How about the goal of reproducing whatever audio signal it’s given.
A loudspeakers ability to do that can be measured in a number of ways separately but better yet for the person fine tuning something they built, in addition to measurements, the flaws can be listened to with music of your choice using a measurement grade microphone.
To make them more obvious you can do a generation loss recording. Using a sound card or better with 24/96 recording capability and measurement mic, you can record stuff in your living room first to get the feel for how faithful it is (best with headphones)..
Do that generation loss recording with the loudspeaker outside and you get rid of the room, leaving only the speaker.
The more faithful it is to the signal, the more generations one can undergo, just like with electronics, recording tape, the AD/DA process and so on.
No kidding guys, this doesn’t tell you what to fix but you can sure hear it and often figure out where to look.
Each generation becomes a more exaggerated version of “the flaws” which alter the signal, where it is faithful to the signal, there is no alteration.
Go just three generations (outdoors) and still be listenable and you have a much better than average loudspeaker.
Best
Tom Danley
Danley Sound Labs
The more DSP oriented might suggest that one compose an impulse to convolve which had X generations of degradation applied but for me, the recording was satisfying to do and hear.
A loudspeakers ability to do that can be measured in a number of ways separately but better yet for the person fine tuning something they built, in addition to measurements, the flaws can be listened to with music of your choice using a measurement grade microphone.
To make them more obvious you can do a generation loss recording. Using a sound card or better with 24/96 recording capability and measurement mic, you can record stuff in your living room first to get the feel for how faithful it is (best with headphones)..
Do that generation loss recording with the loudspeaker outside and you get rid of the room, leaving only the speaker.
The more faithful it is to the signal, the more generations one can undergo, just like with electronics, recording tape, the AD/DA process and so on.
No kidding guys, this doesn’t tell you what to fix but you can sure hear it and often figure out where to look.
Each generation becomes a more exaggerated version of “the flaws” which alter the signal, where it is faithful to the signal, there is no alteration.
Go just three generations (outdoors) and still be listenable and you have a much better than average loudspeaker.
Best
Tom Danley
Danley Sound Labs
The more DSP oriented might suggest that one compose an impulse to convolve which had X generations of degradation applied but for me, the recording was satisfying to do and hear.
This is obviously a much more ideal way of doing things, but I've found an easy shortcut is to use very 'difficult' recordings - as an example, old, poorly transcribed from shellac copies where one is struggling to hear anything that translates to soundstage, etc on normal systems; these strongly emphasise everything that's wrong with the playback, if you know what the recordings can sound like from hearing them on suitably optimised setups. Real gains from 'correct' changes are easily heard, there is no need for measurement gear here ...
That is a very interesting point, fas42.
Tom Danley, as a fan of MST3K, will know that when a film is sufficiently bad, the audience stops suspending disbelief in the film itself and start to focus on the faults.
I think something similar happens with music playback. We've all heard systems and recorded music that get you totally involved. We've also heard stuff that is plain irritating.
For sure good vinyl playback is much better than CD or MP3.
Sometimes the room acoustic or speaker is very boomy.
Some speakers are overly bright and harsh.
The recording engineer or radio station may be applying too much compression and uniform loudness.
There may be too much background noise.
I could list a lot more typical faults, but the fact is that none of them are fatal to the listening experience. But beyond a certain level, the whole experience is tipped over the edge into listener fatigue, and you lose that precious involvement.
That's an excellent viewpoint - for me, the goal is to sufficiently refine everything about the playback so that that key level is never broached, the involvement is consistent. The more difficult recordings make it much, much harder to not overstep the mark - the optimising of everything, about every aspect of the playback's components and environment becomes ever more important.
The realisation that even the worst of the worst recordings can be held back from that tipping point, if enough effort goes into the refinement was a key "discovery" for me. The next big hurdle is then knowing what has to be done, for a particular system, to achieve that goal ...
Hi system7
I think you have hit on a key thing, where we can get totally involved or not, the suspension of belief.
