Follow along with the video below to see how to install our site as a web app on your home screen.
Note: This feature may not be available in some browsers.
Welcome to ASR. There are many reviews of audio hardware and expert members to help answer your questions.
Click here to have your audio equipment measured for free!
Brightness is related to resolution, but resolution implies other than frequency gains or dips.
Comparing frequency domain of two speakers is as useful as trying to understand a video by comparing two frames.
DSP helps to fix some speaker tendencies, or room modes, but the time domain and the behavior of the speakers in other than the listening spot, is part of the “character” of the speaker.
I got the chance to hear both, the Linton and the SuperLinton in the same spot with the same music and the same gear, and they are quite different animals. To me, the Linton is closer in character to the Dovedale, than the SuperLinton. You could compare the frequency chart of those 3 speakers, but is live auditions, where you notice how the speakers interact to the room.
These "voiceprints/soundprints", just like "fingerprints", would contribute to the audible difference between the two, but I know well many other factors including recording and mastering differences would also more-or-less affect.
Okay extreme case when bassoon and cello hitting the same notes obviously they sound very different. On frequency response measurement alone it would be still similar when they produce the same decibel woudn't it? In this case whats causing the soundprint difference?
Part of my question is to understand if frequency response measurement alone is able to print the full picture evaluating speaker character.
Yes. But we do not have a perfect and cover-all way to measure speakers. But as I said, even though we don't, still all speakers measure differently within the measurements we have available.
Okay extreme case when bassoon and cello hitting the same notes obviously they sound very different. On frequency response measurement alone it would be still similar when they produce the same decibel woudn't it? In this case whats causing the soundprint difference?
You need to learn/understand the essences and meanings of the "voiceprint/soundprint" objective 3D charts; it represents Time (X-axis), Fq (Y-axis) and gain (color scale), ref. chart under the spoiler cover in my post #17.
When the base/main keynote tone Fq is similarly represented on the two charts (for one violin and another), the odd-and-even higher Fq (sometimes lower Fq too) of harmonic tones are different between the two, and it can be objectively seen on the chart!
This is true for not only the two similar violins but also for guitar and lute, for another example.
Still be quite easy to tell the different tone or timbre due to different brand, making, material or the layers of paint. This vilion sound different to that one.
If you would have no interest on learning and understanding the essence of the 3D color "voiceprint/soundprint" charts and meanings of "even-and-odd harmonic tones" thereof, I would have no further advice/suggestion for you. Then it will be the time for me to get out from this thread.
I foundthis WiKi page would fit for your very preliminary learning and understanding on "sound spectrograms" or "voiceprints", i.e. almost identical to "sonographs" and/or "voicegrams".
If you would have no interest on learning and understanding the essence of the 3D color "voiceprint/soundprint" charts and meanings of "even-and-odd harmonic tones" thereof, I would have no further advice/suggestion for you. Then it will be the time for me to get out from this thread.
I foundthis WiKi page would fit for your very preliminary learning and understanding on "sound spectrograms" or "voiceprints", i.e. almost identical to "sonographs" and/or "voicegrams".
Used to play for a bit. When it comes to the shop that sells violins at different price point it's easy to tell the diffs. Like they all been tuned properly. I mean human ears are amazing lol ...
Used to play for a bit. When it comes to the shop that sells violins at different price point it's easy to tell the diffs. Like they all been tuned properly. I mean human ears are amazing lol ...
I can well understand at least your above point on violins, and this is why I mentioned about two of Stradivarius in my post #17!
For me, since I used to play piano a lot, I can rather easily identify the sound differences between Steinway & Sons, Bösendorfer, Yamaha, Kawai, and even Fazioli...
I can well understand at least your above point on violins, and this is why I mentioned about two of Stradivarius in my post #17!
For me, since I used to play piano a lot, I can rather easily identify the sound differences between Steinway & Sons, Bösendorfer, Yamaha, Kawai, and even Fazioli...
If there is one take away today would be "the essences and meanings of the "voiceprint/soundprint"". Frequency response would be much of a 2D footprint of sound where as 3D version explains a lot more.
Exactly what I am looking for from a scientific prespective. If human ears can hear the diffs there has to be some form of measurement to reflect it. And expand it to the different soundprint of loudspeakers, amps tho playing the same track ...
For example in the case of Linton and Super Linton since it's closely related. By bringing up the circled area to match each other's response would they ultimately sound the same? Would anyone be able to tell they are the same or not if frequency response matches by DSP one of them? View attachment 482251 View attachment 482252
I think they would sound very similar. If you click through the various normalized measurements, the directivity is very close (normalization allows the off-axis response to be compared without being compromised by the magnitude of the on-axis response, so it can stand in for the assumption that the on-axis response can be EQed the same).
I think they would sound very similar. If you click through the various normalized measurements, the directivity is very close (normalization allows the off-axis response to be compared without being compromised by the magnitude of the on-axis response, so it can stand in for the assumption that the on-axis response can be EQed the same).
Generally and essentially, I would like to agree with you.
