Hiquphon OW1/OW4 comparison



Introduction

Ever since using the OW1 in my Hyperion design it has become one of my favorite tweeters. Since the Hyperion was a design commisioned by a friend it didn't take me long to use it again in what became the Asterion design which teams the OW1 up with the Vifa XT18WH woofer. This has been and still remains my favorite all around speaker and my reference speaker which I compare new designs to. The OW1 is a 3/4" fabric dome tweeter that has more excursion capabilities than most tweeters which allows you to use lower crossover frequencies than you would normally associate with a 3/4" tweeter. The advantage of a smaller dome is better dispersion at high frequencies. If you go to the Asterion project there are off axis measurements which demonstrate this.

I recently had the opportunity to evaluate the newest tweeter in the Hiquphon line, the OW4. The OW4 is similar to the OW1 except the dome is silver (still fabric, just a different color coating) and the sensitivity is a bit higher. David Ellis (www.ellisaudio.com) is a distributor of Hiquphon tweeters and was looking for people to evaluate the new OW4 tweeter. John Krutke ( www.zaphaudio.com offered to give his objective evaluation of them by taking frequency response and distortion measurements of the OW4 (as well as the OW1). I agreed to give my subjective impressions of the OW4 tweeter with a comparison to the OW1 by adjusting the Asterion crossover to accept them. Go to John's page for frequency response, distortion and cumulative spectral decay plots of these drivers. My subjective impressions of the OW4 relative to the OW1 in my Asterion design can be found in what follows.

Asterion OW1/OW4 designs

Above is the schematic for the original Asterion design using the OW1 tweeter. After measuring both tweeters in the Asterion cabinet I determined what crossover modifications would be required to get a similar frequency response using the OW4 tweeters. The only modifications that were required were to increase the 4uF capacitor to 4.7uF (to compensate for the sooner rolloff of the OW4 tweeter) and to add a 20 ohm resistor in parallel with the tweeter (which attenuated the output and matched the sensitivities). I measured each design with the associated tweeter/crossover swap. The results of the measurements are shown below.
As you can see from the plot above the responses are fairly well matched between these tweeters. The top plot shows the overall frequency response of the Asterion design with each tweeter and respective crossover. The bottom plot shows the relative difference between the output of the two designs. Any peak in this response represents an area where the OW4 version is stronger. I may have been able to match them even better but I felt that it was better to only have to change two components during the comparison so the sound would be less influenced by the differences in the crossover topology. Even with the simple crossover modifications that I made I probably couldn't get the responses to match much better considering the peaks and dips are centered around the center line in the bottom plot and any attenuation or boost of a peak or dip would probably make the nearby dip or peak worse. I was surprised at how well I was able to get them to match considering I only changed one capacitor and added a resistor. Most of the differences occur above 5kHz and you'll probably see this much variation between different batches of the same model of tweeter so it is pretty remarkable that two different tweeters were able to match up so closely.

RJBAudio Harmonic Distortion Measurements

I decided that it was a good time for me to learn how to take distortion measurements with SoundEasy. Keep in mind that this is the first time that I've attempted this so I'm still not entirely confident with my methods but still hope to be able to see a difference between these two tweeters relative to each other. Note that these measurements should NOT be compared with anything else but each other due to the measurement conditions.

For these distortion measurements the respective crossovers were connected to each tweeter while testing. The microphone was placed 6" away from the tweeter which may or may not be close enough for the distortion measurements to dominate any room reflection influences. I worried that placing the microphone any closer might cause interference from reflections off of the microphone but lacking knowledge in this area this is just speculation. Another area of uncertainty while taking these measurements was setting the levels. I tried two different levels. The lower level measurement is equivalent to the speaker output with 1 watt of input power which reflects a sensitivity of 85 dB at 1 meter. I verified this level with a Radio Shack SPL meter. The higher level measurement should be closer to 95 dB at 1 meter but was not verified with my SPL meter. My typical listening levels probably fall somewhere around the lower measurement level, maybe lower at times although transient peaks in the music could boost some tones to the level of the higher level distortion measurement. I don't think I've ever listened at average levels comparable to the higher level distortion measurement with these speakers because they are in a rather small room with the speakers separated by about 6' and a listening distance of about 6' away. After taking these measurements I realized that most distortion measurements are taken based on a level of 90 dB from 1 meter away so my measurements fall on either side of that level. John Krutke mentioned that his distortion measurements at the link above are based on a 90 dB level.

