 |
|  What sample rate/bit depth do you use? [MUSIC] | |
|
 |
minomus
  Member
Topics: 7
Replies: 251
Registered: 16.Nov.05 |
I meant that VSTis are having problems earlier with 44.1kHz than 48kHz.. it's obvious that they have problems earlier than Reason  Maybe I said it a bit unclearly...
2GB RAM, shouldn't be a problem..
http://www.myspace.com/andypiney
Upcoming release: Valve - Farlight (Original & Andy Piney remix) [Neuroscience Recordings] |
|
EE Jeffrey
Member
Topics: 14
Replies: 583
Registered: 01.Feb.06 |
 |
We (E/E) track and mix in renoise at 48khz, sometimes go back to 44.1 to prevent stuttering when cpu load is too high.
Render wav at 48khz 24bit, sometimes 96khz 24bit but usually this screws things up with certain (buggy?) vst, so 48khz is usually the way to go for us.
Then I take the render post production in Acid, mastering it at the rendered sample rate/bit depth, but afterwards I render the master from Acid to 44.1 16bit wav, before I encode to mp3. Which is what I'm doing only as of lately.
I used to stick with the 48khz or 96khz / 24bit and directly render it to mp3 from Acid. Actually all E/E tracks here on CTG where rendered directly. However, with our latest 'Heart Wired' I found the result better if I give lame a 44.1khz/16bit wav. Only 44.1/16 and only that. Other input in lame goes wrong like Acid. I don't know why, but this just seems to perform better for me. Didn't took the time to research what is happening and what I like better about it 
Easily Embarrassed Check out our Darkened Emotion mix yay yay! |
|
Atlantis
Member
Topics: 84
Replies: 3227
Registered: 14.Jan.03 |
|
minomus wrote on 10 Dec. (13:05) :
48kHz/24-bit from the beginning till the end. So, no dithering at all. 24-bit makes it slightly easier for mp3 encoder to make a good quality mp3 since lower noise level, and as I said in the other topic, 48kHz keeps full sounding, crispy trance slightly brighter and better in high frequencies. |
Are you sure that's a good idea? MP3 is 16 bit (just convert an MP3 to WAV, or use a bit meter on an MP3 file encoded from a 24 bit WAV), and so you're truncating the lower 8 bits, introducing harmonic distortion. It's better to dither to 16 bit first. Strangely enough, I remember reading on the Studio-Central forums some time ago to use 24 bit, but all my current knowledge points to the fact that this is just plain wrong - MP3 is 16 bit, so dither to 16 bit.
As for using 48 kHz, I agree, but I personally find 48 kHz annoying as a final sample rate, especially considering it would almost always end up at 44.1 kHz (CD) anyway. But of course it's good to process higher, and you could always still resample just before dithering, though I still question whether using 48 kHz in this sense is worth it. But of course if it's all your equipment can handle...
My mind is drifting... |
|
Atlantis
Member
Topics: 84
Replies: 3227
Registered: 14.Jan.03 |
|
Nekativ wrote on 11 Dec. (9:11) :
Then I take the render post production in Acid, mastering it at the rendered sample rate/bit depth, but afterwards I render the master from Acid to 44.1 16bit wav, before I encode to mp3. Which is what I'm doing only as of lately. |
Are you using a dithering plugin before this, or is ACID set to dither the output? If not, you're truncating the (possibly) 32 bit float output, and ending up with only 16 bits, resulting in harmonic distortion. It's important you dither the output first, using the Waves L3 or iZotope Ozone, etc. on the master channel. Read all about dither here. Also, I doubt ACID applies an anti-alias filter before resampling, so you're still ending up with aliasing, and all that processing at 48 kHz probably wasn't worth it.
|
I used to stick with the 48khz or 96khz / 24bit and directly render it to mp3 from Acid. Actually all E/E tracks here on CTG where rendered directly. However, with our latest 'Heart Wired' I found the result better if I give lame a 44.1khz/16bit wav. Only 44.1/16 and only that. Other input in lame goes wrong like Acid. I don't know why, but this just seems to perform better for me. Didn't took the time to research what is happening and what I like better about it |
Again, that's truncating, and should result in the same MP3 output if ACID indeed doesn't apply dither. And if the 44.1/16 file is dithered, it should sound better. I don't know much about sample rates and MP3 though, but I thought it only supported up to 48 kHz.
My mind is drifting... |
|
Atlantis
Member
Topics: 84
Replies: 3227
Registered: 14.Jan.03 |
All those answering 24 bit, do you use dithered 24 or truncated 24?
By the way, I was hoping for this to become an educational topic, so it's not like I'm bashing everyone or anything.