Nevertheless, there could be possible subtle audible difference between the two, even if the Fq-SLP curves would be almost completely the same with each other after applying proper and suitable EQs.
Such subtle difference can be objectively and visually observed in carefully measured 3D color spectrum "voiceprint/soundprint" showing all the even-and-odd harmonics sound in Time(X-axis) - Fq(Y-axis) - Gain(color scale), I believe. Of course, you need excellent measurement microphone (like EARTHWORK M50, I use rather affordable BEHRINGER ECM8000 cross-calibrated against precisely-calibrated M50, ref. here) and its proper identical setting as well as excellent sound recording/analysis system, i.e. excellent ADC and analysis software; I believe ADOBE Audition (ref. #377 for ver.3.0.1) or SOUND FORGE Pro Suite or STEINBERG SpectraLayers or even MusicScope 2.1.0 (free installer for Windows available) will be good enough.
Generally and essentially, I would like to agree with you.
Nevertheless, there could be possible subtle audible difference between the two, even if the Fq-SLP curves would be almost completely the same with each other after applying proper and suitable EQs.
Such subtle difference can be objectively and visually observed in carefully measured 3D color spectrum "voiceprint/soundprint" showing all the even-and-odd harmonics sound in Time(X-axis) - Fq(Y-axis) - Gain(color scale), I believe. Of course, you need excellent measurement microphone (like EARTHWORK M50) and its proper identical setting as well as excellent sound recording/analysis system, i.e. excellent ADC and analysis software; I believe ADOBE Audition or SOUND FORGE Pro Suite or STEINBERG SpectraLayers will be good enough.
I took a look at that thread. I know you are serious and spend a lot of time on your system and comparisons, but I do not think your specific method makes sense. A similar attempt with similar problems is found in Sergei Smirnoff's DF metric, which must be acknowledged as mathematically rigorous, on https://soundexpert.org/vault/dpa.html
To keep it short, the main issue is that this kind of data is more or less uninterpretable. It conflates too much. You will surely find differences in the spectrographic comparisons—that is without question. The main thing ti ask yourself is: are those differences important or audible?
We already have key information in the horizontal and vertical normalized curves and polar maps. These account for the primary sound characteristics. These tell us so much about how the speakers perform, and are so impactful, that everything else, distortion, phase, group delay, falls into a distant second place of importance. Moreover there is a lot of research which helps us sort speaker characteristics by their impact in just this manner.
You will surely find differences in the spectrographic comparisons—that is without question. The main thing ti ask yourself is: are those differences important or audible?
I thank you very much for your kind and comprehensive follow-up!
I too have great respects on all of your points.
My point was/is not whether the differences would be important or not, but my point was/is whether it would be audible or not.
Honestly speaking, I have been actually experiencing many audible cases and many inaudible cases.
In any way, those 3D color spectrum (voiceprint/soundprint) and the FFT Fq-SPL analysis thereof are very powerful tools for objective visual observation of subtle differences (or not) between the two similar recorded sound tracks. You would please find rather intensive and typical example cases here and here.
We already have key information in the horizontal and vertical normalized curves and polar maps. These account for the primary sound characteristics. These tell us so much about how the speakers perform, and are so impactful, that everything else, distortion, phase, group delay, falls into a distant second place of importance. Moreover there is a lot of research which helps us sort speaker characteristics by their impact in just this manner.
On this your specific point, at least I myself, do not always fully agree...
I always would like to observe, in addition to your valid points, the 3D Time-Fq-Gain spectrum (voiceprint/soundprint, showing all the even-and-odd harmonics tones) of the sound from the SP, using some typical tracks of my consistent "Audio Reference/Sampler Playlist", and I would like to compare it with the soundprint of another SP of purchase candidate.
Of course, final subjective comparative listening sessions with my ears and brain at my listening position in my own room acoustic environment would be most critical.
And, please let me repeat and emphasize that Fq-SPL is only one of the various "factors" contributing to the excellency (or not) of a SP system and/or our total audio setup (including our room acoustic environment), as I wrote at the end portion of my post #1,009 on my project thread.
At the end portion of my post #1,009, I wrote: I well understand, however, my Fq-SPL measurements and the results in this post would not fully represent/re-confirm the following aspects of my audio setup in my acoustic environments:
1.0.1 msec time alignments among all the SP drivers (ref. #493, #494, #504, #507)
2. transient (kick-up and fadeout) behavior/characteristic of SWs and WOs (ref. #495, #497, #503, #507)
3. presence and extent of sound reflecting wall(s) and possible reverberation thereof (ref. #498)
4. pros of relative gain (tone) controls in analog domain (ref. #438, #643)
5. favorable effects of wide 3D reflective dispersion of high-Fq ST sound (above ca. 7 kHz) using random-surface hemisphere hard-heavy crystal-glass (ref. #912, #921, #926, #927, #929)
6. size of the “sweet sphere” around my listening position (ref. #926, #927, #931)
7. subjectively felt amazing disappearance of SPs (ref. #520, #687)
8. importance of excellent reproduction of 15 Hz - 40 Hz zone low Fq sound in some (many) music tracks (ref. #782, #588, #591, #641, #650, #63(remote thread) )
9.etc., and so on...
