1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.

    Dismiss Notice

Is burn in real or placebo?

Discussion in 'Sound Science' started by hifi man, May 13, 2013.
37 38 39 40 41 42 43 44 45 46
48 49 50 51 52 53 54 55
  1. bigshot
    I don't think we're calling you a liar. You're clearly saying what you believe. We're just saying that your results aren't necessarily the same as typical results, and there may be an error in testing that caused it. If you want to find out if that's the case, you should try to do a more controlled comparison and see if it comes out the same. Here in Sound Science, that's what we do... test our hypothesis to see if it holds up under controlled conditions and is repeatable consistently.

    If you are interested to see where your threshold is for lossy formats, I have a test I can share with you that includes samples of three different codecs with three different data rates compared against each other and against a lossless sample. It's really handy to figure out what you can hear and what you can't. If you'd like to try it, send me a PM and I will administer it to you.
    Last edited: Mar 13, 2019
  2. gregorio
    1. I didn't call you a liar, I just listed that "Maybe" you were lying as one of several different possibilities and it certainly wouldn't be the first time someone has done that in this sub-forum.

    2. And generally, in most cases, I too wouldn't have expected anyone to be able to tell the difference. However, the other side of that same coin means that there could be certain cases/conditions under which it would be possible to reliably tell a difference. Around the same time (late 1990's, early 2000's) I quite extensively tested the transparency of MP3s and generally it was transparent but there were certain pieces/recordings where I could reliably identify a difference, repeatedly achieve a perfect score with an ABX test. And, that was with 320kbps MP3s, not 192k! Furthermore, all my audio engineer colleagues were also able to reliably tell the difference (at 320kbps) and, this was also the consensus of the audio engineering community as a whole. Of course, we didn't really fall under the banner of "generally", we were all professionals; experienced, trained listeners with listening environments (commercial studios) significantly superior to the "general" public. An MIT educated acoustic engineer would certainly have been aware of the state of MP3 encoders in the early 2000's and therefore should have expected that at only 192kbps most people with most material probably wouldn't be able to tell the difference but some would.

    3. In fact, with some material, I can't even hear the difference between 128kbps MP3's and the original wav files! As ALWAYS though, like with amps and DACs all sounding the same, there are conditions. In this case, there is the stated condition of "some" material, PLUS the non-stated condition of which MP3 encoder, which version and what encoder settings. Many encoders, LAME and Vorbis being two examples, are constantly being tested and developed. Testers report which material (commercial recordings/pieces of music) they are able to differentiate with ABX testing and then a new ".1" version of the encoder is released (with tweaked settings to the employed psychoacoustic model), which eliminates those recordings/pieces from the material which can be reliably differentiated. This has being ongoing for around 20 years and today, the formal testing of MP3 encoders at 320kbps no longer really occurs, because the list of pieces/recordings which can be reliably differentiated at that bit rate has been reduced to zero. Testing today occurs at much lower bit rates, where the list of material that can be differentiated increases as the bit rate reduces (and varies according to the encoder). My answer to your question is therefore BOTH "no I can't and no one else has reliably demonstrated the ability to" AND "yes I can, it's trivially easy and whole communities are also able to". Without the conditions/variables mentioned, your question is meaningless. Let me ask you a question; you can't hear the difference between 320kbps and wav files??

  3. bigshot
    Fraunhofer MP3 isn't nearly as good of a codec as the more modern ones. I'd like to find someone who can consistently identify AAC 256 or 320 from lossless with even the most difficult music to encode. Haven't found any yet.
  4. 71 dB
    Years ago they said Fraunhofer is the best with high bitrates and LAME is the best at low bitrates but I suppose a lot has changed and codecs today are superior to those 15 years ago so who really cares?
  5. bigshot
    I've found that LAME is also better than Frau at high data rates too. There are codecs specifically designed for super low data rate voice applications, but most of the ones we talk about here are for music and higher data rates.

    The reason it matters is because a lot of people seem to have based their opinions on compressed audio based on Fraunhofer MP3, which is almost 30 years old now. What was true in 1991 isn't necessarily true today.
    Last edited: Mar 14, 2019
  6. AKGForever
    I think there is probably a short break-in period for mechanical devices such as speakers or headphones but no more than an hour or two. Parts flex, parts seat.

    I recently read an article for a capacitor that stated that capacitors had to settle down for a week or two to reach their optimum performance. The article was talking specifically about crossover capacitors. Has to be one of the most ludacris thing I every read. Whatever brand they were talking about, I don’t want them.

    The whole burn-in thing comes from the computer industry. With computers it was used to weed out “Infant Mortality” failures and not performance issues. Memory and Hard Drives would last forever if they made it past a few days of hard use
    Last edited: Mar 17, 2019 at 6:49 PM
  7. jagwap
    The capacitor thing is true. I've seen the plots from the manufacturer. The ESR alters in the first 10s of hours when they they have their working voltage applied across them. However crossover capacitors do not have a voltage across them,so this may be different. They do have current passing through them, but I haven't seen how that affects their parameters.
  8. geek707
    And this settling down isn't done by the manufacture in QC under controlled conditions, why?
  9. jagwap
    Who are you asking, and on what subject?

    Answering generally, it makes no financial sense to hold on to stock you could sell on, to do something the customer can. Some products are tested for 12 to 24 hours by manufacturers who want to look for early failures, and this will do the majority of burn in if there is any.

    Speakers definately experience burn-in. It is widely known in the industry and is measurable.

    Headphones - it is fiercely debated here. I have recenly discussed this with driver engineers and they have explained to me that the spider is the main culprit in loudspeakers, so as headphones usually don't have one, maybe the effect is much less or non-existant.

    Electronics: electrolyic capacitors do exhibit burn in, and I've encoutered it often, and seen the manufacturer's data.

    Cables: I see no evidence and find nothing. Seems incredibly unlikely.
    Brooko likes this.
  10. bigshot
    If I was a manufacturer, I wouldn't sell something to a customer that didn't sound the way I intended it to. That's just asking for returns. If all it takes is 30 or 40 hours of playing tones, I'd ship a couple of days later and ship them sounding right.

    It makes no sense that a company would put all that R&D and manufacturing costs into producing a consistent product and then ship it unverified and let the customer find out if it settles into the right place or not. If burn in exists and it is mechanical, it isn't going to be consistent. The outcome is going to be a crap shoot.
    geek707 likes this.
  11. geek707
    Thank you. This is the point I've tried to make. If a company believes that product x requires burn in, why aren't they burning the product in? That said, I think most manufactures believe burn in is nonsense whatever they say publicly.
  12. AKGForever

    Burn-in is a loaded word, I prefer break-in when it comes to mechanic devices. On the contrary, break-in can be very consistent. With motors, as an example, mass produced parts mate with proper break-in. Each part wears a little to match the other and then wear stops
  13. geek707
    I disagree. That's a apple and oranges comparison. When you talk about break-in you think of, for example, a internal combustion engine, break-in was something like piston rings mating with cylinder walls to get the best seal. You ran a cylinder hone into the cylinder to create, among, other things a slightly rough surface for the rings to wear against and create a seal.
    The is nothing in the high-fi world that has the energy or mass levels of what's going on in a engine.
  14. bigshot
    If it was comparable, we would need to do oil changes on our transducers!
  15. AKGForever

    Perhaps not the best example but speaker surrounds soften and spiders flex. I am not saying it is even audible but mechanical devices change with initial use. I think that 100 hour burn-in is ridiculous.
    Last edited: Mar 18, 2019 at 6:01 PM
37 38 39 40 41 42 43 44 45 46
48 49 50 51 52 53 54 55

Share This Page