Some of that is the recording, how well they create the impression of source position, space and so on. Some of it is the loudspeaker and how it interacts with the room and the electronics. For me, the ultimate audio goal is to be able to fool people into thinking it is a real person or whatever via reproduction. The problem for me was if your building a loudspeaker that other people will use to play all kinds of music and voice (not what I like), how do you make one that doesn’t color the sound. Figuring out why the loudspeaker does or doesn’t sound good may or may not show up as obvious on a measurement so why this big gap?
Our hearing system seeks to extract information out of what arrives at our ears and in that process, we are not aware of what it disregards or how it assembles our conscious single image of space from two very different and angular dependant inputs. What we can hear is a sense that “reminds us” of space, an instrument, a voice or whatever “as if” in real life. In the extreme, we have the cocktail effect where “we can hear” a conversation across a noisy room by hearing tiny fragments, watching lips and gestures and filling the blanks and in our mind we know the conversation, very little of that process we are consciously aware of.
What the microphone does is capture from one point in space, the complex two ear information is stripped away, that “hearing through the noise and assembling an image” process to extract information is partially crippled making the flaws much easier to hear.
If you have a measurement microphone and decent headphones, set that up, put the headphones on and spend a while listening in mono to the stuff going on in your house.
Get used to the way things sound, if there are people there that is great just have them talk.
Then when you have a good feel for what the tonality is hearing this way, then, put on one channel of your stereo and listen to your loudspeaker this way.
Tell me if you begin to hear coloration not audible when your automatic binaural processor / information extractor is running. What you hear here, is what stops a loudspeaker from lasting more than a generation or two before being un listenable while any other part of the electronic chain can go many generations.
Listening fatigue is not a feature of a faithful loudspeaker either no more than what you hear daily is fatiguing. Well, on second thought I have cut sheet aluminum on a table saw a few times and heard a few bands that were just as loud and musical.
I think you have hit on a key thing, where we can get totally involved or not, the suspension of belief.
Some of that is the recording, how well they create the impression of source position, space and so on. Some of it is the loudspeaker and how it interacts with the room and the electronics. For me, the ultimate audio goal is to be able to fool people into thinking it is a real person or whatever via reproduction. The problem for me was if your building a loudspeaker that other people will use to play all kinds of music and voice (not what I like), how do you make one that doesn’t color the sound. Figuring out why the loudspeaker does or doesn’t sound good may or may not show up as obvious on a measurement so why this big gap?
Our hearing system seeks to extract information out of what arrives at our ears and in that process, we are not aware of what it disregards or how it assembles our conscious single image of space from two very different and angular dependant inputs. What we can hear is a sense that “reminds us” of space, an instrument, a voice or whatever “as if” in real life. In the extreme, we have the cocktail effect where “we can hear” a conversation across a noisy room by hearing tiny fragments, watching lips and gestures and filling the blanks and in our mind we know the conversation, very little of that process we are consciously aware of.
What the microphone does is capture from one point in space, the complex two ear information is stripped away, that “hearing through the noise and assembling an image” process to extract information is partially crippled making the flaws much easier to hear.
If you have a measurement microphone and decent headphones, set that up, put the headphones on and spend a while listening in mono to the stuff going on in your house.
Get used to the way things sound, if there are people there that is great just have them talk.
Then when you have a good feel for what the tonality is hearing this way, then, put on one channel of your stereo and listen to your loudspeaker this way.
Tell me if you begin to hear coloration not audible when your automatic binaural processor / information extractor is running. What you hear here, is what stops a loudspeaker from lasting more than a generation or two before being un listenable while any other part of the electronic chain can go many generations.
Listening fatigue is not a feature of a faithful loudspeaker either no more than what you hear daily is fatiguing. Well, on second thought I have cut sheet aluminum on a table saw a few times and heard a few bands that were just as loud and musical.