My mind is drifting... |
|
EE Jeffrey
Member
Topics: 14
Replies: 583
Registered: 01.Feb.06 |
|
Atlantis wrote on 11 Dec. (10:59) :
Are you using a dithering plugin before this, or is ACID set to dither the output? If not, you're truncating the (possibly) 32 bit float output, and ending up with only 16 bits, resulting in harmonic distortion. It's important you dither the output first, using the Waves L3 or iZotope Ozone, etc. on the master channel. Read all about dither here. Also, I doubt ACID applies an anti-alias filter before resampling, so you're still ending up with aliasing, and all that processing at 48 kHz probably wasn't worth it. |
I use dithering when rendering from both Acid and also from Renoise ya. Well I don't know what it is all about (will read that wikipedia article you mentioned), but I did know it was to help when converting between sample rates, so I normally have it enabled when rendering.
Is it still a waste then to render and processing at 48khz when I go with 44.1 in the very end? Just curious about that, because my renoise renders at 48khz do seem to sound better in some way. Less dull or something. I rather take that post production in Acid to master than the 44.1khz. 48khz master seems way better, even if I downgrade to 44.1 in the end after mastering is done and compare to doing it with 44.1 all the way.
Just sharing what I experience myself, don't know what is true though What I hear may be clouded by what I believe I want to hear or so 
As for mp3 stuff. I did search some stuff up now. Acid uses Lame encoder too btw, so if I render an acid project directly to mp3 that is on 48khz/24, my guess is Acid will use the 48khz for input in Lame I guess.
I googled some stuff about Lame encoding and found a lot of forums claiming 44.1 will sound good with lame, while 48 may not, as Lame's presets (which I use when doing it myself and Acid probably uses too) are more tuned for 44.1 it seems
Example the chatter on the subject I found by google search in here :
http://www.hydrogenaudio.org/forums/index.php?showtopic=9105
I guess that's why I think rendering from Acid to 44.1/16 wav and then encoding it with Lame sounds better then directly encoding it from Acid where I hear a (minor) difference between the mp3 and when playing the acid project or listening to a wav render. I do am wondering now if I tried encoding 44.1/24bit with Lame. Can't remember that one. I guess it would be better to try and keep using 24bit all the way? Or not? Well, I guess I will check it out with next project anyway.
Easily Embarrassed Check out our Darkened Emotion mix yay yay! |
|
minomus
Member
Topics: 7
Replies: 251
Registered: 16.Nov.05 |
|
Atlantis wrote on 11 Dec. (10:11) :
minomus wrote on 10 Dec. (13:05) :
48kHz/24-bit from the beginning till the end. So, no dithering at all. 24-bit makes it slightly easier for mp3 encoder to make a good quality mp3 since lower noise level, and as I said in the other topic, 48kHz keeps full sounding, crispy trance slightly brighter and better in high frequencies. |
Are you sure that's a good idea? MP3 is 16 bit (just convert an MP3 to WAV, or use a bit meter on an MP3 file encoded from a 24 bit WAV), and so you're truncating the lower 8 bits, introducing harmonic distortion. It's better to dither to 16 bit first. Strangely enough, I remember reading on the Studio-Central forums some time ago to use 24 bit, but all my current knowledge points to the fact that this is just plain wrong - MP3 is 16 bit, so dither to 16 bit.
As for using 48 kHz, I agree, but I personally find 48 kHz annoying as a final sample rate, especially considering it would almost always end up at 44.1 kHz (CD) anyway. But of course it's good to process higher, and you could always still resample just before dithering, though I still question whether using 48 kHz in this sense is worth it. But of course if it's all your equipment can handle...
My mind is drifting... |
Well, there has been a debate about 16-bit vs. 24-bit for encoding and I've been told that it would be better because it would provide less noise to encoder (especially good when using ABR with lame)...? even though the final result is 16-bit, but it would get better source somehow.. I don't have any facts for this to prove it.
Anyways, I'll continue using 24-bit 48kHz in the future, too, because it has been working fine for me. Why to change something which works? 
Btw, just installed 2496, preamp and sennheiser hd-600! My hearing and therefore mixing and mastering abilities just raised one level now when I actually can trust what is comin' to my ears
http://www.myspace.com/andypiney
Upcoming release: Valve - Farlight (Original & Andy Piney remix) [Neuroscience Recordings] |
|
peter.not
Member
Topics: 22
Replies: 53
Registered: 20.Jan.05 |
minomus wrote on 10 Dec. (13:05) :
48kHz/24-bit from the beginning till the end. So, no dithering at all. 24-bit makes it slightly easier for mp3 encoder to make a good quality mp3 since lower noise level, and as I said in the other topic, 48kHz keeps full sounding, crispy trance slightly brighter and better in high frequencies.