Now before I present my distortion measurements I would like to make some comments regarding harmonic distortion that relate to the frequency ranges of these tweeters. Harmonic distortion occurs when frequencies are produced by a driver other than the fundamental tone sent to the driver and these frequencies occur at integer multiples of the fundamental tone frequency commonly known as the "order" of the distortion. So if the fundamental tone is 1kHz, the second, third, fourth and fifth order harmonics will show up at 2kHz, 3kHz, 4kHz and 5kHz respectively. If we assume that the human hearing range is limited to 20kHz this can help us limit the frequency range where each order of distortion should be noticeable. For example, 2nd order harmonic distortion should only be considered up to 10kHz because beyond this fundamental tone frequency the 2nd order harmonics will be above 20kHz. If we carry this over to other orders, 3rd order harmonic distortion should only be noticeable when induced by fundamentals below 6.7kHz. 4th order distortion and 5th order distortion should only be considered when fundamental tones are below 5kHz and 4kHz respectively. Keep this in mind when viewing the charts below. However, also keep in mind that intermodulation distortion could be produced from these tones that are produced beyond the typical human hearing range but my lack of knowledge in this area prevents me from elaborating on this.

Below is a plot of the harmonic distortion at the lower level of 85 dB at 1 meter. Note that the fundamental is approximately 10 dB lower than that of the higher level distortion plot so all distortion levels are actually 10 dB higher if you want to compare them with the next set of plots. The plot below shows the harmonic distortion of the OW1. Move your mouse over the plot to see the distortion of the OW4.



Below is a plot of the harmonic distortion at the higher level of 95 dB at 1 meter. As noted in the previous paragraph the fundamental is 10 dB higher in this plot so you'll have to scale everything down by 10 dB to compare with the above plot. The plot below shows the harmonic distortion for the OW1. Move your mouse over the plot to see the distortion of the OW4.

Listening Impressions

First I will discuss the methods used to compare the two tweeters. The primary method was to install one set of tweeters in the Asterion cabinets along with the associated crossovers and then manually switch the tweeters and crossovers which takes about 5 minutes total. The second method was to place the left and right speakers side by side with the OW1 in one speaker and the OW4 in the other speaker. I sent a mono signal to the speakers and used a relay switchbox to instantaneously switch between them. In order to confirm that the differences I heard weren't related to differences in the crossovers or woofers in each speaker I swapped the tweeters in this configuration and verified that the differences moved with the tweeters. Although I was able to hear differences with the time delayed swap of the stereo comparison I found that it was easier to define these differences once I compared the tweeters side by side with an immediate switch between them. So when I went back to the stereo comparison I had a better idea of what to listen for during the time delayed swap of tweeters/crossovers but had a better impression of how they compared in a real application. I spent about a week comparing the two tweeters and then my friend came over to compare them. My friend has a much better "acoustic memory" than I do and didn't need to do the instantaneous switch mono comparison to define the differences that he noticed between them. While I used several dozen CD's during my comparison of the tweeters I chose two specific tracks that seemed to expose the differences rather well when my friend evaluated them.

