A Subjective Blind Comparison of 2in to 4in drivers - Round 4

Select the driver that you think sounds the best.

  • A

    Votes: 10 24.4%
  • B

    Votes: 5 12.2%
  • D

    Votes: 3 7.3%
  • E

    Votes: 11 26.8%
  • F

    Votes: 12 29.3%

  • Total voters
    41
  • Poll closed .
Founder of XSA-Labs
Joined 2012
Paid Member
Any FU10RB in this winner list ? Already made eslwhere ?

Never had a chance to use an FU10RB - but will gladly test it and add to next round if someone provides me one.

Nice looking driver, reminiscent of a ferrite motor 10F, if such a driver where available, or should I say, an alloy framed TG9FD?

Good specs too: https://www.madisoundspeakerstore.com/approx-4-fullrange/seas-prestige-fu10rb-h1600-04-4-full-range/

h1600-04.jpg


As I seem to be getting a lot of questions if a certain driver was part of the original list of entrants, perhaps it would be useful to list the entire set of drivers from all 3 rounds that these drivers came from.


I updated Post 1 to include info below:


Here is a list of all drivers tested in all 3 rounds leading up to this finale. We are now auditioning and selecting the top 3 (plus 1 that I selected) from a total of 21 drivers. In general, the driver with the flattest response has taken the top spot - so now we are listening to the best of the best in terms of flat frequency response so it is not surprising that it is hard to hear a difference as that is the principal means that our ears differentiate sound voice signature.

Round 1: Winner = TC9FD

1. Vifa TC9FD (3.5in paper cone 8 ohm)
2. Faital Pro 3FE22 (3in paper cone 16 ohm)
3. Dayton Audio RS100-4 (3.5in aluminum cone 4 ohm)
4. Dayton Audio RS100P-4 (3.5inpaper cone 4 ohm)
5. Dayton Audio PS95-8 (3in paper cone 8 ohm)
6. MA CHN-70 (4in paper cone 8 ohm)
7. PRV 5MR450NDY (5in 'midrange' pro audio paper cone 8 ohm)

Round 2: Winner = 10F/8424

8. Vifa TG9FD10-8 (fiberglass cone)
9. Tang Band W4-1320SB (paper cone)
10. Peerless P830986 (aluminum cone)
11. Fostex FF105WK (paper cone)
12. ScanSpeak 10F/8424G00 (fiberglass cone)
13. Mark Audio Alpair A7.3 (silver cone) (aluminum cone)
14. AHE 3in (from Taiwan) (paper cone) http://m.ruten.com.tw/goods/show.php?g=21206258966085

Round 3: Winner = B80

15. SB Accoustics SB65WBAC25-4 (2.5in aluminum cone Nd magnet)
16. Peerless P830983 (2in aluminum cone Nd magnet)
17. Tang Band W3-1364SA (3in bamboo cone Nd magnet)
18. GR Research LGK (3in paper cone ferrite magnet)
19. Mark Audio Alpair A7P (4in paper cone ferrite magnet)
20. Visaton B80 (3.5in paper cone Nd magnet)
21. Visaton FRS5X (8ohm 2in paper cone ferrite magnet)
*REPEAT* Mark Audio Alpair A7.3 (aluminum cone ferrite magnet - previously tested in Round 2)
 
Last edited:
Founder of XSA-Labs
Joined 2012
Paid Member
BMS,
You are voting for most accurate vs what you think sounds best? You are making quick work of this. Very tough for me. I am glad that clip 1 seems to do the weed out trick. It was mentioned in Round 3 that ACDC track was key for easy selection. It's interesting how rock over drive guitar which is full of distortion in the source is key for picking a driver that sounds accurate. :)
 
Would like to try to vote as I have a 5" (or 4" if measurement is without the middle of the suround rubber: just the cone) aluminium LR2 from 125hz to 2600hz then a 1" aluminium dome (unluckilly no Stax headphones)

Any flac or wav ? I clearly hear the 320 mp3 as climbing in the mid-highs on my system when I rip from a 16/42 brand new cd (bit acurrate) in mp3 320 instead flac/wav ?

mp3 320 has no impact on this FR ?

Is there an average Sd for the 4" drivers, please ?

