Discussion in 'Measurement Techniques Discussion' started by atomicbob, Feb 11, 2018.
Please reference the intro which appears at the top of many of my technical measurement posts:
If you are unfamiliar with audio measurements please use a search engine with the query:
"audio measurements" or "audio measurement handbook"
Look for publications by Richard C. Cabot and also by Bob Metzler, both from Audio Precision. There are other useful publications as well. These will provide basic knowledge.
The choice involves compromise between resolution and time required. High resolution measurements can take 30 minutes each. Not helpful when you need to make several hundred high quality measurements in a day. My typical measurements utilize scripts such that consistent measurement methodology is applied. That is why @Ali-Pacha could pull up those measurements for a direct comparison.
I have put a lot of those graphs in the measurement section on several forums. Most of mine are labeled Device Model Techincal Measurements. "Really bad" is relative. Modi Multibit and Bifrost MB are using a 16 bit ladder DAC intended for instrumentation use which require level accuracy. As BiMB has 2Vrms output at 0 dBFS, -78 dBFS is the last 3 bits to produce approximately 244 uVrms. This is about 0.2 nanowatts for an HD650 or HD800. at 300 ohms.
Not easily. At this time none of the audio measurement system tools support linux for their measurements. I would need to create each signal I wished to use for stimulus as a wavefile, load it onto a linux system with a media player and (this is the big one) synchronize playback of the linux system with the measurement system data acquisition on the Windows PC.
@atomicbob Thanks for the litterature references. I'll have a look at all this to have a better understanding of the meaning of those two kinds of graphs (glitch at zero cross and linearity analog vs digital). Even if you've already given some intel on this specific stuff :
Ironically, it was some experimentation with plug-ins that finally made me throw up my hands in surrender and buy jrmc. As soon as it feels like I am spending more time f'ing with a computer than enjoying music, I am out. Computers are only a means to an end for me, and my knowledge in this area is for shit. In fact, in college I took a programming class using a card punch/reader, which explains both my age disadvantage and disinclination to computer-tweak fuckery.
Now Amir says that Bifrost Multibit is a 12 ENOB DAC......
So are we saying that the last 3 bits on the Bifrost Multibit are not usable? This is certainly not what I was lead to believe...
I think he's saying the Bifrost Multibit won't reproduce sounds at inaudibly quiet parts of the signal. I could be wrong, but it seems to me that this is fine. If it were screwing stuff up at -40dBS or higher, the DAC would sound like shit.
It doesn't sound like shit. It sounds like Schiit. I'm ok with that.
Yeah, I completely understand that one.
I just wanted to have a play with some EQ plugins and Sonarworks for fun, but the flakiness and special cases in the Foobar VST hosts drove me mad too- didn't want to babysit broken stuff, even if it was free.
That is a good question, thank-you for asking. The following answer is not directed specifically at you, but at many who ask a similar question or focus on a single issues.
It is not a question of usability, it is a question of perspective. Can you hear 0.2 billionths of a Watt on an HD8x0 or HD6x0 for example. The "massive glitch" expounded by a few is another good example of "is it audible?"
Attached is a wave file to test with your own ears.
WARNING. This wavefile contains a 440 Hz sine at -1 dBFS. It will be LOUD. Turn down your volume before listening. Then adjust volume such that the first 5 seconds plays at a comfortable level. After the first five seconds level will change suddenly to a much lower level for the remaining 10 seconds. Can you hear what is playing during the remaining 10 seconds? This is a listening test. Do not listen with eyes, looking at a level meter, RTA, spectrum analyzer etc.
I love measurements and they do allow us to examine audio equipment with a microscope, but at the end of the day, what is and isn't audible matters. Context and perspective are important.
@atomicbob - While I appreciate your efforts, and respect an artist's right to creative expression, I have to say... your attached song was boring and short. I regret to inform you we can't publish an entire CD of this, and a record deal is not feasible at this time.
What dBS are the last 10 seconds?
That file is possibly the best audio file I've heard all year. Even cheating, there's a zero percent chance that I could hear this.
You just don't get electronic music.
Spoiler: Wave file details - do NOT read until after listening
5 S 440 Hz -1 dBFS sine
5 S 440 Hz -84 dBFS sine
5 S 440 Hz -84 dBFS square - far worse than the Modi MB and Bifrost MB "massive glitch"
I thought maybe it was a joke so I opened it up in audition, turned on SFD and sure enough...
edit: p.s. I couldn't hear it AT ALL so yeah, well done.
Here is another one for fun. Can anyone hear the Hidden Message in the attached wavefile. It is only 20 dB below the main signal which is pink noise.
WARNING - this file is also LOUD - lower volume and bring up to a comfortable level.
Here are some visual details for guidance:
Just adding some perspective. This doesn't mean I don't believe in higher resolution in DACs but want people to think about when it matters vs. when it doesn't. The "massive glitch" is a great example of not being the problem extolled. If a perfect sine at -140 dB was the only attribute that mattered, I'd only have to make one measurement per device and be done with it.
I'm sure folks will provide arguments as to why you can't hear it in these sorts of tests, at least when factoring in the rest of the gear chain, but can in actual music. Hence why measurements matter to some extent and are absolutely worth posting about religiously online. Or that if it doesn't matter, spend less on a product that measures better for peace of mind. Or just stick to what's on your phone or motherboard. All of which are both contradictory arguments, and, yet, often come from the same sources at various points in time, thus leading me to think we're all just silly lunatics, a fact of life I've already accepted of myself.
The real truth is that everything simultaneously does and does not matter, therefore insert some arbitrary point here that makes me feel powerful about myself and helps me sleep well at night, before I, some day, eventually die and turn to dirt.
Or, you know, just, like, EQ the bad sound in, man.
Nwavguy is a sad sod. Not only he uses atomicbob's testing output to validate his own measurements*, but in the same post he then proceeds to denigrate him and the quality of his work. The prick considers himself a sage or something.
* I'm still not taking his pronouncements of "same results" at face value, and hoping more competent folks can fill in the gaps. I'm pretty confident though that we still don't know the precise conditions that Amir used to obtain his results; it's hard to take him at face value for anything.
Well, this is just masking.
Just be careful. This is the same arguments FOR mp3.
But yes, 20db down is huge.
First off thanks for all the good work, it is appreciated by people with a brain. That being said, what you mentioned basically that you have keep measurements in perspective and prioritize them is such a good point. Some poor guy on the shithouse or was it shithole amir fan board tried to explain this and a few other clear thoughts to the torch carrying mob. Did not go over very well. They have the mindset of the walkers from the Living Dead. He even asked amir if he listened to the multi-bit, the dumb excuse was he only listened for 2 hours for noise or dropouts! lmao
The amir is afraid to say the DAC sounds good because it would go against his smear campaign.
Below is a video with the amir at a show asking questions of Jonathan Novick AP, at the end of a presentation. The amir starts at the 57 min mark. and you can hear how much of a douch he is. Jonathan knew it but was trying to be polite to him.
Well, he is saying something like can you please make a model for idiots. Or, How about a consumer-oriented model.
I (much as I hate to say anything non-negative about Amir) don't think those would be douche questions, although the right answer might be "Because we make gear for scientists. That is our market. Sorry, but if you want to use it, you are going to have to learn how. Just like everyone else." I gather that that stuff is pretty expensive: do people who don't have a clue really buy it?
Well, apparently yes. And doesn't Amir represent himself as something of an electrical engineer?
OKOK.... They are douche questions.
Separate names with a comma.