I'll start with my impressions of the tweeters with one in each Asterion cabinet evaluated side by side with an instantaneous switch. The first thing that I noticed was that the midrange character with voices and certain instruments was quite different between these two tweeters despite how similar they are in the lower range based on the on axis frequency response measurements shown above. The OW4 seemed to have more presence in this region despite the frequency response measurement revealing a slight dip around 4-5kHz. In comparison the Asterion with the OW1 had more of an emphasis on the mid to lower midrange which I attributed to less presence in the upper midrange. This was when I decided to swap the tweeters to make sure that differences in the speakers other than the tweeters weren't influencing my impressions. When the tweeters were swapped the differences moved with the tweeters so it was safe to assume that the differences that I heard were due to the tweeters and not other factors. I next disconnected the woofers to compare the tweeters by themselves and noticed that vocals were more intelligable with the OW4 due to this extra presence. In this design I actually felt that this extra presence made the sound more neutral because the Vifa XT18 drivers tend to have a slightly warm, bloated midrange sound. However, the downside was that the Asterion design with the OW4 had a more forward sound while the OW1 had more depth.

Under the same conditions I compared the rest of the frequency range above the region previously discussed. The main thing that I noticed was that it was easier to distinguish sounds with the OW1 and the depth of the presentation (even in the mono configuration) was greater which enhanced this aspect. Once again the OW4 had a more forward, in your face sound relative to the OW1 and when there were several different types of high frequency producing instruments like cymbals it was more difficult to distinguish them with the OW4. Some of this could be due to the differences in frequency responses between the two tweeters which are more prominent in this frequency range but some of it could be due to the differences in the lower end of the tweeters' range discussed in the previous paragraph.

Next I'll discuss my impressions with the speakers set up in my normal stereo listening configuration and the same tweeter model installed in each speaker. Similar conclusions were drawn in this comparison but I got a better idea of which one I might prefer in the Asterion design. The added upper midrange presence of the OW4 was appealing to me because I felt that it actually sounded a bit more neutral in this design but did have some negative side effects as well. As I said before this extra presence seemed to provide more compensation for the slightly bloated lower midrange that the XT18 drivers suffer from and increased the clarity of vocals. Some acoustic guitar recordings also sounded a bit cleaner to me due to this while the sound with the OW1 could sometimes seem a bit more muted and resonant except the extreme high guitar harmonics were still present but with greater emphasis not because they were stronger but because the range of frequencies just below them had less emphasis. I noted similar things with vocals where the stronger upper midrange presence of the OW4 sounded more neutral to me while the OW1 sounded like there was something missing in this range which made the low-to-mid midrange seem stronger as well as the extreme vocal harmonics. In subjective terms the OW1 seemed to have less edge and breath in the vocals while the higher frequency sibilant tones were just as strong as those with the OW4 but were more noticeable because of the reduced presence in the frequency range just below this. This was most noticeable with female vocals. However, the consequence of this greater upper midrange presence with the OW4 was a more forward, less smooth midrange sound.

At the extreme top end the OW1 seemed to have a bit more sparkle than the OW4 but some of this could be influenced by less presence in the low end of the OW1 relative to the OW4. Once again the OW4 seemed more balanced and neutral because of this. However, the OW4 seemed to have less distinction between instruments in this upper range both spatially and sonically. The bottom line was that the Asterion design with the OW4 seem more forward in its presentation while the OW1 implementation seemed to have a more three dimensional sound with more distinction between instruments. This was one area where I clearly preferred the sound of the OW1 because it had more clarity with these high frequency sounds due to both spatial and tonal distinction between instruments.

I'll briefly discuss a theory as to why I might be hearing some of the differences in the lower range of the tweeters despite them not showing up in the frequency response plots. I suspect that some of this could be due to distortion but am not entirely confident. If you look at the distortion plots above you'll notice that the OW4 has more 3rd and 5th order distortion in the 3kHz to 5kHz range and I'm guessing that these differences could increase the presence of the OW4 in that range. Although my design has a crossover point above 2.5kHz it can be seen in the 95dB level distortion measurement that the OW4 should be able to handle a lower crossover frequency with less consequences from distortion. The thing about speakers is I don't think there will ever be enough measurements to adequately predict or justify the differences that your ears notice and it is easy to misinterpret measurements when trying to translate them into listening impressions.

