2006.02.26 — "Weeded" this section of the site, removing reviews for products that no longer exist. Also removed some other fully obsolete material. If a product is still available, the technical parts of the previous review stand unchanged. The analysis parts did have to change in places where they were based on comparisons to products that are no longer available. Also, some product prices changed, which further changed the conclusions.
Updated prices and URLs for products that remain available.
2001.12.05 — Updated the PVR review: I test the new v1.5 driver which adds SVCD and WinXP support, I corrected a few problems with the article, and I cover how to get the advanced settings dialog back (they took it away!). More important than all of that, however, I finally have some decent frame grabs that truly show what the board is capable of, instead of merely highlighting my inability to deinterlace the PVR's output streams! These improvements added another point to the PVR's final score, causing it to jump up one slot in the ratings, ahead of the primitive BMK.
2001.11.15 — Vitec Multimedia released a major upgrade to their DVD Cut Machine package, so the review has been updated to cover the improvements. The improvements were large enough to require a rebalancing of the scores, causing the relative rankings of some of the encoders to change.
2001.10.25 — Tweaked the scoring system to penalize unusable products more. Consider a product that scores a '0', because it is completely unusable. Previously, that would only count for 30% of the score, so the product could still score a respectable 7.0 if everything else was perfect. Now the usability score counts for 50%, so the top score for an unusable product is 5.0. Also notice that a score of '0' is legal again.
2001.09.13 — I have changed my software encoder testing methodology. Instead of using a HuffYUV-coded AVI, I'm now using an MJPEG AVI. After assuring myself that I couldn't tell the difference between the two AVIs even when examined frame by frame, I decided to do this because it speeds most encoding tasks up significantly. The simple reason is that PICVideo's MJPEG codec is faster than HuffYUV. Most encoders improved markedly. The only exceptions are bbMPEG which didn't change much at all, and Studio DV which is now measured as slower. However, the latter is probably because I also upgraded to a new version in the interim.
By "can't tell the difference", I mean that the MJPEG AVI — which used the codec's lowest compression level — had only insignificant differences with respect to the HuffYUV-coded original. When considered as still images, these differences were so minute as to be literally insignificant. In full motion, these minute differences can't possibly be detected.
Another good reason to change over to a lossy codec is that most people will be using a lossy codec of some sort anyway. Lossless codecs are merely a nice conceit, not an accurate reflection of the way video is handled in the real world.
I've also changed over to reporting encode rates in terms of frames per second, rather than as a fraction of real time or as an absolute time. This makes comparing the offline encoders to the real-time encoders more direct. Also, it shouldn't be long now that we'll see software encoders that can encode faster than real time on the fastest single-processor PCs. It's time to shift to a measure that makes it clear when encoders begin outpacing the real-time hardware encoders.
2001.08.23 — Major work to the Summary Tables page. First, I split the features table into two tables, one for software encoders and another for hardware and hybrid encoders. If you need hardware for video capture, a software encoder won't do, no matter how good it is. Similarly, if you only need a software encoder, why buy hardware? Second, I added "scores" tables: instead of visiting each page to see each encoder's detailed scores, you can see them all in one table, and compare them. This made it obvious that the old scores were broken; one category had a range of only 6-10, for example. Now, every category has a full 1-10 range, and the scores in between are relative to these "anchor" scores. I am now happier about the relative rankings of the encoders. The scores match my overall impression of the encoders, which is exactly what you'd expect.
Major work on the MPEG Maker, DVD Cut Machine, CCE SP, MegaPEG and LSX-MPEG reviews. I have learned things about all of these encoders since the last update to each.
2001.08.22 — Reworked the Ligos LSX-MPEG Premiere plugin review after receving an unlocked version. I discovered many inaccuracies in the review now that I no longer have the demo version's limits. Previously, I thought the standalone version to be better, but now I see that there are many situations where the plugin is far superior. Read the review for details!
