--- title: "Video Quality Analyzer: Frame-by-Frame HLS & DASH Bitstream Analysis" url: https://ireplay.tv/tools/video-quality-analyzer markdown_url: https://ireplay.tv/tools/video-quality-analyzer.md description: "Open any HLS or DASH segment frame by frame: bits per frame, I/P/B frame types, scene cuts and keyframes that waste bitrate, on a per-frame bitrate chart with the picture for every frame. Free no-reference bitstream analysis. Not VMAF, PSNR or SSIM: no source master needed." type: tool site: iReplay.tv pricing: "Free, no account required to run an analysis." keywords: - "video quality analyzer" - "bitstream analyzer" - "frame by frame video analysis" - "bitrate per frame chart" - "bitrate viewer" - "per frame bitrate analysis" - "frame type analysis" - "GOP analyzer" - "I-frame analysis" - "IDR keyframe placement" - "scene cut detection" - "VMAF" - "PSNR" - "SSIM" - "no-reference video quality" - "HLS segment analysis" - "DASH segment analysis" - "keyframe interval" - "B-frame pyramid" - "VBV microburst" - "encoding ladder optimization" --- # Video Quality Analyzer: frame-by-frame HLS and DASH bitstream analysis The iReplay.tv Video Quality Analyzer opens a real segment of your HLS or DASH stream and shows where every bit went. One bar per frame, sized by the bits that frame cost and coloured by whether it is an I, P or B frame, the per-frame bitrate view that desktop bitstream analyzers are normally bought for. Beneath the chart sits the actual picture for every frame, so each verdict can be checked by eye rather than taken on trust. ## This is no-reference analysis, not VMAF, PSNR or SSIM VMAF, PSNR and SSIM are full-reference metrics: each one compares an encode against the original master, frame against frame, and none of them can be computed without that master. You do not upload a master here, so this tool reports none of them. What it does instead is read the encoded bitstream itself and explain the structure behind a quality problem. VMAF tells you how good an encode is; this tells you why, and what to change. ## What it finds - Keyframes sitting mid-segment with no scene change behind them, with the waste quantified in KB, kbps and percent of the segment. - Keyframes correctly placed on a scene cut, which are explicitly *not* counted as waste: at a hard cut a P frame has almost nothing to predict from and costs nearly as much. - Scene cuts the encoder could not afford, where rate control ran out of headroom at the exact moment the picture changed. - IDR versus open-GOP I frames, read at NAL level, and whether B frames are used as references. - Microbursts over 100, 250, 500 and 1000 ms windows, measured against the BANDWIDTH your manifest declares. - Gradual transitions such as dissolves and fades, which no scene-cut detector will flag and which are often the most expensive stretch in a segment. - Frame rate and bitrate the manifest declares against what the media actually contains. ## VOD and live are judged differently The analyzer detects whether the playlist is VOD or live and changes its recommendations accordingly. On VOD a long GOP, open GOP inside segments and a B-pyramid are all free efficiency, and mid-segment keyframes are pure waste. On live the keyframe interval has to be locked to the segment duration and identical across the ladder, B frames cost latency, and the VBV buffer has to be constrained or microbursts reach the player. It supports HLS with MPEG-TS or fMP4/CMAF segments, DASH with SegmentTemplate, SegmentList or SegmentBase addressing, and direct segment URLs. Analysing the manifest instead of the frames? The [Streaming Analyzer](https://ireplay.tv/tools/stream-analyzer) checks your encoding ladder, codec profiles, target duration and DRM signalling. This tool was built by [Sylvain Corvaisier](https://ireplay.tv/hire-me/sylvaincorvaisier), a streaming and video engineer with 20 years in the industry. iReplay.tv is run by a collective of streaming and broadcast engineers; you can also [hire streaming and broadcast specialists](https://ireplay.tv/media-streaming-broadcast-talents). ## What it does - Per-frame bitrate chart coloured by I, P and B frame type - Thumbnail of every frame, aligned to the chart - Scene-change detection with a score per frame - Detects mid-segment keyframes that buy nothing and quantifies the waste - Distinguishes IDR from open-GOP I frames at NAL level - B-pyramid detection - Microburst analysis over 100, 250, 500 and 1000 ms windows - Measured bitrate against the manifest BANDWIDTH declaration - Separate VOD and live recommendations - Generates the ffmpeg / x264 settings that fix what it finds ## Pricing Free, no account required to run an analysis. ## Questions **Is the iReplay Video Quality Analyzer a VMAF score?** No. VMAF is a full-reference metric: it needs both the encode and the original master, and scores how close one is to the other. This tool only ever sees the encode, so it reports no VMAF value. It gives you the reason behind a quality problem rather than a number describing it. **What is the difference between VMAF, PSNR, SSIM and bitstream analysis?** PSNR and SSIM compare an encode against its source mathematically, pixel by pixel. VMAF does the same with a model trained to track human opinion. All three are full-reference and all three need the master file. Bitstream analysis is no-reference: it reads the compressed data itself and reports structure, which is how you find a keyframe that bought nothing or a scene cut the encoder could not afford. **Can I measure video quality without the source file?** You can gather evidence, not a reference score. Any tool claiming a VMAF-equivalent number from the encode alone is estimating. This analyzer reports what it can actually observe: bits per frame, frame types, scene-change scores, keyframe placement and burst rates. **Is there a free browser-based per-frame bitrate analyzer?** The iReplay.tv Video Quality Analyzer is free and runs in the browser. It shows one bar per frame sized by its bits and coloured by frame type, marks scene cuts, and adds the thumbnail of every frame beneath the chart. It works directly on an HLS or DASH URL with no install. **Why does a keyframe in the middle of a segment waste bitrate?** HLS and DASH players switch renditions and seek at segment boundaries, not inside them, so a keyframe placed mid-segment buys no switch point. Unless it lands on a scene change, where a P frame would have cost nearly as much anyway, its extra bits over an ordinary P frame are spent for nothing. On live streams it also suggests the keyframe interval is not locked to the segment duration, which can leave renditions cutting at different points. **Does the analyzer work on DRM-protected streams?** No, and it refuses rather than guessing: a decoder fed encrypted samples without a licence still emits pictures, but they are error-concealment artefacts, so any measurement taken from them would be meaningless. Package a clear, unencrypted copy of the same encode and analyse that instead. Encryption is applied at packaging, after the encoder has made every decision the tool reports on, so the clear copy has identical frame types, frame sizes and keyframe placement. --- Source: [Video Quality Analyzer: Frame-by-Frame HLS & DASH Bitstream Analysis](https://ireplay.tv/tools/video-quality-analyzer) on iReplay.tv. The HTML page is the canonical one to cite, and the only place the tool actually runs. Free to quote and cite with attribution and a link to the source URL. Every iReplay.tv tool, one line each: https://ireplay.tv/llms.txt