Second Opinion

In order to give a more balanced review of these tweeters I invited my friend Pete D. over to compare them. He is very familiar with my Asterion design and I trust his evaluation skills when it comes to speakers because he is very good at picking out differences especially after a time delayed swap between systems. Rather than try to interpret our discussions after the comparison I asked him to email his review to me. Note that I didn't share any of my opinions of the tweeters with him prior to him listening to them and comparing them. Below are his comments...

On Grey Eye Glances' "Float" the extended treble was stronger on the silver-domed tweeters (I'll assume that they're the OW4's from here on), but the really significant difference was in the midrange, especially the vocals. I thought that the OW4's had less "throat" and more "edge" and "breath" to the vocals as compared to the OW1's. But it's the OW1's that are unique in that respect: the OW4's have a more typical west-coast tonal balance. I can recall that for some vocals (like Natalie Merchant) the OW1's have too much midrange, but I liked them a lot on "Float".

On Supertramp's "Breakfast in America" the OW4's were really harsh to the point that the upper midrange detail was totally obscured when there was a lot of energy there: they were awful on the cymbals (you couldn't hear any of that tinkley sizzle; it was just noise, and if it was louder it would have hurt). The piano sounded really flat, no "body" in the midrange. The trombone was OK, but it didn't have the life that it did with the OW1s. The sax and guitars were about the same for either tweeter, with a bit more high-end detail for the OW4s. The vocals had a bit more "throat" with the OW1's, but that wasn't as profound as it was for "Float".

In general I liked the OW1's better, even though the OW4's had more extension and detail in the highs. The OW1's seemed to compliment the woofers better for the cuts we listened to, but for some music that could be considered too euphonic in the midrange. I think that the OW4's have significantly more Intermodulation distortion in certain parts of the midrange, and that's what's making them sound more harsh in some cases, and more "detailed" in others. The OW1's midrange "bloom" doesn't show up in your data at all, but it's quite noticeable (and pleasant) to listen to in most cases, and it might be dependent upon the crossovers and woofers.

OW1 vs. OW4 Conclusions

I'll start by saying that subjective impressions of speakers can be very difficult to write as well as interpret because different people have different concepts of what certain subjective terminology means. Personally I found it difficult to determine which frequency ranges were influencing the sound when comparing tweeters because I'm usually more concerned about the midrange region when tuning crossovers. You'll notice that in some ways my impression of the tweeters contradicts Pete's impression partially due to the differences in preferences and partially due to differences in our interpretation of what we heard. The bottom line is that we both heard differences that weren't revealed by the frequency response measurements of the two tweeters. This is an important thing to note outside of the scope of this evaluation. Even if you equalize two drivers to have the same exact frequency response it is very likely that they will still sound very different. The fact that the design of these two tweeters is so similar regarding dome size and type and the fact that they are from the same manufacturer makes this argument even more compelling.

Also, keep in mind that this evaluation is based on the application of the Hiquphon tweeters in my Asterion design which can influence the preference based on the other factors of the design outside of the tweeter's range. If a design was used that had a lower crossover point or required a higher sensitivity tweeter the OW4 might be the clear winner. Also, even though my friend preferred the OW1 in the Asterion design I felt that some of the negative aspects in his mind led to a more neutral sounding speaker when combined with the XT18 but still felt that each tweeter had its strong and weak points. As with most aspects of speaker design there are a lot of tradeoffs involved and it really comes down to which benefits you prefer and which sacrifices you can live with.

I still feel that both tweeters are high on my list of favorite tweeters but some of the things that I like about these tweeters could easily be criticized by someone who has a different taste in tweeters which could be based on music preferences or other factors. The bottom line is that you shouldn't let measurements or somebody else's opinion of a driver persuade you into thinking you will or won't like it. Always trust your ears when making these decisions because they are the only thing that won't lie.