(Out off topic in relation to your red advert, but for later if generous members could send it to you : http://www.visaton.com/en/chassis_zubehoer/tiefmittelton/ti100_8.html )

at least very interrested by the result to come... do you give your own vote after ? (without any added impedance, cool to have your feeling of a direct listening sessions while there is no easy ABX if you made it alone ! I mean after all the vote! Why ? It is said than placebo effect works also when you say before (and after) to the subject is trying a placebo. Rephrasing it , it could be fun you make a further test with the same winner of the tests 3 + the second winners of each list in a further test (without saying which one if test 5 or 6 or 7) to see if there is a concistency in the result)

In medecine, there is a concrete result with placebo because there is a real production of chimical substance of the body without added chemical (because placebo). As we know perfectly ears and eyes are looking for Something when the subject has an information to influence him, it could be interresting... I mean if if you want of course !
regards
 
Last edited:
A most interesting round. These drivers all sound very different.

In particular the room response exemplifies how differently they are behaving.To demonstrate this each track of clip-1 was mixed down to mono. All tracks were brickwall high pass filtered 550Hz. All tracks displayed in multi-track view with Cool Edit, and time aligned to within one sample. 231 milliseconds of tailing response at end of tracks were selected. Each selection was faded in and out with identical fade functions. Each faded selection was then normalized to -6dB peak. Each processed selection was then pasted seven times into a new track such that when played 7 repetitions of tailing room response of speaker A,B,D,E and F are heard.

Track was saved as wave file and is attached as tails.asc; change to .wav for listening.

With time aligned tracks, identical segments may be highlighted and looped and played with track solo button. With different segments different speakers seem to sound more like reference track. With this observation combined with above observation it appears that speaker directivity at different frequencies plays a big part in listening perception.
 

Attachments

  • Tails x7.asc
    1.5 MB · Views: 59
Founder of XSA-Labs
Joined 2012
Paid Member
Would like to try to vote as I have a 5" (or 4" if measurement is without the middle of the suround rubber: just the cone) aluminium LR2 from 125hz to 2600hz then a 1" aluminium dome (unluckilly no Stax headphones)

Any flac or wav ? I clearly hear the 320 mp3 as climbing in the mid-highs on my system when I rip from a 16/42 brand new cd (bit acurrate) in mp3 320 instead flac/wav ?

mp3 320 has no impact on this FR ?

Is there an average Sd for the 4" drivers, please ?

(Out off topic in relation to your red advert, but for later if generous members could send it to you : TI 100 - 8 Ohm )

at least very interrested by the result to come... do you give your own vote after ? (without any added impedance, cool to have your feeling of a direct listening sessions while there is no easy ABX if you made it alone ! I mean after all the vote! Why ? It is said than placebo effect works also when you say before (and after) to the subject is trying a placebo. Rephrasing it , it could be fun you make a further test with the same winner of the tests 3 + the second winners of each list in a further test (without saying which one if test 5 or 6 or 7) to see if there is a concistency in the result)

In medecine, there is a concrete result with placebo because there is a real production of chimical substance of the body without added chemical (because placebo). As we know perfectly ears and eyes are looking for Something when the subject has an information to influence him, it could be interresting... I mean if if you want of course !
regards

Average Sd is about 36cm2

I vote after listening to tracks without thinking about which letter it is. I am test administrator so tough for me to be blind. But I do my best to vote objectively.

My original plan was to have winner and second place from each round but that is just too many drivers.

320k MP3 is quite good and probably may be a factor when we are trying to separate the creme de le creme apart. FLAC or wav would of course, be better but they are too large to upload to DIY Audio servers. I don't want to place them in a separate file server because the link will eventually be broken and the clips won't be here in perpetuity for others after the poll is over.

If there is enough interest in FLAC files maybe someone can offer to host on their website and I can post link.
 
I can also guess I won't pick 10F. Some time ago I made remark about a Reed/Vaughn song, and 10F having more detail in playback than original I own multiple copies and listened to for many years, on too many occasions. That was weird.
On one of these new soundclips 4 out of 5 speakers sound more detailed than reference clip, the one I would prefer the least, sounds like reference clip the most. Only one close to reference, the others sound too good, weird again, this time I can't say I have reference clips myself to check, but weird choice it is, picking the worst sounding speaker. Will postpone voting till I have rechecked couple of times:)
 
Average Sd is about 36cm2

I vote after listening to tracks without thinking about which letter it is. I am test administrator so tough for me to be blind. But I do my best to vote objectively.

