DSP hardware and Software Comparison
by Stu Olson, N7QJP
From: Stu Olson, N7QJP
Well, I got a number of suggestions to put these comments out on the reflector. No problem, here it is. Before I get started, let me state that I have no loyalties (or family connections) to Brian, K6STI (DSP Blaster software developer) or TimeWave Tech. I own a copy of the DSP Blaster software and a TimeWave DSP-9 (recent firmware upgrade to 3.0).
I also recently had the opportunity to borrow a TimeWave DSP-59+ and see how it worked compared to my two DSPs.
Here are my thoughts, in no particular order.
To run the DSP software, you must have a 486 or better machine, VGA, math coprocessor, mouse, and a 16-bit Creative Labs sound card (Sound Blaster 16, Vibra 16, or AWE32). I can not verify proper operation on clone 16 bit sound cards, although I do not think they will work. I am using a Pentium 75 and Win95. The DSP Blaster software is the latest release, version 1.12. There are no special digital modes that I am aware of. It is my understanding that K6STI has other software for this (or is currently working on it).
My first impression of the software was that I did not like it. Why...it was different. I had my DSP-9 for about 18 months and was comfortable using it. The software DSP was new, different, etc. After I had it for a few days, I started to get over that "new" feeling and got into using it. I hooked up my FT-736R so I could do side-by-side comparisons...A/B switching, etc. I wanted to be able to switch back and forth between both DSPs and hear the difference, if any. I ran none of these processed signals through an analyzer, o'scope, or even a cheap DVM (grin). This was all done with my two ears (which were working OK the last time I had them checked... would someone please answer the phone that keeps on ringing!).
One thing I immediately noticed is that the software does not decrease the "clear channel" noise nearly as much as does the DSP-9. By clear channel, I mean that no one is talking...just good old background noise. With my DSP-9, I routinely dial up the calling frequency and then turn on noise reduction. The noise goes away (almost like having squelch without the drawbacks). I spoke with K6STI about this and he gave me a long and quite complete explanation for this....having to do with the way noise cancellation works, that the DSP-9 is also killing part of the weak signal when it does this (sometimes called soft-squelch), etc. I can not prove or disprove his explanation. I am not a DSP engineer. All I can tell you is that if you want don't wish to hear noise when no one is talking, don't get the software DSP.
Heterodyne reduction....they both do a very nice job. I could not really tell any difference between either unit. I went down on 40 meters in the evening to check this out, since I couldn't find that many heterodynes up around 144.200. I did, however, kill a couple of birdies that I found lurking down on 6 meters. As I said before, they both did a good job in this area, without screwing up the desired signal.
Noise reduction on a really weak SSB signal.....I tried this on 2 meters and also on 432 during the contest last weekend. After spending significant time evaluating both, I think the software DSP wins this one by a little bit. There were a several times when I could pick out more words, and when this was the case, the overall sound of the audio signal was more pleasing to listen to. (Of course, I was not running the signals through the same audio amps and speakers either, and this could account for the difference in the sound as well.) I believe the software DSP digs a bit further into the mud and brings up a better sounding signal.
I do not do a lot of CW work, although I spent some time trying out both in this mode. The software version also takes the honors in CW weak signal as well. It has a very nice CW peaking filter, which has an automatic fine tuning option to really tweak the signal. I was able to track a operator with a rather drifty VFO without having to fool with the controls, nice touch!
Low and High pass filters.....they can be controlled separately in the software. On the DSP-9, you must use them as a single function. I liked being able to switch in a 250 Hz HPF while leaving the audio up at 1500 Hz untouched. I found that a lot of nasty noise is taken out when turning on the high pass filter. I used this more than I did the low pass filter. Having two independent controls, with multiple settings each, is much better than the DSP-9 with only a couple of bandwidth selections.
Automatic Gain Control....the software has it, the DSP-9 does not (the DSP-59+ did though). It seemed to work quite well in the software, although I really don't care to use it. Perhaps it would be more useful in a faster paced contest were you have lots of signals with widely varying signal strengths. I prefer to use the volume control myself.
Support....K6STI normally answered my e-mail requests the same day...often times within an hour or two (this was over the holiday season). The DSP-9 folks always answered my e-mail within 1 business day (and even once on a Sunday). So, I guess they both get a high rating here. Nice to see that customer support still means something to some people! FYI...I have already received one update from K6STI, no charge. He distributes them via e-mail attachments (hard to do with the EPROM in my DSP-9).
So...at this time, it appears that the software version is quite a bit ahead, eah? Well, yes and no. One thing that I found that I REALLY HATE is the fact that I can't run the K6STI software in a window in Win95. The computer does not give it enough time, and the software bogs down...distorts, stops processing, complains about interrupts, etc. If I run it full screen, it is fine. It also does not do well at all running as a background process. In defense of the software, let me state that it is written in assembly language and directly accesses the cards hardware. As such, running it in any kind of Windows environment would not be the preferred method to use. It is a DOS based program, and from what the manual says, does best running that way. It can also be loaded as a TSR and is suppose to run well in the background in this manner. (It is only about 35K bytes in size...ah yes, those good old days of compact assembly language code!) Anyway, if you plan on running your Windows based logging program and this at the same time...good luck. Every time you switch over to log a contact, your DSP goes south!
I usually don't take my PC workstation roving with me...so using the software in the Jeep is a no go for sure. However, it is really easy to toss the little old DSP-9 in the front seat and use it. Score a point for the DSP-9!
Bottom line....PRICE always means something here! For the price, you can not beat the software DSP. It sells for $100, and when compared to a hardware DSP, that is quite a bit cheaper. It sure is not portable, unless you have a notebook with compatible sound card to support it. I have had a couple of people come over and use my station, and I fire up the software DSP for them to try. They were very impressed with it...having never used a DSP before. If you are a Windows person, and you MUST use other software at the same time, you may not be very happy. If you are a good old DOS person, this should not cause a problem.
Last comment for the road. This evaluation sure as hell was not scientifically conducted. It was never meant to be. If you have more questions, drop me a note. If you DSP is better than either of mine, good. I won't dispute your claims. If you want to flame about these contents... vent your frustrations on the calling frequency thread and spare me. I wrote this up because several people wanted my comments on this. I hope you have may have gained something from these paragraphs. I wish I had known some of these things before I spent my money (on either one for that matter). Good luck and happy DSPing!
73 and best to all from Phoenix AZ
Stu Olson, N7QJP DM33
Comments: Rein, W6/PA0ZN