Extra Innings... OW1-FS evaluation

After completing the OW4 evaluation Dave asked me to check out the OW1-FS tweeter. The box of tweeters that was sent to me included two OW4's, one OW1-FS and one OW1. Initially I hadn't planned on evaluating the OW1-FS because I didn't have a pair to work with. However, after realizing that I found it easier to pick out the minute differences between tweeters using my relay switchbox for an immediate switch between one Asterion cabinet with one tweeter model and the other cabinet with the other tweeter model I decided to give it a shot. Keep in mind that during the OW4 evaluation above and this evaluation I used the OW1's that I bought from Dave a few years ago and not the one that he sent. This time I evaluated them first with the OW1-FS in the left speaker and my original OW1 in the right speaker with the speakers placed side by side so both would be influenced by the extended baffle width. I later reversed the tweeters and continued the evaluation and noticed that the notable differences between the tweeters moved with the tweeters.

Measurements

I decided to take some measurements of each tweeter using the same speaker (cabinet/woofer/crossover/etc.) and just swapping the tweeter between measurements. These measurements use a slightly different gate time and maybe some small differences in level settings and mic distance than the measurements shown above so there may be some minor differences despite the fact that the same cabinet, crossover, and OW1 tweeter was used in these measurements. However, these measurements will provide an accurate representation of the differences between the OW1 and OW1-FS in the Asterion design without any crossover changes. The comparison is shown below.
The top plot gives the total response of the Asterion design with each tweeter installed and using the same crossover with no modifications for the OW1-FS. The bottom plot shows the relative difference between the two responses. The regions where this trace is above zero represent the areas where the OW1-FS is stronger. As you can see from these plots the response differences between the OW1 and OW1-FS are greater than those between the OW1 and OW4 shown further up the page. This should be kept in mind when you read my listening impressions below because I feel that these response differences were reflected in the comparison of the tweeters.

Listening Impressions

Like the OW1 vs. OW4 comparison I was able to hear a difference between the tweeters in this evaluation. However, the differences seemed to correspond more closely to the frequency response differences and there were some similarities in other aspects. I'll start with the similarities. Both the OW1 and OW1-FS had a similar spatial presentation and both seemed to have similar depth to the presentation unlike the OW4 which had a more forward sound. However, it was more difficult to distinguish different instruments with the OW1-FS than the OW1 because they seemed to blend together a bit more. I have a feeling that some of this has to do with some of the other differences that I will discuss next.

There was definitely a different character to the midrange sound between these two tweeters. The OW1-FS sounded much more laid back in the upper midrange than the OW1. This was noticeable with vocals where OW1 had more breath and edge to the vocals. Also, the higher frequency sibilant tones were less pronounced with the OW1-FS. This could also be heard with acoustic guitar where the strumming, plucking and sliding across the strings sounded more muted with the OW1-FS. Unlike the OW4 where it sounded like there was something extra in the frequency range near and below this the OW1-FS just sounded like it was missing something here which is exactly what the frequency response comparison reveals. Now the question is whether a 0.5 dB difference in the 3kHz to 7kHz range is enough to justify the differences that I heard in that region. Since it is such a wide range of frequencies (over an octave) it is more likely that a small difference in output would be more noticeable. Since the XT18 already has a slightly warm midrange sound I preferred the more detailed midrange sound with the OW1 in this comparison. I'm sure you might be able to adjust the crossover with the OW1-FS to compensate for this but I'm not sure you could get them to sound the same as the OW1. In general the OW1-FS sounded more veiled in the midrange region compared to the OW1.

I noticed similar differences in the higher frequencies between these two tweeters. Cymbals seemed to have a crisper sound with more sparkle with the OW1. This was very present with tambourines which sounded a bit laid back and muted with the OW1-FS. Ride cymbals sounded similar in their fundamental tones but the fading shimmer that followed was more present with the OW1. Similar to what Pete noticed above with the OW4, crash cymbals sounded a bit more like noise with less of the tinkly sizzle (to borrow his words). The bottom line is there was more definition with the OW1 in this range. It is very probable that these differences could be explained by the stronger output of the OW1 above 12kHz. However, the differences were similar in nature to those in the midrange region and the OW1-FS sounded more veiled across the entire range.