http://www.myspace.com/andypiney
Upcoming release: Valve - Farlight (Original & Andy Piney remix) [Neuroscience Recordings] |
you can hear the difference?? you bat !!  |
|
EE Jeffrey
Member
Topics: 14
Replies: 583
Registered: 01.Feb.06 |
minomus wrote on 11 Dec. (13:31) :
Anyways, I'll continue using 24-bit 48kHz in the future, too, because it has been working fine for me. Why to change something which works? |
Can always try and see what happens? I think changing stuff that already works, maybe might work even better Never know if you don't try
Easily Embarrassed Check out our Darkened Emotion mix yay yay! |
|
Atlantis
Member
Topics: 84
Replies: 3227
Registered: 14.Jan.03 |
|
Nekativ wrote on 11 Dec. (11:57) :
Atlantis wrote on 11 Dec. (10:59) :
Are you using a dithering plugin before this, or is ACID set to dither the output? If not, you're truncating the (possibly) 32 bit float output, and ending up with only 16 bits, resulting in harmonic distortion. It's important you dither the output first, using the Waves L3 or iZotope Ozone, etc. on the master channel. Read all about dither here. Also, I doubt ACID applies an anti-alias filter before resampling, so you're still ending up with aliasing, and all that processing at 48 kHz probably wasn't worth it. |
I use dithering when rendering from both Acid and also from Renoise ya. Well I don't know what it is all about (will read that wikipedia article you mentioned), but I did know it was to help when converting between sample rates, so I normally have it enabled when rendering. |
Renoise will dither to 16 bit if enabled, so even if you render in 32 bit float (I thought the dither disables then?), you're only rendering in 16 bit (dithered), and the lower bits are just 0's. Disable the dither, and render in 32 bit float. Dither doesn't have anything to do with the sample rate from my understanding, though.
Is it still a waste then to render and processing at 48khz when I go with 44.1 in the very end? Just curious about that, because my renoise renders at 48khz do seem to sound better in some way. Less dull or something. I rather take that post production in Acid to master than the 44.1khz. 48khz master seems way better, even if I downgrade to 44.1 in the end after mastering is done and compare to doing it with 44.1 all the way.
Just sharing what I experience myself, don't know what is true though What I hear may be clouded by what I believe I want to hear or so |
It's not a waste if you resample down to 44.1 kHz with an anti-alias filter. When rendering at a sample rate of 44,100 Hz, any frequencies above the Nyquist frequency (22,050 Hz) will be reflected onto the lower frequencies to produce alias components. So even when you render at 48 kHz, resampling to 44.1 kHz without first lowpassing off these high frequencies with an anti-alias filter will result in aliasing, as with rendering directly at 44,100 Hz.
|
As for mp3 stuff. I did search some stuff up now. Acid uses Lame encoder too btw, so if I render an acid project directly to mp3 that is on 48khz/24, my guess is Acid will use the 48khz for input in Lame I guess. |
That makes sense, but the bit depth will still be truncated to 16 bit, so you end up with harmonic distortion, unless you have a dithering plugin at the end of your mastering chain, or you have ACID set to dither the output.
I googled some stuff about Lame encoding and found a lot of forums claiming 44.1 will sound good with lame, while 48 may not, as Lame's presets (which I use when doing it myself and Acid probably uses too) are more tuned for 44.1 it seems
Example the chatter on the subject I found by google search in here :
http://www.hydrogenaudio.org/forums/index.php?showtopic=9105
I guess that's why I think rendering from Acid to 44.1/16 wav and then encoding it with Lame sounds better then directly encoding it from Acid where I hear a (minor) difference between the mp3 and when playing the acid project or listening to a wav render. I do am wondering now if I tried encoding 44.1/24bit with Lame. Can't remember that one. I guess it would be better to try and keep using 24bit all the way? Or not? Well, I guess I will check it out with next project anyway. |
Well I don't know enough about MP3, but that could be so.
What you want to do however is render from Renoise in 32 bit float and at the highest sample rate that works without giving problems (it's not really a bug, but some plugins just don't have support for high sample rates).
Then master this in ACID (I really recommend an audio editor though), resample it either by rendering at the new sample rate and at the highest bit depth possible (but bypass the dither, since you'll want to import the file again and dither it afterwards, before saving it in 16 bit), or, render at the highest bit depth and sample rate as possible, then resample it in Sound Forge (possible even upsampling it to 64 bit float first), apply dither, and save it as 44,100 Hz, 16 bit.
I hope that wasn't too confusing, but ACID just isn't a program you want to be resampling in. There are audio editors for that and programs like Voxengo r8brain.
My mind is drifting... |
|
|
|
|
|
Latest Topics
![[80.220.112.222]](pics/ip.gif){kind=small}
