Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jul 20, 2020 9:22:13 GMT -6
Not sure if this is the appropriate forum for this, but I suspect it is....
Hoping to get a more technical explanation of if/why using a dedicated line-level attentuator, like the DIYRE L2P (which I have), or the A-Designs ATTY is a better choice for bringing line-level signals down to utilize preamps as line amps during mixing, rather than just using clip gain or the trim plugin in PT before sending out of your DA and/or just using the pads on your preamps.
Is there a difference, or is it all just padding things down? The L2P is a pretty cool and easy to build box, but is it necessary?
|
|
|
|
Post by svart on Jul 20, 2020 9:40:47 GMT -6
Not sure if this is the appropriate forum for this, but I suspect it is.... Hoping to get a more technical explanation of if/why using a dedicated line-level attentuator, like the DIYRE L2P (which I have), or the A-Designs ATTY is a better choice for bringing line-level signals down to utilize preamps as line amps during mixing, rather than just using clip gain or the trim plugin in PT before sending out of your DA and/or just using the pads on your preamps. Is there a difference, or is it all just padding things down. The L2P is a pretty cool and easy to build box (even though the switch on mine broke almost immediately), but is it necessary? All stems from the theory that the math used for gain/attenuation in the DAW is ultimately lossy. The multiplication and division needed to change gain creates words that are too big to process in real time, so either you need to buffer large amounts of data or cut the amount of information down (truncation). The first leads to large latencies and processor/RAM intensive workloads and the latter can cause loss of fidelity. The second part of the theory is that when doing offline rendering, the DAWs tend to use higher precision processing since it's not real-time, and that's why some folks swear they hear a difference. Of course, some say they don't hear a difference and claim that the plugins used probably alter the sound of the signal more than the truncation does anyway. So, the purists believe that the only precise way to change gain is to do so in the analog domain since it only modifies the voltage. The fact that nobody has really been able to reliably provide proof that the digital method adversely affects the audio to any reasonable degree should say all you need to know. That and every single thing you touch in the DAW (pan, gain, sends, receives, plugins (both active and inactive), etc) modifies the signal going through it to some degree and probably have a few magnitudes more effect than whether or not you use an analog attenuator or just drop the gain a little via a slider in the DAW. Choose whichever works better for your workflow and you're good to go. |
|
|
|
Post by Blackdawg on Jul 20, 2020 9:53:34 GMT -6
I was always under the impression that you would use more bits and there fore a larger scale from your DA output as opposed to a very small section of the DA output which helps with noise and things. So if you are coming out of the DAW it's better to be higher in the range then knock it down to drive back up via whatever process you want.
|
|
|
|
Post by EmRR on Jul 20, 2020 10:17:42 GMT -6
To me the significant difference is impedance match. One or the other may plot a more or less desirable frequency response, and that may be another crayon in the box.
Bit depth arguments pale in comparison, unless maybe the converter is ancient or very cheap.
|
|
|
|
Post by drbill on Jul 20, 2020 19:57:56 GMT -6
I agree on the impedance thing, but IMO an ATTY or other device will not necessarily get you where you want to go. it will level wise, but perhaps not impedance wise. In point, Dan Deurloo spent a LOONG time designing a set of pads for me to mix into his ZOD DI's. He tested lots of different levels and impedances before setting on the right one.
Personally, for line level gear, I just use trim before I exit via PT Hardware Insert. For mic preamps, I generally use the Avedis pad going into them. I still might have to trim one way or the other as I like to leave my input levels set on my outboard, but I've heard no degredation that I can notice via the trims feeding my outboard. And often, I'll have trim, outboard a, return, trim, outboard b, return, trim, outboard c, return, trim, outboard d return on a single track. If there was significant degredation in real world practice I'd notice.
Good luck! Trims for line level outboard. Avedis pads for mic inputs!!
|
|
|
|
Post by johneppstein on Jul 21, 2020 18:44:05 GMT -6
To me the significant difference is impedance match. One or the other may plot a more or less desirable frequency response, and that may be another crayon in the box. Bit depth arguments pale in comparison, unless maybe the converter is ancient or very cheap. Erm, dithering.