Not if the recording engineer and producer are any good. The "final step" (from the "mix room" to the end user) is the understanding that the sound in (and even the purpose of) the "mix room" very likely does not match the final "listening environment" . . . and "corrections" will be made to the final mix to accommodate that (to "translate" is the common term of art). This does not always result in the best mix for a "high end" system being what makes it to market. More often with "classical" recordings, perhaps . . . there's a presumption that that tiny fraction of the market wants some kind of "fidelity" . . . but while studio recordings may be mixed on studio monitors they are seldom mixed for studio monitors.
The "generation" test makes a lot of sense for tape recorders but no sense for speakers which must be tested out of doors and hoisted high in the air by a crane... and in each and every succeeding generation recording likewise. But they are not designed to play that way and out of doors testing applies meaningfully to some parameters of speaker performance like beam-angle or distortion but not to some of the more subtle parameters.
A perception factor ignored by many is "learning" the sound of a room both as a means of interpreting sound in it as well as for setting a baseline or what psychologists call an "adaptation level." So when demo'ing your system, you always have your guest sit around a while in the room before playing serious test music.
Ben
A perception factor ignored by many is "learning" the sound of a room both as a means of interpreting sound in it as well as for setting a baseline or what psychologists call an "adaptation level." So when demo'ing your system, you always have your guest sit around a while in the room before playing serious test music.
Ben
Last edited:
In my experience most mixers will play back on a couple of systems to try and get a sense of how the system might sound for various consumer playbacks. I know they are thinking in terms of auto playback, headphone playback, small narrow bandwidth system playback. Sound is frequently heavily mucked with for the sake of "optimizing" to these possibilities.
I've been in studios where the NS-10s or Auratones were used heavily in the mixing and the big soffit mounted primary systems were just used to give the band a dB fix after the session.
Between the loudness war and assumptions of the playback system we are lucky to get any semblance of fidelity.
We need to start boycotting bad recordings.
Regards,
David S.
I've been in studios where the NS-10s or Auratones were used heavily in the mixing and the big soffit mounted primary systems were just used to give the band a dB fix after the session.
Between the loudness war and assumptions of the playback system we are lucky to get any semblance of fidelity.
We need to start boycotting bad recordings.
Regards,
David S.
Here is one place to go to get an indicator that isn't subject to audiophile opinion:
Album list - Dynamic Range Database
This database won't necessarily tell you if some recording "sounds good" or not, but it will tell you if a recording has been compressed to an overall DR rating below ~12.
I've found that all recordings below this threshold, i.e., a DR rating of 12 for the album, have shortcomings--almost without exception. YMMV.
The idea here is that if the dynamic range of the album is squashed the entire recording has probably been irreparably damaged in other ways. Note that my view is "do no damage" from the musician's creation, not from the idea of fame for the people that are charged with recording and producing copies of the music for sale.
Chris
Last edited:
I think both of the above are a good way determine if a disc suffers badly from the loudness wars. If you have never used a digital audio editor then try Audacity which is, I think, free.
I use cool edit pro (now Adobe Audition?) and can easily rip and view any track from a CD.
Lets start returning any CD that is compressed to nothing!
David
I use cool edit pro (now Adobe Audition?) and can easily rip and view any track from a CD.
Lets start returning any CD that is compressed to nothing!
David
Fascinating Chris.
Most people think that Steely Dan made some of the best sounding records ever, and they are around Green 15:
Album list - Dynamic Range Database
Dire Straits in the Green 15 too and sounding good:
Album list - Dynamic Range Database
Supertramp "Crime of the Century" was a great sounding album on Vinyl, Green 16:
Album details - Dynamic Range Database
And Lady Gaga, bless her, makes some hideously poppy compressed recordings. I hate listening to them. Red 7:
Album list - Dynamic Range Database
Seems to work! Compression is BAD!
Last edited:
if your ears cannot and do not find the recording bad, its a good recording!
I hear so often in diy forum about bad system then I begin to wonder if most people that DIY have indeed bad system
Hi
Well there is a big problem as soon as you want to record stereo and no quick answer.