I have already objectively and/or semi-objectively measured and discussed above aspects/features in each of the attached reference posts.
After having completed present intensive re-confirmation of Fq-SPL in SP high level signals as well as in actual room sound at my listening position, therefore, I re-confirmed these other “aspects/features” through careful subjective listening session using my consistent “Audio Reference/Sampler Music Playlist” consists of excellent-recording-quality 60 music tracks selected from various genres (summary ref. #669, actual playlist #670, and my independent thread here).
I always prefer rather subjective re-confirmation of those aspects through my daily very much enjoyable music listening sessions in my listening room, in my acoustic environments. Furthermore, it would be always a little bit hard for me describing those wonderful subjective acoustic impressions by words/sentences, in English and/or Japanese.
I took a look at that thread. I know you are serious and spend a lot of time on your system and comparisons, but I do not think your specific method makes sense. A similar attempt with similar problems is found in Sergei Smirnoff's DF metric, which must be acknowledged as mathematically rigorous, on https://soundexpert.org/vault/dpa.html
To keep it short, the main issue is that this kind of data is more or less uninterpretable. It conflates too much. You will surely find differences in the spectrographic comparisons—that is without question. The main thing ti ask yourself is: are those differences important or audible?
We already have key information in the horizontal and vertical normalized curves and polar maps. These account for the primary sound characteristics. These tell us so much about how the speakers perform, and are so impactful, that everything else, distortion, phase, group delay, falls into a distant second place of importance. Moreover there is a lot of research which helps us sort speaker characteristics by their impact in just this manner.
I thank you very much for your kind and comprehensive follow-up!
I too have great respects on all of your points.
My point was/is not whether the differences would be important or not, but my point was/is whether it would be audible or not.
Honestly speaking, I have been actually experiencing many audible cases and many inaudible cases.
In any way, those 3D color spectrum (voiceprint/soundprint) and the FFT Fq-SPL analysis thereof are very powerful tools for objective visual observation of subtle differences (or not) between the two similar recorded sound tracks. You would please find rather intensive and typical example cases here and here.
On this your specific point, at least I myself, do not always fully agree...
I always would like to observe, in addition to your valid points, the 3D Time-Fq-Gain spectrum (voiceprint/soundprint, showing all the even-and-odd harmonics tones) of the sound from the SP, using some typical tracks of my consistent "Audio Reference/Sampler Playlist", and I would like to compare it with the soundprint of another SP of purchase candidate.
Of course, final subjective comparative listening sessions with my ears and brain at my listening position in my own room acoustic environment would be most critical.
And, please let me repeat and emphasize that Fq-SPL is only one of the various "factors" contributing to the excellency (or not) of a SP system and/or our total audio setup (including our room acoustic environment), as I wrote at the end portion of my post #1,009 on my project thread.
At the end portion of my post #1,009, I wrote: I well understand, however, my Fq-SPL measurements and the results in this post would not fully represent/re-confirm the following aspects of my audio setup in my acoustic environments:
1.0.1 msec time alignments among all the SP drivers (ref. #493, #494, #504, #507)
2. transient (kick-up and fadeout) behavior/characteristic of SWs and WOs (ref. #495, #497, #503, #507)
3. presence and extent of sound reflecting wall(s) and possible reverberation thereof (ref. #498)
4. pros of relative gain (tone) controls in analog domain (ref. #438, #643)
5. favorable effects of wide 3D reflective dispersion of high-Fq ST sound (above ca. 7 kHz) using random-surface hemisphere hard-heavy crystal-glass (ref. #912, #921, #926, #927, #929)
6. size of the “sweet sphere” around my listening position (ref. #926, #927, #931)
7. subjectively felt amazing disappearance of SPs (ref. #520, #687)
8. importance of excellent reproduction of 15 Hz - 40 Hz zone low Fq sound in some (many) music tracks (ref. #782, #588, #591, #641, #650, #63(remote thread) )
9.etc., and so on...
I have already objectively and/or semi-objectively measured and discussed above aspects/features in each of the attached reference posts.
After having completed present intensive re-confirmation of Fq-SPL in SP high level signals as well as in actual room sound at my listening position, therefore, I re-confirmed these other “aspects/features” through careful subjective listening session using my consistent “Audio Reference/Sampler Music Playlist” consists of excellent-recording-quality 60 music tracks selected from various genres (summary ref. #669, actual playlist #670, and my independent thread here).
I always prefer rather subjective re-confirmation of those aspects through my daily very much enjoyable music listening sessions in my listening room, in my acoustic environments. Furthermore, it would be always a little bit hard for me describing those wonderful subjective acoustic impressions by words/sentences, in English and/or Japanese.
What I am interested is under what measurement can tell the difference between guitar and lute when their frequency reponses are almost identital when play the same note. 3D Time-Fq-Gain spectrum method mentioned above seems to be one of them. Apart from that is there any other full spectrum measurement would be as close as how human hearing processes sound signal?