2001.08.11 — Updated the PVR review, noting that the board can be used on low-end Pentiums; this in spite of the fact that Hauppauge demands at least a Pentium II-300 on their web pages.
2001.08.07 — Updated the TMPGEnc review. It now covers version beta 12i, up from beta 12f. Also, I now report its speed using the "normal" motion search speed, which runs twice as fast as the slowest speed. I found that the slow speed doesn't help all that much, and I have been using other adjustable encoders' normal speeds, too, rather than their slowest speeds. I also go into more detail on TMPGEnc's various encoding modes.
Updated the WinTV PVR review. I re-did the capture with Hauppauge's 1.4a drivers (rather than the beta 1.4 drivers I used before) and did a better job of deinterlacing the frame grabs, so now it's possible to compare the frame grabs with those of other encoders. Lots of general text fixups as well.
Updated the All-in-Wonder review. I have now tested the lowest-end current AIW board, the PCI AIW 128, and have found that it performs just as well as the top-end AIW Radeon for MPEG encoding. I'm sure the 128 Pro, between these two boards in the product line, performs well, too. Also did lots of general updates to this review, with the things I've discovered since the previous update, including new info on CPU usage and encoding quality.
I have discovered that Ligos' GoMotion technology is behind the ATI All-in-Wonder real-time encoder, Matrox's low-end encoders, Ulead's video editing products, and Ligos' own LSX-MPEG Premiere plugin. (This explains a lot. :)) Updated the LSX-MPEG review with these and other observations.
2001.07.15 — For each encoder, noted whether it has scene change detection. (Encoders with this feature will cut a GOP short and start a new one when a frame is dramatically different from the previous ones. This improves quality because the encoder doesn't have to use a difference frame (P or B) to encode an entire frame's worth of changes; difference frames are not the most efficient way to encode an entirely new frame. It also improves the quality of subsequent difference frames, since they now have a clean reference frame to encode their differences against.)
For each software encoder, noted whether it supports multi-CPU encoding.
Added the Pinnacle Systems Studio DV review.
2001.06.29 — Updated the AIW review, adding discussion of the AIW's ambiguous Macrovision situation. (It gets detected with some people's systems, but not on all.)
2001.06.28 — Added the CinemaCraft Encoder SP review.
Added a review of bbMPEG.
2001.06.22 — Expanded the Analysis Method page to include more detailed information on duplicating my results.
2001.06.19 — Major update to the Ligos LSX-MPEG encoder review. I am now testing with the registered LSX-MPEG standalone encoder instead of the demo Premiere plugin. I've added some new things I've discovered about this encoder since the original review, and I added a table showing the differences between the plugin and the standalone encoder.
2001.06.07 — Added a review of the BMK-Elektronik Kfir-based Encoder. This is a specialty board similar to the Hauppauge WinTV PVR, except that it works under Linux. It doesn't have the pretty GUI stuff of the WinTV, but it's a little better quality, and it's the only board under $1000 available for Linux.
2001.05.20 — Added the Conclusions page to summarize the reviews in a more expository way than the Executive Summary table. Also touched up virtually all reviews to be more accurate, more informative, and comparable to the reviews of the closest products in the roundup.
2001.05.19 — Updated the TMPGEnc and Ligos LSX-MPEG reviews. I have recently spent a lot of time running the TMPGEnc encoder off against the Ligos LSX-MPEG standalone encoder; these two encoders can be so similar at times that I thought deeper testing would help me to isolate more differences between the two. In my original review, I simply stated that the two encoders were identical, but this new testing has shown that there are indeed many more differences than speed and price. This update also reflects some new information that I discovered about the standalone Ligos encoder: the original review was done with a demo version of the Premiere plugin, which has turned out not to be entirely representative.
2001.04.21 — Initial release.
| Updated Mon Sep 22 2008 12:15 MDT | Go back to MPEG Encoder Reviews | Go to my home page |