|
|
Post by bluesholyman on Jun 17, 2024 12:30:45 GMT -6
I am trying out an interface for my "office" computer where I listen to music and do zoom, etc. When I travel, the interface acts as my portable recording solution for ideas, etc. Mostly a scratch recording scenario when I travel because hotel rooms, etc. aren't exactly acoustics friendly.
The interface I am trying now claims a dynamic range (A-weighted) of 120dB on inputs and 127dB on monitor outputs.
A far less expensive one (1/3 -1/4 the cost) claims a dynamic range (A-weighted) of 104dB on inputs and 102dB on monitor outs.
That seems like a considerable difference to me on paper, but what is the practical difference in use?
|
|
|
|
Post by svart on Jun 17, 2024 12:44:18 GMT -6
I am trying out an interface for my "office" computer where I listen to music and do zoom, etc. When I travel, the interface acts as my portable recording solution for ideas, etc. Mostly a scratch recording scenario when I travel because hotel rooms, etc. aren't exactly acoustics friendly. The interface I am trying now claims a dynamic range (A-weighted) of 120dB on inputs and 127dB on monitor outputs. A far less expensive one (1/3 -1/4 the cost) claims a dynamic range (A-weighted) of 104dB on inputs and 102dB on monitor outs. That seems like a considerable difference to me on paper, but what is the practical difference in use? Almost none. When was the last time you listened to anything at -100dB? Never. That's lower than the ambient noise in the room you're sitting in. That's lower than the sound of your blood in your veins when you have your earbuds in your ears.. My tinnitus is probably around -60dB! They call those things a "Figure of merit" because a lot of them are simply theoretical numbers. They might denote an amount of performance but in a secondary way. For instance, in audio they don't actually mean for you to use the devices at -100dB in any real capacity, but they are telling you that the ADC self-noise will be lower than your circuitry. It might also be important if there's a lot of channels summed together and some of the noise is correlated (technically it can't be truly noise if it's correlated in any way, but..) and it might sum, but you're still talking about only a few dB of noise gain, etc. So honestly there really isn't a good technical reason to worry about stuff that low. Some well-regarded converters just a decade ago might not even make it to -90dB and people love them. In any modern converter I'd worry about coloration and clarity (things coming from the analog aliasing filters) before ever worrying about the noise floor. |
|
|
|
Post by Dan on Jun 17, 2024 13:57:49 GMT -6
The only thing that matters are the jitter and distortion measurements. The SNR is mostly how new the chip is. There is no way in hell a UAD Apollo X is as good as a Lynx Hilo 1 but it uses newer chips so it has a lower dither noise floor that's still below the noise of a quiet mic amplified by a quiet pre. The delta sigma converters also auto mute to prevent chirping. You cannot take a simple noise measurement with silence. The reviewers are mostly shills and refuse to test them for IMD or put in high frequency tones into them. Then idiots on certain websites, mainly Audio Science Review and Gearslutz, go on a field day about how X converter is best when it's worse in every single way than Y converter and most of the posters are listening to recordings done to tape with hiss or loudness war bullshit that could fit into 8-bit audio with a noise shaped dither and probably has insane noise not on silence from being overmodulated unless they izotoped it out or used expanders on everything but they do not unless it's movie dialogue. That used to be all CEDAR DNSed and now it's all Izotoped to death by fools and the lazy.
|
|
|
|
Post by bluesholyman on Jun 17, 2024 14:01:12 GMT -6
In any modern converter I'd worry about coloration and clarity (things coming from the analog aliasing filters) before ever worrying about the noise floor. What metric would I be looking at for coloration and clarity, apart from the one my ears can give me? I guess at some given price point, the "chips" are going to be of comparable quality, no? I mean if its an Apogee, I'd expect it all to be very solid. If its say an M-Audio, Behringer, etc., I'd expect it all to be just "meh." I guess specs are specs, but price point seems to be a universal indicator for an overall quality level, no? I realize I am sorta answering my own question indirectly, but this is a bit of technical curiosity for me also. |
|
|
|
Post by svart on Jun 17, 2024 14:52:38 GMT -6
In any modern converter I'd worry about coloration and clarity (things coming from the analog aliasing filters) before ever worrying about the noise floor. What metric would I be looking at for coloration and clarity, apart from the one my ears can give me? I guess at some given price point, the "chips" are going to be of comparable quality, no? I mean if its an Apogee, I'd expect it all to be very solid. If its say an M-Audio, Behringer, etc., I'd expect it all to be just "meh." I guess specs are specs, but price point seems to be a universal indicator for an overall quality level, no? I realize I am sorta answering my own question indirectly, but this is a bit of technical curiosity for me also. There really isn't any kind of metric for coloration or clarity since those are somewhat based on opinion, i.e., one man's "muddy" is another man's "warm" and so forth. Almost all modern converter IC's are nearly identical in performance when external factors are equal. Luckily for us living in the future, most IC's also integrated a lot of what was external just a decade ago. Mostly it's just the aliasing filters and power supplies that are external influences apart from the obvious clocking system. With that, as long as the designers don't cock up the signal, it should be good. I think a lot of the cheapo manufacturers have learned some lessons by copying other, better, designs over the years, so I wouldn't write off Behrry and those guys on the audio quality, which leads me to the next thought. I think that even the cheapest converters these days are pretty usable. Where they tend to go wrong is in the drivers and cheapness of the pots/connectors/cables and they tend to break pretty easily or the power supplies will die faster over time than something more pro. Chinese manufacturing was once really trashy, but it's come a very long way and the quality is becoming less of an issue, but it still follows the throw-away-society where they don't expect anyone to fix anything so they don't design/build anything to be repaired either. I'd just take a listen to what you want to buy and keep or return as you see fit. One side effect of the throw-away-society we live in is the ability to return just about anything for any reason. |
|
