|
Exporting from Ableton (pg. 3)
|
View this Thread in Original format
| Kismet7 |
| quote: | Originally posted by echosystm
why, because the bit resolution loss on 32bit float audio is going to be so big? :wtf:
if we were talking 16bit fixed point, it would be a different case, but such loss is irrelevant here.
that is assuming that the engineer would attenuate the signal digitally... it is more likely the case that the signal would be pushed out the DAC at it's existing volume (hence a lower noise floor out of the computer) and then be attenuated passively. |
So you do use a DAW for mixing, right? Remember im telling you what ME's like to be handed, you can mix to 0dbfs on DAW, but why would you when ME's ask for headroom to work with for various reasons? Even though its 24-bit floating point, there is obviously reasons why they still ask for the headroom.
With everything said, the most important thing is the relative volumes of the different channels/tracks, and that can be done at -6dbfs on a DAW or at 0dbfs, though the latter leaves very little space for the ME to do his magic. But whatever man, do what you want with your mixes, I suggest talking to a few ME's or seasoned producer's first though. |
|
|
| echosystm |
so... the only rational reason for engineers to demand headroom is to make sure noobs don't have the signal clipping?
the only other thing i can think of is that perhaps the amplification circuits in their analog gear doesn't work so well near the extremes. in this case they would want to boost it a bit more to avoid any distortion. on high quality mastering gear, you wouldn't really expect this though? |
|
|
| Kismet7 |
| quote: | Originally posted by echosystm
so... the only rational reason for engineers to demand headroom is to make sure noobs don't have the signal clipping?
the only other thing i can think of is that perhaps the amplification circuits in their analog gear doesn't work so well near the extremes. in this case they would want to boost it a bit more to avoid any distortion. on high quality mastering gear, you wouldn't really expect this though? |
since your obviously a blatant noob (LOL sorry, its getting ridiculous), I really suggest talking to a professional ME who works on DAW mixes, see what they want/prefer from you. |
|
|
| echosystm |
| quote: | Originally posted by Kismet7
since your obviously a blatant noob (LOL sorry, its getting ridiculous), I really suggest talking to a professional ME who works on DAW mixes, see what they want/prefer from you. |
if i'm such a noob, why have i foiled every "reason" that you have come up with, for leaving headroom? why has everyone agreed with me?
;)
i'm waiting for someone who actually knows what they are talking about to come in and give me a reason other than "because an engineer said so" and "because they need room, but i don't know why".
please stay out of this thread, it is obviously beyond your knowledge.
for the record, amplification circuits DO tend to not work optimally at their extremes - particularly shifts in stereo image and so on. if you didn't already know that, you really should have never come into this discussion. my bet is that you don't even know the difference between fixed point and floating point numbers. if you don't, take some time to sit down and think about your life - you're like a 3 year old chasing a nobel prize with "facts" daddy told you.
i've now heard two hilarious quotes from you:
"vsts and anlog synths might be doing the same maths"
"24-bit float" - this isn't even supported by X86 CPUs
:haha: |
|
|
| Kismet7 |
| quote: | Originally posted by echosystm
if i'm such a noob, why have i foiled every "reason" that you have come up with, for leaving headroom? why has everyone agreed with me?
;)
i'm waiting for someone who actually knows what they are talking about to come in and give me a reason other than "because an engineer said so" and "because they need room, but i don't know why".
please stay out of this thread, it is obviously beyond your knowledge.
for the record, amplification circuits DO tend to not work optimally at their extremes - particularly shifts in stereo image and so on. if you didn't already know that, you really should have never come into this discussion. |
They'd be wasting their time, since you don't even mix on an Analogue Console, yet you were talking about Noise Floor being a big issue for not having headroom, taking the discussion somewhere that hardly pertains to you, does it?. I think you just felt like nerdin out but failed?
Besides, I dont think anyone here is going to supercede the knowledge of a professional mastering engineer like Robert Babicz who asks for -6db peaks for headroom. You haven't even admitted you use a DAW for mixing yet because it makes most of your arguments as worthless as every other argument you've had with me.