Any time you change level digitally you should dither. And dithering does change the tone quality. (If you don't believe that, go through Chris Johnson's (Airwindows) set of dithering plugins and compare. He has a lot of them and all sound different.)
|
|
|
|
Post by svart on Jul 22, 2020 8:19:06 GMT -6
To me the significant difference is impedance match. One or the other may plot a more or less desirable frequency response, and that may be another crayon in the box. Bit depth arguments pale in comparison, unless maybe the converter is ancient or very cheap. Erm, dithering.
Any time you change level digitally you should dither. And dithering does change the tone quality. (If you don't believe that, go through Chris Johnson's (Airwindows) set of dithering plugins and compare. He has a lot of them and all sound different.)
If that were true, you'd possibly have thousands of instances of dither in your DAW then as it passes through all processes.. Even uncorrelated noise would add up in this case eventually since uncorrelated noise will add 3dB per doubling of additional sources and correlated noise will double with every additional source. This would quickly swamp all quantization and truncation noise. Luckily, random system noise through active devices will usually be higher than both quantization noise and truncation noise and it's only a matter of academics, especially when working with 24bits. Besides, most DAWs use floating point math these days, so the native bit depth is never changed once encoded, only the wordlength needed to maintain the representation changes. The output of the DACs has a reconstruction filter that removes quantization noise on the output. The only real place it matters is when rendering down a mix in the DAW and you're changing bit depths but not going through an intermediate hardware conversion stage. |
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jul 22, 2020 14:01:52 GMT -6
Dither me timbers!
|
|
|
|
Post by craigmorris74 on Jul 23, 2020 8:04:51 GMT -6
If a person is paranoid about degrading the sound by adjusting volume in the DAW, Airwindows Purest Gain addresses this problem.
|
|
|
|
Post by EmRR on Jul 23, 2020 8:36:48 GMT -6
The ATTY is more likely to be same as what they did in the 1940's-1970's on broadcast equipment for adjustable bridging line input to preamps, but not certain of the approach used.
The Avedis LPZ is also pretty much on target, and decently priced too. Just not adjustable.
Many converters don't really have the necessary low impedance drive to feed a preamp directly without treble losses, that's one side of the impedance question. Others are very low, really too low for direct interface nd will make some smiley face EQ driving many vintage color preamps.
|
|
|
|
Post by brenta on Jul 26, 2020 10:02:41 GMT -6
I tried a few different methods to run line level through preamps and wasn’t happy with the results. The pads built into preamps usually sound like crap. What finally made me happy was running the signal through a Radial JDI Duplex, which has a configuration for this purpose. It pads the input, converts the impedance, and also passes the signal through the Jensen transformers, which is probably a big reason I like the way it sounds. It’s also a great DI.
|
|
|
|
Post by svart on Jul 26, 2020 10:14:59 GMT -6
I tried a few different methods to run line level through preamps and wasn’t happy with the results. The pads built into preamps usually sound like crap. What finally made me happy was running the signal through a Radial JDI Duplex, which has a configuration for this purpose. It pads the input, converts the impedance, and also passes the signal through the Jensen transformers, which is probably a big reason I like the way it sounds. It’s also a great DI. Most line level outputs are only designed to drive 10k-100k loads. Preamps tend to have 600-4k loads, so having an impedance transformation designed to lighten the load on the source is important. A resistor pad can do it, if it's designed to adapt impedances as well, but almost all of them are designed to maintain the same impedance through them, not change impedance. |
|
|
|
Post by EmRR on Jul 26, 2020 14:28:58 GMT -6
What svart said ^ The pads built into preamps usually sound like crap. What finally made me happy was running the signal through a Radial JDI Duplex, which has a configuration for this purpose. It pads the input, converts the impedance, and also passes the signal through the Jensen transformers, which is probably a big reason I like the way it sounds. The DI; the transformer is doing the impedance conversion. The -20 dB pad on a preamp is typically about 1500Ω, which some things won't like, and usually not enough with line level signals to get you in the right gain zone. The -20dB pad here is 10KΩ. You can also put (the right) 3 resistors in a barrel and you have the Avedis pad at 10KΩ, meant for direct connection to a mic pre. And that's a different pad in a different situation to arrive at the same result. |
|
|
|
Post by Guitar on Aug 6, 2020 5:23:08 GMT -6
|
|
|
|
Post by matt@IAA on Aug 6, 2020 8:11:15 GMT -6
If you want a 10k input impedance for your line level and a 200 ohm source impedance for your mic pre you end up at an attenuation of 34 dB. Close enough to use 4k7s for the legs and 200 ohm for the shunt.
|
|