If I had to pick one word to describe the OW1-FS compared to the OW1 in this application I would have to say "veiled". The OW1 could also be considered as sounding a bit more dynamic as well but that could be a misinterpretation of what I heard. I definitely prefer the sound of the OW1 along with the XT18 with my crossover because I feel that it compliments the smooth lower midrange by adding some edge and detail while the OW1-FS lets the slight lower midrange resonance dominate a bit more. I prefer the crisper sound of the OW1 in general but some of that is personal preference. Some people might find the OW1 to have too much sparkle for their tastes. It reminds me of a comparison I did a while back with the Vifa XT19 tweeter where I thought that the XT19 lacked some of the top end sparkle that the OW1 has. I've also found that a lot of metal dome tweeters that I've used seem to have even less of this sparkle than the XT19. I suspect some of the differences between the OW1 and OW1-FS can be justified by the differences in frequency response with the Asterion crossover but I have no way of telling if the OW1-FS response can be compensated to provide a similar sound as the OW1 or if some of the differences are inherent to the physical differences between them which involves the use of ferrofluid in the OW1-FS. I don't think this evaluation is thorough enough to draw those conclusions due to the fact that some of the differences that I noticed seemed to coincide with the measured differences in frequency response.

Final Conclusions

As I said above in the OW4 conclusions subjective impressions can be easily misconstrued. It is also different to convey how great the differences I heard between the tweeters evaluated in this study. I am convinced, however, that comparing two similar items with an instantaneous switch between them allows for a more critical evaluation of minor differences which can make them stick out a bit more. This is especially true with me because I don't have to rely on my "acoustic memory" or associated psychoacoustic influences. However, long term listening sessions do have their value as well when considering which tweeters might be preferred.

One of the issues that was discussed above, mostly with the OW1-FS comparison, was whether the differences I heard were due to frequency response issues and if they could be equalized out. This is one area where I am very uncertain. However, production variances with tweeters can lead to differences of this magnitude depending on the quality control of the manufacturer. I've found that this is especially true with fabric dome tweeters above 10kHz. The nice thing about the Hiquphon tweeters is that Oskar sells them in pairs matched within 0.5 dB. The point that I'm trying to make is that if the differences that I'm hearing are solely due to frequency response differences then just buying a different batch of the same model tweeter could lead to a speaker design having a different character on the order of the impressions I made between these tweeters. I have a hunch that there is more going on with these tweeters than the frequency response indicates and this seemed more prevalent with the OW4 comparison because the differences I heard didn't exactly coincide with the differences in frequency response.

It seems that this evaluation may have led to more questions than answers and you should keep in mind that these are my impressions of these tweeters in my design. I suspect that if I had started out with the OW4 or OW1-FS while designing the Asterion crossover I might have tweaked it a bit differently. However, this has given me more confidence that I got the Asterion crossover right with the OW1 based on my listening preferences and still think the OW1 is my favorite tweeter based on its performance in this design. All of the Hiquphon tweeters that I evaluated share some fundamental similarities and these similarities seem to match well with my preference in tweeters. I believe that there is no such thing as a perfect speaker or driver and it is all about choosing components that compliment each other and have qualities that suit your tastes. It seems like most drivers have good and bad qualities both from a technical and a subjective perspective. I think the key is to let your ears and not somebody else's judge whether a certain driver is good or bad. This tweeter evaluation still has some value in revealing some of the relative differences I noticed between these tweeters and may give you insight into which one you might prefer but should not lead to conclusions as to whether you would prefer any of these tweeters over another tweeter not included in this evaluation. As I said before you should choose the tweeter that compliments the other drivers in the system and suits your tastes.

I would like to thank Dave Ellis for allowing me to evaluate these tweeters and learn quite a bit in the process. If you decide you would like to try any of the Hiquphon tweeters I highly recommend buying them from him at www.ellisaudio.com. Any comments or questions regarding my subjective impressions of these tweeters or the Asterion design can be sent to my email address under the "About Me" section of my site.