Now stop acting like you know , and go learn something. Or better yet, learn it and actually implement it correctly. Whats the point of knowing or thinking you know this technical things when you probably suck at implementing it. |
|
|
| echosystm |
| quote: | Originally posted by Kismet7
You haven't even admitted you use a DAW for mixing yet because it makes most of your arguments as worthless as every other argument you've had with me. |
of course i use a daw... guess what... when audio comes out the DAC, suddenly we have a noise floor.
what? NO WAY!
anyway man, i cbf wasting any more time on your stupidity. you're on the ignore list so i don't even have to hurt my brain with your retarded posts anymore. you should feel pretty bad, i haven't ever put anyone on ignore... not even rubez!
for the record (to everyone else), i am still interested to know why engineers want headroom. if any of you know, by all means chime in! ...but please come up with something better than "because robert babicz says so" and "because they do, but i don't know exactly why". :haha: |
|
|
| Acton |
| quote: | | Originally posted by echosystm i am still interested to know why engineers want headroom. if any of you know, by all means chime in! ...but please come up with something better than "because robert babicz says so" and "because they do, but i don't know exactly why". :haha: |
I'm not exactly a sound engineer, but I just assumed it was because it allows plenty of room to play with. It leaves space to try various types of EQ'ing and compression whilst at the same time leaving room for the dynamics of the track to still shine through.
Personally I leave a little headroom for final EQ/Compression touches, but not massive amounts. |
|
|
| kitphillips |
| quote: | Originally posted by echosystm
"24-bit float" - this isn't even supported by X86 CPUs
|
Really? Why is that? Hadn't actually heard that before...
theres a lot of misinformation in this thread.
1/ Its always best to mix to 0 DBfs
2/ That means PEAKING at 0dbfs, so a lot of engineers say -3 so that if theres a bit of a clip it doesn't matter.
3/ -0.3 is standard in CD mixing, so perhaps thats where the number comes from. I suspect its -0.3 rather than 0 to allow for errors in decoding and the equipment on the consumer end.
4/ The higher the bit resolution, the lower you can peak, because theres more quantisation levels. But if your eventually bouncing to 16 bit, you really want to mix into all that headroom so that the music is as dynamic as possible. Otherwise, you arent' making use of all the quantisation levels. |
|
|
| kitphillips |
| quote: | Originally posted by Acton
I'm not exactly a sound engineer, but I just assumed it was because it allows plenty of room to play with. It leaves space to try various types of EQ'ing and compression whilst at the same time leaving room for the dynamics of the track to still shine through.
Personally I leave a little headroom for final EQ/Compression touches, but not massive amounts. |
See you'd be better normalising during the bounce and then dropping the overall level on the first effect in the mastering chain. |
|
|
| Kismet7 |
| quote: | Originally posted by echosystm
of course i use a daw... guess what... when audio comes out the DAC, suddenly we have a noise floor.
what? NO WAY!
anyway man, i cbf wasting any more time on your stupidity. you're on the ignore list so i don't even have to hurt my brain with your retarded posts anymore. you should feel pretty bad, i haven't ever put anyone on ignore... not even rubez!
for the record (to everyone else), i am still interested to know why engineers want headroom. if any of you know, by all means chime in! ...but please come up with something better than "because robert babicz says so" and "because they do, but i don't know exactly why". :haha: |
At this point, you are 0 for 3, running away and putting me on ignore is the best move for you, highly recommended. Stick to that, because frankly im tired of your bs technical nerdery. Can only hope it nets actual results for you though, because i'd feel stupid debating about this stuff, and you can't even implement technical knowledge to actually make good music. By the way I already covered why its good to leave headroom, between being easier to add compression, to adding eq, and of course why would you question someone like Robert Babicz? Your like those dumb boxers that dont listen to their boxing coach and get knocked out in the first round. Listen to seasoned masters whenever you get the chance to hear what they have to say. |
|
|
| Acton |
| quote: | Originally posted by kitphillips
See you'd be better normalising during the bounce and then dropping the overall level on the first effect in the mastering chain. |
Can't say I have tried this. But to me it seems like dropping the level on the first effect and then potentially boosting it with other effects would have an effect on the dynamics. |
|
|
| echosystm |
| quote: | Originally posted by kitphillips
Really? Why is that? Hadn't actually heard that before...
|
the IEEE specification is something like...
half precision (16 bit)
single precision (32 bit)
double precision (64 bit)
etc.
anyway... it's all about what the FPU inside the processor was designed to handle. im not sure why it is like this, but it probably has something to do with the binary number system (radix of 2). eg 2^4=16, 2^5=32, 2^6=64, etc. if you deviate from this, then you can't rely on hardware for FPU calculations and it will be slow as hell. i don't know of any CPUs that have hardware support for 24bit float. 24 bit fixed point is obviously a different ball game entirely though.
the IEEE has extended precision (80 bit) too. i don't really know how this fits into the grand scheme of things. probably alanzo would know this.
edit: x86 word length is 16 bits, so any number less than 16 bits is wasting space. all the above are multiples of 16, as is 80. so yeah, it's something to do with that. i cbf looking further into it haha. |
|
|
|
|
