理說就沒有失真。
而且對於DAW的電壓summing，儘量保證大電平是降低噪音一個最主要的手段。
大家繼續討論。
另外，個人覺得聲音劣化的重災區其實在A to D的過程中
不要以為在DAW中0dBFS是失真點，AD轉換也0dB就是失真點，那就大錯特錯了。
很多ADDA芯片的性能其實在還沒到18dBu時就已經開始劣化，不同的生產商又做不同的電路優化，但基礎沒變。
所以錄音的時候，在高電壓狀態下，進入DAW的信號就已經失真了。
另外，所有的數字設備最大的問題便是高頻失真，普通的ADDA高頻失真更是嚴重。
這便是我認為Analoge味道丟失的最主要環節。
所以我的看法是：錄音時在進入DAW前就讓電平 控制 在-3dBFS以下，
Mixing時在保證各軌不失真及母線不失真的情況下（不是用limit去限制），儘量提升各軌電平，以保證母線的動態範圍被完全利用。
這樣mastering提升響度后也能有較低的底噪。
[ 本帖最后由 margindge 于 09-11-27 22:48 编辑 ]
不要以為在DAW中0dBFS是失真點，AD轉換也0dB就是失真點，那就大錯特錯了。
很多ADDA芯片的性能其實在還沒到18dBu時就已經開始劣化，不同的生產商又做不同的電路優化，但基礎沒變。
所以錄音的時候，在高電壓狀態下，進入DAW的信號就已經失真了。
另外，所有的數字設備最大的問題便是高頻失真，普通的ADDA高頻失真更是嚴重。
這便是我認為Analoge味道丟失的最主要環節。
所以我的看法是：錄音時在進入DAW前就讓電平 控制 在-3dBFS以下，
Mixing時在保證各軌不失真及母線不失真的情況下（不是用limit去限制），儘量提升各軌電平，以保證母線的動態範圍被完全利用。
這樣mastering提升響度后也能有較低的底噪。
[ 本帖最后由 margindge 于 09-11-27 22:48 编辑 ]
原帖 由 margindge 于 09-11-27 22:17 发表
我想瞭解的是，爲什麽要讓每個軌道的電平都在-10dBFS左右（DAW中）。
所以我想知道，到底是什麽讓DAW中的聲音變”數字“的，因為如果所有的電平表都沒clip，按理說就沒有失真。
而且對於DAW的電壓summing，儘量保證大電平是降低噪音一個最主要的手段。...
ＰＡＵＬ 上面說的正好解釋了其中疑問
In the analogue world, losses here and there in the signal chain could often have the effect of providing a degree of this HF loss during production, multiple passes of tape, bouncing down etc etc. could sometimes effectively put the 'backing into the background'..
But in the lossless 'what you hear is what you get - always and forever' digital world no such effect happens by itself. So it's necessary for people to be fully aware of the art of mixing and apply it themselves :-)
在模擬世界, 訊號鏈的過程中, 多次的磁帶往返、軌道合拚等等. 都會導致高頻失掉 (其實也是失真的一種), 但很多時候卻讓一些背景的聲音更往後靠去, 而數字世界本身就沒有這樣的事情發生, 在沒有人為下的高頻本已比模擬偏多, 再加以不恰當的處理就會更糟,...
[ 本帖最后由 himhui 于 09-12-9 10:35 编辑 ]
有一點我也想說說， 就是這個 ０ｖｕ／＋４／１．２３ 或 －１８ｄｂｆｓ 這個建議直應該是看成是一個起點， 進入效果器處理再出來時應該還要選擇聲音的ｓｗｅｅｔ ｓｐｏｔ （理想點）， 出來的聲音就不一定硬要停留在這個值。
在混音時，一開始我會在先平衡各通遁音量， 盡量保持 ｍａｓｔｅｒ ｂｕｓｓ的推子在０ｄｂ位置而電平不過載，逼不得矣才會在 ｍａｓｔｅｒ ｂｕｓｓ 拉下一點． 這樣有個好處， 就是可以用最少自動化來達到最終的平衡，而我再翻看我過往一年的工程， 其實各軌道的電平也跟這個建議值不謀而合， 至於處理低噪在數字年代比模擬年代要輕鬆得多。
．．
[ 本帖最后由 himhui 于 09-11-28 01:36 编辑 ]
在混音時，一開始我會在先平衡各通遁音量， 盡量保持 ｍａｓｔｅｒ ｂｕｓｓ的推子在０ｄｂ位置而電平不過載，逼不得矣才會在 ｍａｓｔｅｒ ｂｕｓｓ 拉下一點． 這樣有個好處， 就是可以用最少自動化來達到最終的平衡，而我再翻看我過往一年的工程， 其實各軌道的電平也跟這個建議值不謀而合， 至於處理低噪在數字年代比模擬年代要輕鬆得多。
．．
[ 本帖最后由 himhui 于 09-11-28 01:36 编辑 ]
看了一下,这个讨论很有意义.但感觉有个重点没有强调,作者原话里也没很好的区分开来,就是电平的PEAK值和RMS值,以及电平表工作在这两者中的哪个状态.许多疑惑也出在这个上面了.举个例子(呵呵,不是故意挑刺呀)"电脑的电平表有一个重大的缺陷就是0电平是在最高处的，其实应该是在-20dB的地方的，就和老式的vu表一样，这样，新手就不会盲目追求大电平。" 象这个电脑的电平表0电平是在最高处没错,因为这是峰值电平,而老式VU表是平均电平.此贴大家的讨论中,相当多的地方都没区分开PEAK和RMS,造成很多混乱.
顺带说下,0vu / +4 / 1.23 的标准源头,应该是来自最早的有线电话线路. 如有错误欢迎指正.
顺带说下,0vu / +4 / 1.23 的标准源头,应该是来自最早的有线电话线路. 如有错误欢迎指正.
在DAW中 RMS的意義在於控制歌曲響度的標準，不像模擬時代的VU表直接控制信號是否失真。在DAW中，PPM才是控制失真的唯一標準，但在AD設備上並不是。
說以我覺得HIMHUI老師說的很對，錄音的時候就要控制進來的數字電平，要找到sweet spot。mixing的時候只要保證各軌的PPM不爆即可，然而如果保證了main buss的PPM不爆，實際上各軌的電平就已經蠻小的了。如果需要，可以在各個軌道上加壓縮后再Gain，但一定要保證Main buss的PPM不爆。
[ 本帖最后由 margindge 于 09-11-28 10:26 编辑 ]
說以我覺得HIMHUI老師說的很對，錄音的時候就要控制進來的數字電平，要找到sweet spot。mixing的時候只要保證各軌的PPM不爆即可，然而如果保證了main buss的PPM不爆，實際上各軌的電平就已經蠻小的了。如果需要，可以在各個軌道上加壓縮后再Gain，但一定要保證Main buss的PPM不爆。
[ 本帖最后由 margindge 于 09-11-28 10:26 编辑 ]