We don’t hear how two microphones pick up the sound. If you put microphones in a dummy head or in the ears, you get some of the artifacts which tell your brain direction etc BUT when you play that back though loudspeakers instead of headphones, your own head and ears are there doing that again.
This is why the stereo image on nearly all recordings is produced at the mixing desk and not a captured event.
The recording process can’t easily capture the sense of “outside” etc, things
I wish to capture and reproduce, I want to fool people and how they do it now, doesn’t’ do it. In some months “an attraction” in Florida with demonstrate sound floating in front of you a stereo image, produced by sources 500 feet away. That part is getting easier.
While 90% of my focus is on loudspeaker design, I have been working on a way to capture a live events.
These are the front two channels spanning more or less your visual field, the idea being if you can create an image anywhere between two speakers like this, then all you need are more speakers to make a larger dome of reproduction.
Call me a mad inventor but at least for me, when I can get to where it really sounds like somewhere else, it gives me goose bumps. Do that with a great conventional recording, and I have seen people brought to tears including myself.
Anyway, pop on headphones DO LISTEN WITH THEM FIRST so you get a sense of space.
The parade is the mk4 version of the thing from last summer, recorded about 60 feet back from a road lined with noisy people.
The train starting up (with bugs) was in my backyard about 200 feet from the rail road tracks which are on a diagonal and done with an old version as was the bbq. That was at a friend’s when his son and the kids in the Irish folk music group played. I left a bunch of lead in because I loved all the sounds, the neighbors air conditioner, the child’s toy starting up and a mother (someone) stopping it and the kids who were going to play talking.
The problems which prevent one from being fooled into believing you’re in another acoustic environment are at both ends.
But when one can reproduce a voice played through both speakers (in Mono) and only hear the phantom image in front of you and not the right and left speakers, then when you play these recordings, it will probably sound in front of you, very much like it does when you use headphones.
https://www.dropbox.com/s/vlo7p9koy4f3f15/parade section4.wav
https://www.dropbox.com/s/jq5n4gj4mpptjpn/TrainStart.wav
https://www.dropbox.com/s/c3c0si0r7giud2w/Johns bbqTrack 04.wav
Best
Tom Danley
Danley Sound Labs
Again, most people attempt to achieve this through fiddling with speaker technology, and the way the recording is done ... but I've found that if the overall system integrity is to a high enough level, then setups with nominally very ordinary components will produce this with completely conventional recordings - easily.
What needs to happen is that what many people perceive as a noise floor, but what to me is good ol' distortion at a less obvious level - the villain of the piece - needs to be markedly reduced ... and this is purely electrical behaviour issues ...
Last edited:
Fascinating Chris.
Most people think that Steely Dan made some of the best sounding records ever, and they are around Green 15:
Album list - Dynamic Range Database
Dire Straits in the Green 15 too and sounding good:
Album list - Dynamic Range Database
Supertramp "Crime of the Century" was a great sounding album on Vinyl, Green 16:
Album details - Dynamic Range Database
And Lady Gaga, bless her, makes some hideously poppy compressed recordings. I hate listening to them. Red 7:
Album list - Dynamic Range Database
Seems to work! Compression is BAD!
More on this subject: "Loudness War" Dynamic Range Compression & The DR Database - Observations - diyAudio
Never thought about reproducing live music because it could be live in so many different venues. My goal was always to have total control over the frequencies that appeal to me. Always needed to be able to highlight vocals for example, drive the women crazy with highs, have the base do chest compressions whenever the mood might strike.
Perhaps one more link concerning this subject of preserving the original dynamics in recordings (which I feel is at least as important important as talking about DIY sound reproduction systems and at the same time very under-appreciated by DIYers) by one of the most visible practicing proponents of preserving the "liveness" and emotion of recordings through original recorded dynamics:
Metro Times - Music: The mastering master
Chris
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.