My original plan was to have winner and second place from each round but that is just too many drivers.

320k MP3 is quite good and probably may be a factor when we are trying to separate the creme de le creme apart. FLAC or wav would of course, be better but they are too large to upload to DIY Audio servers. I don't want to place them in a separate file server because the link will eventually be broken and the clips won't be here in perpetuity for others after the poll is over.

If there is enough interest in FLAC files maybe someone can offer to host on their website and I can post link.

Ok I will listen tommorow Sunday the files to be on pair with the test rules. (But I really hear a difference - a subjective airier and climbing mid-treb with mp3 320 with my system - : it's not if chantilly was the same as English creme ! )

you have in USA, many way to upload & share files with simple links you can publish everywhere: mail, forums, etc... though if not on diyaudio "server" ("":because now all is virtualised : sever , disks, networks... sorry for the off topic) it will not as simple with concistency in time : choose your poison : validity or concistency in time :) )
 
Last edited:
Founder of XSA-Labs
Joined 2012
Paid Member
A most interesting round. These drivers all sound very different.

In particular the room response exemplifies how differently they are behaving.To demonstrate this each track of clip-1 was mixed down to mono. All tracks were brickwall high pass filtered 550Hz. All tracks displayed in multi-track view with Cool Edit, and time aligned to within one sample. 231 milliseconds of tailing response at end of tracks were selected. Each selection was faded in and out with identical fade functions. Each faded selection was then normalized to -6dB peak. Each processed selection was then pasted seven times into a new track such that when played 7 repetitions of tailing room response of speaker A,B,D,E and F are heard.

Track was saved as wave file and is attached as tails.asc; change to .wav for listening.

With time aligned tracks, identical segments may be highlighted and looped and played with track solo button. With different segments different speakers seem to sound more like reference track. With this observation combined with above observation it appears that speaker directivity at different frequencies plays a big part in listening perception.

You lost me here. Not sure what all your processing was supposed to do but do recall one of the requests from round 1 was not to use a digital audio workstation to de-construct the sound tracks and compare side by side. Converting to mono is fine but I am not sure what you gained by putting brick walls and tails with 7 copies? To amplify the off axis content that then makes its way back to the mic via room reflections?
 
Founder of XSA-Labs
Joined 2012
Paid Member
I can also guess I won't pick 10F. Some time ago I made remark about a Reed/Vaughn song, and 10F having more detail in playback than original I own multiple copies and listened to for many years, on too many occasions. That was weird.
On one of these new soundclips 4 out of 5 speakers sound more detailed than reference clip, the one I would prefer the least, sounds like reference clip the most. Only one close to reference, the others sound too good, weird again, this time I can't say I have reference clips myself to check, but weird choice it is, picking the worst sounding speaker. Will postpone voting till I have rechecked couple of times:)

Interesting how you say it sounds better than original in some cases. There is a bit of room gain which boosts the bass a few dB - that may change overall perception compared to reference. But that bass boost is same for all. The extra detail could be due to RESA - resonance enhanced selective amplification. A term I coined to describe extra detail generated from resonance peaks which cause ringing. Ringing makes certain frequencies persist longer and are thus more audible. If ringing occurs in certain bands where we expect detail - cymbals, snares, high hats, etc there may be some enhancements. Most of these drivers have very little ringing. They do show some low level energy storage in the waterfalls though.
 
Interesting how you say it sounds better than original in some cases......

.....The extra detail could be due to RESA - resonance enhanced selective amplification. A term I coined to describe extra detail generated from resonance peaks which cause ringing.....

Not that i suspect you haven't concentrated best possible when making settings to setup :), but not absolut perfect acoustic slopes for both LF and HF part at XO point that replicate their textbook targets plus if little error in time offset for acoustic center will probably provoke low level details not heard before, but maybe more dangerous with higher XO point than in this test, else agree the RESA stuff.

Below is blacksmith five different hammer on the anvil when looking into Barleywaters shared file with free Audacity, took the first knock of the seven and united them side by side :).
 

Attachments

  • Barleywater.PNG
    Barleywater.PNG
    86.8 KB · Views: 484